R1 (D09) - Data support to COVID-19 RCT (RECOVERY)

The University of Oxford requires access to data for a study entitled Randomised Evaluation of COVid-19 thERapY (RECOVERY).

In 2019 a novel coronavirus-induced disease (COVID-19) emerged in Wuhan, China. A month later the Chinese Center for Disease Control and Prevention identified a new betacoronavirus (SARS (Severe Acute Respiratory Syndrome) coronavirus 2, or SARS-CoV-2) as the aetiological (causing or contributing to the development of a disease or condition) agent. The clinical manifestations of COVID-19 range from asymptomatic infection or mild, transient symptoms to severe viral pneumonia with respiratory failure. It is estimated that by December 2021, 18 million people worldwide had died from COVID-19. RECOVERY was set up in March 2020 as the UK’s national platform trial assessing potential treatments for patients hospitalised with COVID-19. Since then it has delivered over 13 practice changing results (see benefits section) with further results expected.

This study is supported by a grant to the University of Oxford from UK Research and Innovation/National Institute for Health Research (NIHR) and by core funding provided by NIHR Oxford Biomedical Research Centre, the Wellcome Trust, the Bill and Melinda Gates Foundation, Health Data Research UK (HDRUK), and the Medical Research Council Population Health Research Unit, and NIHR Clinical Trials Unit Support Funding. The new trial has been classed as an Urgent Public Health Research Study. It is one of a round of projects to receive £10.5 million as part of the £20 million rapid research response funded by UK Research and Innovation, and by the Department of Health and Social Care through the National Institute for Health Research.

This study aims to compare several different treatments that may be useful for patients with COVID-19. These treatments have been recommended by the expert panel that advises the Chief Medical Officer (CMO) in England.

The study protocol allows reliable assessment of the effects of multiple different treatments (including re-purposed and novel drugs) on major outcomes in COVID-19. All patients will receive usual care for the participating hospital. From version 6.0 of the protocol, a factorial design is being used such that eligible and consenting participants may be randomised simultaneously to one or more of the study treatment arms (depending on location and infection).

Randomisation part A: Eligible patients may be randomly allocated between the following treatment arms (although not all arms may be available at any one time):

• No additional treatment:

• Corticosteroids (children ≤44 weeks gestational age, or >44 weeks gestational age with PIMS-TS only): Children >44 weeks gestational age with SARS-CoV-2 infection without PIMS-TS will be excluded from this arm (as it is standard practice to treat with low-dose corticosteroids).

This comparison has now reported results for adults (see benefits section).

• Azithromycin: Azithromycin is a macrolide antibiotic. Azithromycin has immunomodulatory properties that has shown benefit in inflammatory lung disease.

This comparison has now reported preliminary results (see benefits section).

• Intravenous immunoglobulin (children >44 weeks gestational age with PIMSTS only): a 2 g/kg single dose has been shown to be effective in immunomodulation and preventing cardiovascular complications in Kawasaki disease, an inflammatory condition with overlapping clinical features with PIMS-TS.

• Colchicine (adults ≥18 years old only)

This comparison has now reported results (see benefits section).

• Dimethyl fumarate (UK adults ≥18 years old only; early phase assessment)

This comparison has now closed. Preliminary results have been reported.

Randomisation part B (UK only): Simultaneously, eligible patients will be randomly allocated between the following treatment arms (provided there are no contraindications and the appropriate consent has been given):

• Convalescent plasma: Plasma from patients who have recovered from SARS-CoV2 infection may contain antibodies that can bind to and neutralise the virus.

This comparison has now reported results (see benefits section).

• Synthetic neutralising antibodies: Synthetic human monoclonal antibodies have been developed to bind to and neutralise the virus.

This comparison has now reported results (see benefits section).

Randomisation part C (adults ≥18 years old only): Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• Aspirin

This comparison has now reported results (see benefits section).

Randomisation part D (adults, and children ≥2 years old with COVID-19 pneumonia [UK only]): Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• Baricitinib

This comparison has now reported results (see benefits section).

Randomisation Part E (adults ≥18 years old with hypoxia only). Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• High-dose corticosteroids

This comparison has now reported results for a sub-set of randomised participants (see benefits section).

Randomisation Part F (adults ≥18 years): Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• Empagliflozin

This comparison has now reported results (see benefits section).

Version 9 of the agreement in October 2022 added:

Randomisation Part J (UK only, patients ≥12 years old): Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• Sotrovimab (a monoclonal antibody treatment against the spike protein)

Randomisation Part K (UK only, patients ≥18 years old): Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• Molnupiravir (an antiviral treatment)

Randomisation Part L (UK only, patients ≥18 years old): Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• Paxlovid (an antiviral treatment)

Second randomisation for patients with progressive COVID-19

Participants with progressive COVID-19 (as evidenced by hypoxia and an inflammatory state) may undergo an optional second randomisation between the following treatment arms:

• No additional treatment

• Tocilizumab: (an immunosuppressive drug, mainly for the treatment of rheumatoid arthritis) or control (both in addition to the treatment assigned at the first randomisation).

This comparison has now reported results (see benefits section).

Second randomisation for children with PIMS-TS

Children (at least 1 year old) with PIMS-TS (as evidenced by an exaggerated inflammatory state) may undergo an optional second randomisation between the following treatment arms:

• No additional treatment

• Tocilizumab (children ≥1 <18 years old only)

• Anakinra (children ≥1 <18 years old only)

Modifications to the number of treatment arms: Other arms can be added to the first or second randomisation if evidence emerges that there are suitable candidate therapeutics. Appendix 1 of the trial protocol sets out the rational for each individual treatment comparison currently being evaluated ((https://www.recoverytrial.net) and previous versions of the protocol do the same for treatment comparisons which have been completed.

Conversely, in some patient populations, not all trial arms are appropriate (e.g. due to contraindications based on co-morbid conditions or concomitant medication); in some hospitals or countries, not all treatment arms will be available (e.g. due to manufacturing and supply shortages); and at some times, not all treatment arms will be active (e.g. due to lack of relevant approvals and contractual agreements). The Trial Steering Committee may elect to pause one or more of the arms in order to increase trial efficiency during a fluctuating epidemic. In any of these situations, randomisation will be between fewer arms. Depending on the availability and suitability of treatments, it may be allowed for participants to be randomised in only one part (A or B) of the main randomisation. This will not impact or change the level of data that is required from NHS England - as the cohort in the trial remains the same, regardless of what drug is being trialled.

Data from the trial will be regularly reviewed so that any effective treatment can be identified quickly and made available to all patients. The RECOVERY Trial team will constantly review information on new drugs and include promising ones in the trial.

Patients are eligible for the study if all of the following are true:

(i) Hospitalised

(ii) SARS-CoV-2 infection (clinically suspected or laboratory confirmed)

(iii) No medical history that might, in the opinion of the attending clinician, put the patient at significant risk if he/she were to participate in the trial

In addition, if the attending clinician believes that there is a specific contra-indication to one of the active drug treatment arms or that the patient should definitely be receiving one of the active drug treatment arms then that arm will not be available for randomisation for that patient. For patients who lack capacity, an advanced directive or behaviour that clearly indicates that they would not wish to participate in the trial would be considered sufficient reason to exclude them from the trial.

Informed consent should be obtained from each patient 16 years and over before enrolment into the study. However, if the patient lacks capacity to give consent due to the severity of their medical condition (e.g. acute respiratory failure or need for immediate ventilation) or prior disease, then consent may be obtained from a relative acting as the patient’s legally designated representative or independent doctor. Further consent will then be sought with the patient if they recover sufficiently. For children aged <16 years old consent will be sought from their parents or legal guardian. Where possible, children aged between 10-15 years old will also be asked for assent. Children aged ≥16 years old will be asked for consent as for adults. Witnessed consent may be obtained over the telephone or web video link if hospital visiting rules or parental infection mean a parent/guardian cannot be physically present. Where Local Clinical Centre's (LCC's) plan to recruit children the principal investigator will co-opt support from a local paediatrician and/or neonatologists to oversee the management of children and infants in the trial. PLEASE NOTE: Only data for participants 16+ at randomisation is being requested from NHSD.

By April 2023, the RECOVERY trial had successfully recruited over 48,500 participants, making it one of the largest trials assessing treatments for COVID-19 worldwide. This rapid recruitment and resulting large sample size has enabled the study to produce important results (see yielded benefits section). This remarkable participant recruitment is the result of a number of factors. Eligibility criteria are simple and trial processes (including paperwork) are minimised. Critically, the trial (and subsequent data collection) is designed to minimise the burden on front-line hospital staff working within an overstretched care system during a major pandemic.

The large sample size allows the focus of the COVID-19 Core Protocol to be the impact of candidate treatments on mortality, the key outcome which smaller studies are not able to assess.

For each pairwise comparison with the ‘no additional treatment’ arm, the primary objective is to provide reliable estimates of the effect of study treatments on all-cause mortality at 28 days after first randomisation (with subsidiary analyses of cause of death and of death at various timepoints following discharge).

The secondary objectives are to assess the effects of study treatments on duration of hospital stay; the use of (and duration of) ventilation; and, among patients not on ventilation at baseline, the composite endpoint of death or need for mechanical ventilation or ECMO (extracorporeal membrane oxygenation). Other objectives include the assessment of the effects of study treatments on the use of any ventilation (and duration of invasive mechanical ventilation), acute kidney injury, renal replacement therapy, thrombotic events, and infections. Study outcomes will be assessed based on data recorded up to 28 days and up to 6 months after the main randomisation.

Data from routinely-collected healthcare records (including linkage to medical databases held by organisations such as NHS England) will allow subsidiary analyses of the effect of the study treatments on particular non-fatal events (e.g. ascertained through linkage to Hospital Episode Statistics), the influence of pre-existing major co-morbidity (e.g. diabetes, heart disease, lung disease, hepatic insufficiency, severe depression, severe kidney impairment, immunosuppression), and longer-term outcomes (e.g. six month survival) as well as in particular sub-categories of patient.

Follow-up information is to be collected on all study participants, irrespective of whether or not they complete the scheduled course of allocated study treatment. Study staff will seek follow-up information through various means including medical staff, reviewing information from medical notes, routine healthcare systems, and registries.

ADDITION OF A SUB-STUDY IN VERSION 10 OF THIS AGREEMENT -

A Trial Steering Committee approved sub-study aims to further explore the clinical and biological effects of immunomodulatory therapies within the RECOVERY trial in order to:

• Understand the full spectrum of host and pathogen risk factors which predict outcomes and modify the effects of immunomodulatory therapy in COVID-19

• Describe the long-term clinical effects of immunomodulatory therapy in COVID-19

• Identify potential causal pathways through which the beneficial effects of immunomodulatory treatment for COVID-19 are mediated.

The work will use data from the RECOVERY trial dataset; derived data from RECOVERY’s linked national routine healthcare datasets, including those obtained through NHS England; and from the ISARIC4C external research consortium (https://isaric4c.net/).

The data for ISARIC4C consists of clinical, biological and genetic data on recruited participants to this research consortium and sits within the Scottish National Safe Haven Trusted Research Environment (Public Health Scotland, https://www.isdscotland.org/products-and-services/edris/use-of-the-national-safe-haven/).

The RECOVERY team will combine data from datasets held by Nuffield Department of Public Health (NDPH) to create a project-specific analysis dataset (“The RECOVERY extended analysis dataset” - contents described in Section 5b Processing). This new dataset will be linked with the ISARIC4C datasets by Public Health Scotland’s electronic Data Research and Innovation Service (eDRIS) (https://www.isdscotland.org/Products-and-Services/eDRIS/).

The amendment for Version 10 of this agreement saw the addition of Public Health Scotland as a data processor so they can receive the data for this purpose. Access to NHS England data will be restricted to named researchers approved by the University of Oxford RECOVERY team, who are all substantive employees of the University of Oxford, and system administrators only. The University of Oxford will remain the sole Data Controller. An appropriate controller-processor agreement is in place between the University of Oxford and Public Health Scotland prior to any data transfer.

The study team require the following information from NHS England at the following frequencies (where an end date is given the frequency will be re-evaluated at that time unless otherwise stated):

*** FREQUENCIES UPDATED FOR VERSION 11 OF THIS AGREEMENT***

- SUS (Secondary Use Services) Admitted Patient Care (APC) on hospital discharge. Under v8 this is provided on a monthly basis from September 2021 to September 2022. From v9 this will be provided on a Quarterly basis until March 2023. From v11 one further data drop will be provided in approximately September 2023 after which it will cease flowing.

- HES (Hospital Episode Statistics) APC data. Under v8 this is provided on a Quarterly basis from September 2021 to September 2022. From v9 this will continue to be provided on a Quarterly basis until March 2023. After this (from Amendment v11) data is being requested on an ad-hoc basis in approximately July 2023, September 2023 and March 2024 after which it will become Annual. Each Data Dissemination will include historical data back to March 2015.

- HES Critical Care data Set (CC). Under v8 this is provided on a Quarterly basis from September 2021 to September 2022. From v9 this will continue to be provided on a Quarterly basis until March 2023. After this (from Amendment v11) data is being requested on an ad-hoc basis in approximately July 2023, September 2023 and March 2024 after which it will become Annual. Each Data Dissemination will include historical data back to January 2020.

Emergency Care Data Set (ECDS). Under v8 this is provided on a Quarterly basis from September 2021 to September 2022. From v9 this will continue to be provided on a Quarterly basis until March 2023. After this (from Amendment v11) data is being requested on an ad-hoc basis in approximately September 2023 and March 2024 after which it will become Annual. Each Data Dissemination will include historical data back to January 2020.

- Civil Registrations (Deaths) data (death certificate information). Under v8 this is provided on a monthly basis from September 2021 to September 2022. From v9 there will be three ad-hoc data drops between Sept 2022 and March 2023. After this (from Amendment v11) data is being requested on an ad-hoc basis in approximately September 2023 and March 2024 after which it will become Annual.

- CHESS (COVID-19 Hospitalization in England Surveillance System) Data. Under v8 this is provided on a monthly basis from September 2021 to September 2022. This ceased flowing after September 2022.

- SGSS (COVID-19 SGSS First Positives (Second Generation Surveillance System)) Data. Under v8 this is provided on a monthly basis from September 2021 to September 2022. From v9 there will be three ad-hoc data drops between Sept 2022 and March 2023. From v11 one further data drop will be provided in approximately September 2023 after which it will cease flowing.

- GDPPR Data (General Practice Extraction Service Data for Pandemic Planning and Research). Under v8 this is provided on a monthly basis from September 2021 to September 2022. From v9 there will be three ad-hoc data drops between Sept 2022 and March 2023. After this (from Amendment v11) data is being requested on an ad-hoc basis in approximately September 2023 and March 2024 after which it will become Annual.

- Cancer Registration Data. Under v8 this is provided on an annual basis from September 2021 to September 2022. This continues to be provided on an Annual basis.

- Medicines dispensed in Primary Care NHS Business Services Authority data (NHS BSA). Under v8 this is provided on a monthly basis from September 2021 to September 2022. From v9 there will be three ad-hoc data drops between Sept 2022 and March 2023. After this (from Amendment v11) data is being requested on an ad-hoc basis in approximately September 2023 and March 2024 after which it will become Annual.

- COVID-19 Vaccination Status Data. Under v8 this is provided on a Quarterly basis from September 2021 to September 2022. From v9 this will continue to be provided on a Quarterly basis until March 202. After this (from Amendment v11) data is being requested on an ad-hoc basis in approximately September 2023 and March 2024 after which it will become Annual.

- COVID-19 UK Non-hospital Antigen Testing Results (Pillar 2). Under v8 this is provided on a Monthly basis from September 2021 to September 2022. From v9 this will continue to be provided on a Quarterly basis until March 2023. After this (from Amendment v11) data is being requested on an ad-hoc basis in approximately September 2023 after which it will cease flowing.

- Electronic Prescribing and Medicines Administration (EPMA) data in Secondary Care for COVID-19. Under v8 this is provided on a Quarterly basis from September 2021 to September 2022. From v9 this will continue to be provided on a Quarterly basis until March 2023. After this (from Amendment v11) data is being requested on an ad-hoc basis in approximately September 2023 and March 2024 after which it will cease flowing.

- One off Demographics linkage (Agreed and received under v7).

- Demographics data set (added under version 9 of this agreement) - On a Quarterly basis until March 2023. After this (from Amendment v11) data is being requested on an ad-hoc basis in approximately September 2023 and March 2024 after which it will become Annual. *** NEW FOR VERSION 11 OF THIS AGREEMENT *** The last SUS+ data drop will be in September 2023. After this the study team continue to require deaths data to be contained within the Demographics dataset. The study team request that variables 'Fact of death', ‘Formal date of death’ and ‘Informal date of death’ are included with the Demographics data provision. ***

JUSTIFICATION FOR DATASETS

The GDPPR and Primary Care Medicines data sets are requested for two key reasons. First they will be used to assess whether the effects of the treatments being tested in RECOVERY are different in people with certain medical conditions (e.g. lung disease or diabetes) or who were receiving particular medications (e.g. immunosuppression) prior to joining the trial. To avoid burdening front-line staff, only limited data about other medical conditions is collected from the site staff at the point of enrolment and information about usual medications is not collected. Therefore these data sets will be invaluable in identifying relevant subgroups and will supplement information derived from the HES data regarding prior hospitalisations.

Second, information in these data sets after randomization will be used to determine the effects of the treatments on outcomes after discharge from hospital. For example, effects on the development of long term lung problems, diabetes (particularly relevant to the steroid comparison), liver disease (particularly relevant to the lopinavir/ritonavir comparison) or on infections (particularly relevant to the tocilizumab comparison). It is understood that the GDPPR data is available for a limited time period only but will allow analysis of these and other outcomes following discharge from hospital.

Identifiable data is necessary for the purpose of data linkage and quality assurance of that linkage – the Data Sharing Agreement (DSA) requires measures are taken to ensure subsequent uses of the data will involve only data that has been appropriately pseudonymised. Data requested includes a range of detailed information that is necessary for analysis and sub-group analysis of the effectiveness of the treatments – in particular to enable accounting for a range of biases and confounding factors relevant to the primary and secondary outcomes, specifically:

• The trial has taken a pragmatic approach to recruitment and as a result may have included some patients who do not have confirmed Covid 19. For sub-group analysis and interpretation of the trial results, they need to know which patients have a confirmed positive Covid 19 test.

• Disease onset, dates for admission into hospital, ICU (Intensive Care Unit) and HDU (High Dependency Units) are important to track progression of the disease to understand the efficacy of treatment.

• These data will allow the trial to assess the impact of trial drugs on the precise care and treatment delivered e.g. the need for mechanical ventilation and any complications such as secondary infections.

• Risk factors and underlying conditions enable sub-group analysis of other factors that may have a bearing on the treatments being tested.

The Cancer Registrations dataset will be important in identifying long-term outcomes in the cohort and will help to reliably assess the long-term safety of the treatments tested in RECOVERY. This is particularly important as treatment arms now include medicines which have not previously been assessed in large trials with long follow-up (e.g. tocilizumab) or are currently unlicensed (e.g. REGN-COV2). Cancer Registration data will be provided on an annual basis for the cohort.

The BSA Medicines dataset will be used to further characterise the cohort at baseline, allowing analysis of the effects of the treatments in different types of patient. In addition, refreshes of these data are requested as part of the long-term follow-up as they help to identify new diagnoses (e.g. diabetes). This will help the team identify long-term outcomes in the cohort and assess the effects of the treatments on these outcomes. The processing of this data and use of the BSA Medicines dataset for RECOVERY is in line with the Direction covering that dataset where the purpose is “…to drive the linkage of medicines data with other data sets to provide intelligence about the safety and effectives of medicines”.

The HES Critical Care Dataset is requested to allow comparison with the more rapidly available SUS data. The SUS Critical Care Dataset is used to identify use of organ support following randomisation including use of invasive mechanical ventilation (the secondary outcome for the study). SUS is timely, allowing rapid generation of practice changing data but its quality is not assured to the same extend as HES. Checking for consistencies between SUS Critical Care Dataset and HES Critical Care Dataset would enable further analyses from SUS to be used with confidence and provide reassurance to regulators (e.g. the MHRA) on the validity of the trial results.

The COVID-19 vaccination data will be particularly relevant to the convalescent plasma and Regeneron monoclonal antibody results since these treatments may be less effective in people with prior vaccination. Obtaining linked vaccination data will help understand which patients are likely to benefit from the Regeneron monoclonal antibody.

Linkage to the Covid-19 UK Non-hospital Antigen Testing Results (pillar 2) testing data will enable the study to reliably distinguish those with laboratory proven COVID-19 from those with suspected COVID-19. By the time patients present to hospital PCR testing may be negative and therefore the study may currently underestimate the proportion of participants who have proven SARS-CoV-2 infection. This is important for the interpretation of the study results and acceptance of the findings by regulators.

The Emergency Care Dataset is requested principally to identify safety outcomes occurring in the study participants after discharge from hospital. Most relevant would be infections presenting to A&E. A number of immunosuppressant treatments have been and continue to be tested in RECOVERY and understanding the effects of these treatments on infections following the COVID-19 admission will help doctors to make informed treatment choices for their patients with COVID-19.

The Electronic Prescribing and Medicines Administration (EPMA) data in Secondary Care for COVID-19 dataset is requested for two purposes. First treatments given in hospital will be used to verify or identify complications occurring in the study participants (e.g. infections requiring intravenous antibiotics). Second, these data will be used to assess adherence to the treatments allocated in the study, which is important in the interpretation of the study results.

A one off Personal Demographics Service matching service was requested and provided under version 7 of this agreement. Among the 40,000 participants recruited to date around 400 were recruited in England but have a missing or invalid NHS number recorded by the site staff. In order to allow data to be collected on these consented participants going forward, the RECOVERY team requested a one off patient matching service. The study team provided date of birth, sex and name.

Ongoing Demographics data is being requested for the purpose of identifying individuals during follow-up who no longer have a registered GP. This is because if a study participant no longer has a GP then no further follow-up data can be obtained, so the date of removal of the study participant is needed to calculate the follow-up time for each participant. The variable fields Reason_for_removal, and Date_of_removal are required for this.

RECOVERY PPIE activities:

When RECOVERY was initially being set up in March 2020, there was public involvement via the mechanism of Oxford Population Health’s existing public advisory panel (Nuffield Department of Public Health (NDPH) public advisory panel). They undertook the following tasks:

• Reviewed patient information sheets – 6 NDPH public advisory panel members reviewed

• Reviewed patient consent form – as above

• Reviewed the storyboard of the RECOVERY patient-facing animation – 4 members of NDPH public advisory panel

• Reviewed the first newsletters to trial participants/parents and carers of child participants – 11 people reviewed and the newsletters were revised considerably as a result, splitting out the most scientific detail into a separate section.

During 2021 the trial set up a ‘bespoke’ public advisory group, comprised of patient volunteers who had been participants in RECOVERY. This group has met, in general, every couple of months. The group consists of 10 females and six males, who range in age from late 40s to early 80s. Prior to RECOVERY, none had previous experience of patient and public involvement.

This advisory group has been consulted and reviewed further newsletters to participants, an animation on RECOVERY-related data sharing and has also been involved in the development of a proposal to adapt the RECOVERY platform to trial treatment for flu.

The group reviewed the adapted patient information sheet and consent form that will be used for RECOVERY flu. This resulted in these being simplified as far as was possible. (NB: The study team don’t yet know if RECOVERY flu is going to be funded)

There has also been public engagement around RECOVERY. This includes:

• Media outreach including liaison with participants for case studies/media interviews, 16,100 items of coverage.

• Social media activity (>19,000 uses of #RECOVERYtrial in 2020)

• Development of case studies, e.g. for Understanding Patient Data

• Public-facing webinar with NIHR (participant, husband and clinician in discussion)

• Swindon Science Festival materials (Martin Landray panel event, and online ‘Anatomy of clinical trials’ resource for schools)

• NIHR Oxford Biomedical Research Centre (BRC) Health Open Day – RECOVERY Trial featured in engagement activities with school age children at this small event in Oxford.

NDPH will continue to engage with the RECOVERY PPI Panel and the public throughout the course of the trial.

LEGAL BASIS FOR THE PROCESSING OF DATA

The University of Oxford, as the Data Controller who is also processing the data will process Personal Data under GDPR Article 6 (1) (e) - Processing is necessary for the performance of a task carried out in the public interest or in the exercise of official authority vested in the controller. As a higher education establishment, the University of Oxford conducts research to improve health care and services, and the data requested is necessary for the performance of a task carried out in the public interest. The University of Oxford hopes that the RECOVERY methodology will to improve the speed, scale and quality of trials in COVID-19 and other diseases, providing reliable information on potential treatments for COVID-19 and potentially changing the standard of care that the NHS offers which could improve outcomes for many thousands of patients in the UK and around the world. The RECOVERY extended analysis of immunomodulatory therapies sub-study is anticipated further to lead of the causal biological pathways through which immunomodulatory therapies benefit patients with COVID-19; aiding research into future therapeutics for COVID-19 and therefore contribute to improvements in risk assessment and treatment of NHS patients in the UK.

Additionally, under GDPR Article 9(2)(j) processing of Special Category Personal Data (of which Health data is one) is necessary for archiving for research p

COVID-19 is an emerging pathogen which presents a significant threat to the population in terms of increased morbidity and mortality, particularly among vulnerable groups such as those with pre-existing disease.

The primary objective is to provide reliable estimates of the effect of study treatments on in-hospital death (with subsidiary analyses of cause of death and death at various timepoints following discharge).

The secondary objectives are to assess the effects of study treatments on duration of hospital stay; the need for (and duration of) ventilation; and the need for renal replacement therapy.

The interim trial results will be monitored by an independent Data Monitoring Committee (DMC). The most important task for the DMC will be to assess whether the randomised comparisons in the study have provided evidence on mortality that is strong enough (with a range of uncertainty around the results that is narrow enough) to affect national and global treatment strategies. In such a circumstance, the DMC will inform the Trial Steering Committee who will make the results available to the public and amend the trial arms accordingly.

The data requested will be used to evaluate the efficacy and safety of the study treatments and will help shape the public health response. The rapid feed will be used to ensure that the trial Data Monitoring Committee have complete, up-to-date information on the major outcomes in the trial on which to base their decisions. If they find compelling evidence of efficacy or safety then that arm may be stopped and – if effective – added to standard care across the NHS.

All outputs produced will be in the form of aggregated reports with small number suppression applied.

Outputs delivered to date include the following results publications

RECOVERY Collaborative Group, Horby P, Lim WS, Emberson JR, Mafham M, Bell JL, Linsell L, et al. Dexamethasone in Hospitalized Patients with Covid-19. N Engl J Med 2021; 384:693-704

RECOVERY Collaborative Group. Effect of Hydroxychloroquine in Hospitalized Patients with Covid-19. N Engl J Med. 2020;383(21):2030-2040

RECOVERY Collaborative Group. Lopinavir–ritonavir in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. The Lancet, 2020;396(10259): 1345-1352

RECOVERY Collaborative Group. Azithromycin in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. Lancet. 2021;397(10274):605-612

RECOVERY Collaborative Group. Tocilizumab in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. Lancet. 2021 May 1;397(10285):1637-1645.

RECOVERY Collaborative Group. Convalescent plasma in patients admitted to hospital with COVID-19 (RECOVERY): a randomised controlled, open-label, platform trial. Lancet. 2021 May 29;397(10289):2049-2059

• RECOVERY Collaborative Group. Colchicine in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. Lancet Respir. Med. 2021 Dec 1;9(12):1419-1426 EPub 2021 Oct 18

• RECOVERY Collaborative Group. Aspirin in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. Lancet. 2022 Jan 8;399(10320):143-151 EPub 2021 Nov 17

• RECOVERY collaborative Group. Casirivimab and imdevimab in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. Lancet 2022 Feb 12;399(10325):665-676

• RECOVERY Collaborative Group. Baricitinib in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial and updated meta-analysis. Lancet 2022 July 30;400(10349):359-368

• RECOVERY Collaborative Group. Dimethyl fumarate in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. medRxiv 2022.09.23.22280285

***** NEW FOR VERSION 11*****

• RECOVERY Collaborative Group. Higher dose corticosteroids in patients admitted to hospital with COVID-19 who are hypoxic but not requiring ventilatory support (RECOVERY): a randomised, controlled, open-label, platform trial. Lancet 2023 Apr 22; 401(10385):1313-1398

• RECOVERY Collaborative Group. Empagliflozin in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. medRxiv 2023.04.13.23288469

By the end of July 2021, the RECOVERY trial had successfully established over 175 sites across the UK and recruited over 40,000 participants treated in hospital for COVID-19. The data linkage already established to received SUS+ and other data is providing important information to the Data Monitoring Committee on a regular basis about patients' recovery (i.e. discharge from hospital), in-hospital death and procedures required. Provision of complete and reliable data to the DMC throughout the trial is critical to allow robust assessment of the effects of the trial treatments with a major contribution to these data expected from analysis of the routine health care data requested under this agreement.

On Thursday 4 June 2020, in response to a request from the UK Medicines and Healthcare Products Regulatory Agency (MHRA), the independent Data Monitoring Committee conducted a further review of the data. The DMC recommended the chief investigators review the unblinded data on the hydroxychloroquine arm of the trial. The results included a total of 1542 patients randomised to hydroxychloroquine compared with 3132 patients randomised to usual care alone. There was no significant difference in the primary endpoint of 28-day mortality (25.7% hydroxychloroquine vs. 23.5% usual care; hazard ratio 1.11 [95% confidence interval 0.98-1.26]; p=0.10). There was also no evidence of beneficial effects on hospital stay duration or other outcomes. These results were released to the public and on the 15 July 2020, the U.S. Food and Drug Administration (FDA) revoked the emergency use authorization (EUA) that allowed for chloroquine phosphate and hydroxychloroquine sulfate donated to the Strategic National Stockpile to be used to treat certain hospitalized patients with COVID-19 when a clinical trial was unavailable, or participation in a clinical trial was not feasible. This result has major implications for countries around the world who were planning to scale manufacturing of these drugs in order to treat COVID patients.

On 8 June 2020, recruitment to the dexamethasone arm was halted since, in the view of the trial Steering Committee, sufficient patients had been enrolled to establish whether or not the drug had a meaningful benefit. On 16 June 2020, preliminary results were released. A total of 2104 patients were randomised to receive dexamethasone 6 mg once per day (either by mouth or by intravenous injection) for ten days and were compared with 4321 patients randomised to usual care alone. Among the patients who received usual care alone, 28-day mortality was highest in those who required ventilation (41%), intermediate in those patients who required oxygen only (25%), and lowest among those who did not require any respiratory intervention (13%). Dexamethasone reduced deaths by one-third in ventilated patients (rate ratio 0.65 [95% confidence interval 0.48 to 0.88]; p=0.0003) and by one fifth in other patients receiving oxygen only (0.80 [0.67 to 0.96]; p=0.0021). There was no benefit among those patients who did not require respiratory support (1.22 [0.86 to 1.75]; p=0.14). Based on these results, 1 death would be prevented by treatment of around 8 ventilated patients or around 25 patients requiring oxygen alone.

On the 16 June 2020, on the basis of these results, the MHRA issued an alert to health care providers in the UK recommending the use of dexamethasone in hospitalised patients with COVID who require oxygen or ventilation. In addition dexamethasone has also been added to the government’s parallel export list, which bans companies from buying medicines meant for UK patients and selling them on for a higher price in another country.

An external analysis estimated that treatment with Dexamethasone may have saved 22,000 lives in the UK and 1 million lives worldwide up to March 2021 (most recent estimates available at: ((https://www.england.nhs.uk/2021/03/covid-treatment-developed-in-the-nhs-saves-a-million-lives/))

Without the results from RECOVERY the effectiveness of this treatment would not be known.

On 11 February 2021, analyses of the RECOVERY trial showed that, among hospitalised COVID-19 patients who had low oxygen levels and elevated markers of inflammation in the blood, the rheumatoid arthritis drug Tocilizumab significantly reduced deaths: 596 (29%) of the patients in the tocilizumab group died within 28 days compared with 694 (33%) patients in the usual care group (rate ratio 0·86; [95% confidence interval [CI] 0·77 to 0·96]; p=0·007), an absolute difference of 4%. This means that for every 25 patients treated with tocilizumab, one additional life would be saved. Tocilizumab also increased the probability of discharge alive within 28 days from 47% to 54% (rate ratio 1·23, [95% CI 1·12 to 1·34], p<0·0001). These benefits were seen in all patient subgroups, including those requiring oxygen via a simple face mask through to those requiring mechanical ventilators in an intensive care unit and those already receiving steroid treatment. Tocilizumab is now standard of care within the NHS. In December 2022, the US Food and Drug Administration approved a new indication for tocilizumab for the treatment of patients with COVID-19 based on data from trials including RECOVERY FDA Roundup: December 23, 2022 | FDA.

On 16 June 2021, analysis of 9785 patients hospitalised with COVID-19 randomly allocated to receive usual care plus the antibody combination treatment (casirivimab 4g with imdevimab 4g by intravenous infusion) or usual care alone produced practice changing results. Among the one-third of participants who were seronegative at baseline (ie they had not mounted a natural antibody response of their own) the monoclonal antibody combination significantly reduced the primary outcome of 28-day mortality by one-fifth compared with usual care alone (24% of patients in the antibody combination group died vs 30% of patients in the usual care group; rate ratio 0·80; 95% confidence interval 0·70–0·91; p=0·001). Thus, for every 100 such patients treated with the antibody combination, there would be six fewer deaths.

On 3 March 2022, analyses of 8,156 patients hospitalized with COVID-19 randomly allocated to receive usual care plus baricitinib, or usual care alone produced positive results. Treatment with baricitinib significantly reduced deaths: 513 (12%) of the patients in the baricitinib group died within 28 days compared with 546 (14%) patients in the usual care group, a reduction of 13% (age-adjusted rate ratio 0.87, 95% confidence interval [CI] 0.77 to 0.98; p= 0.026). The benefit of baricitinib was consistent regardless of which other COVID-19 treatments the patients were also receiving, including corticosteroids, tocilizumab, or remdesivir. This result provides new evidence of the additional beneficial of baricitinib on top of other immunomodulatory treatments. The MHRA have issued an alert to health care providers in the UK recommending the use of baricitinib as an additional treatment option for patients admitted to hospital with COVID-19.

***** NEW FOR VERSION 11*****

On 12 April 2023, analyses of 1,272 patients hospitalized with COVID-19 (with hypoxia receiving no oxygen, or simple oxygen) randomly allocated to receive usual care plus high-dose corticosteroids, versus usual care alone were reported. It was found that participants receiving higher dose corticosteroids had a significantly increased risk of death compared to usual care (which included low-dose corticosteroids).

RECOVERY has also shown that a number of widely used treatments are not effective in COVID-19 including aspirin, colchicine, lopinavir-ritonavir, Convalescent plasma, and empagliflozin. These findings are important since they avoid unnecessary exposure to potentially hazardous treatments and allow healthcare systems and researchers to focus resources on treatments which are effective.

***********

The University of Oxford requires access to data for a study entitled Randomised Evaluation of COVid-19 thERapY (RECOVERY).

In 2019 a novel coronavirus-induced disease (COVID-19) emerged in Wuhan, China. A month later the Chinese Center for Disease Control and Prevention identified a new betacoronavirus (SARS (Severe Acute Respiratory Syndrome) coronavirus 2, or SARS-CoV-2) as the aetiological (causing or contributing to the development of a disease or condition) agent. The clinical manifestations of COVID-19 range from asymptomatic infection or mild, transient symptoms to severe viral pneumonia with respiratory failure. It is estimated that by December 2021, 18 million people worldwide had died from COVID-19. RECOVERY was set up in March 2020 as the UK’s national platform trial assessing potential treatments for patients hospitalised with COVID-19. Since then it has delivered over 11 practice changing results (see benefits section) with further results expected.

This study is supported by a grant to the University of Oxford from UK Research and Innovation/National Institute for Health Research (NIHR) and by core funding provided by NIHR Oxford Biomedical Research Centre, the Wellcome Trust, the Bill and Melinda Gates Foundation, Health Data Research UK (HDRUK), and the Medical Research Council Population Health Research Unit, and NIHR Clinical Trials Unit Support Funding. The new trial has been classed as an Urgent Public Health Research Study. It is one of a round of projects to receive £10.5 million as part of the £20 million rapid research response funded by UK Research and Innovation, and by the Department of Health and Social Care through the National Institute for Health Research.

This study aims to compare several different treatments that may be useful for patients with COVID-19. These treatments have been recommended by the expert panel that advises the Chief Medical Officer (CMO) in England.

The study protocol allows reliable assessment of the effects of multiple different treatments (including re-purposed and novel drugs) on major outcomes in COVID-19. All patients will receive usual care for the participating hospital. From version 6.0 of the protocol, a factorial design is being used such that eligible and consenting participants may be randomised simultaneously to one or more of the study treatment arms (depending on location and infection).

Randomisation part A: Eligible patients may be randomly allocated between the following treatment arms (although not all arms may be available at any one time):

• No additional treatment:

• Corticosteroids (children ≤44 weeks gestational age, or >44 weeks gestational age with PIMS-TS only): Children >44 weeks gestational age with SARS-CoV-2 infection without PIMS-TS will be excluded from this arm (as it is standard practice to treat with low-dose corticosteroids).

This comparison has now reported results for adults (see benefits section).

• Azithromycin: Azithromycin is a macrolide antibiotic. Azithromycin has immunomodulatory properties that has shown benefit in inflammatory lung disease.

This comparison has now reported preliminary results (see benefits section).

• Intravenous immunoglobulin (children >44 weeks gestational age with PIMSTS only): a 2 g/kg single dose has been shown to be effective in immunomodulation and preventing cardiovascular complications in Kawasaki disease, an inflammatory condition with overlapping clinical features with PIMS-TS.

• Colchicine (adults ≥18 years old only)

This comparison has now reported results (see benefits section).

• Dimethyl fumarate (UK adults ≥18 years old only; early phase assessment)

This comparison has now closed. Preliminary results have been reported.

Randomisation part B (UK only): Simultaneously, eligible patients will be randomly allocated between the following treatment arms (provided there are no contraindications and the appropriate consent has been given):

• Convalescent plasma: Plasma from patients who have recovered from SARS-CoV2 infection may contain antibodies that can bind to and neutralise the virus.

This comparison has now reported results (see benefits section).

• Synthetic neutralising antibodies: Synthetic human monoclonal antibodies have been developed to bind to and neutralise the virus.

This comparison has now reported results (see benefits section).

Randomisation part C (adults ≥18 years old only): Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• Aspirin

This comparison has now reported results (see benefits section).

Randomisation part D (adults, and children ≥2 years old with COVID-19 pneumonia [UK only]): Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• Baricitinib

This comparison has now reported results (see benefits section).

Randomisation Part E (adults ≥18 years old with hypoxia only). Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• High-dose corticosteroids

Randomisation Part F (adults ≥18 years): Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• Empagliflozin

Version 9 of the agreement in October 2022 added:

Randomisation Part J (UK only, patients ≥12 years old): Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• Sotrovimab (a monoclonal antibody treatment against the spike protein)

Randomisation Part K (UK only, patients ≥18 years old): Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• Molnupiravir (an antiviral treatment)

Randomisation Part L (UK only, patients ≥18 years old): Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• Paxlovid (an antiviral treatment)

Second randomisation for patients with progressive COVID-19

Participants with progressive COVID-19 (as evidenced by hypoxia and an inflammatory state) may undergo an optional second randomisation between the following treatment arms:

• No additional treatment

• Tocilizumab: (an immunosuppressive drug, mainly for the treatment of rheumatoid arthritis) or control (both in addition to the treatment assigned at the first randomisation).

This comparison has now reported results (see benefits section).

Second randomisation for children with PIMS-TS

Children (at least 1 year old) with PIMS-TS (as evidenced by an exaggerated inflammatory state) may undergo an optional second randomisation between the following treatment arms:

• No additional treatment

• Tocilizumab (children ≥1 <18 years old only)

• Anakinra (children ≥1 <18 years old only)

Modifications to the number of treatment arms: Other arms can be added to the first or second randomisation if evidence emerges that there are suitable candidate therapeutics. Appendix 1 of the trial protocol sets out the rational for each individual treatment comparison currently being evaluated ((https://www.recoverytrial.net) and previous versions of the protocol do the same for treatment comparisons which have been completed.

Conversely, in some patient populations, not all trial arms are appropriate (e.g. due to contraindications based on co-morbid conditions or concomitant medication); in some hospitals or countries, not all treatment arms will be available (e.g. due to manufacturing and supply shortages); and at some times, not all treatment arms will be active (e.g. due to lack of relevant approvals and contractual agreements). The Trial Steering Committee may elect to pause one or more of the arms in order to increase trial efficiency during a fluctuating epidemic. In any of these situations, randomisation will be between fewer arms. Depending on the availability and suitability of treatments, it may be allowed for participants to be randomised in only one part (A or B) of the main randomisation. This will not impact or change the level of data that is required from NHS Digital - as the cohort in the trial remains the same, regardless of what drug is being trialled.

Data from the trial will be regularly reviewed so that any effective treatment can be identified quickly and made available to all patients. The RECOVERY Trial team will constantly review information on new drugs and include promising ones in the trial.

Patients are eligible for the study if all of the following are true:

(i) Hospitalised

(ii) SARS-CoV-2 infection (clinically suspected or laboratory confirmed)

(iii) No medical history that might, in the opinion of the attending clinician, put the patient at significant risk if he/she were to participate in the trial

In addition, if the attending clinician believes that there is a specific contra-indication to one of the active drug treatment arms or that the patient should definitely be receiving one of the active drug treatment arms then that arm will not be available for randomisation for that patient. For patients who lack capacity, an advanced directive or behaviour that clearly indicates that they would not wish to participate in the trial would be considered sufficient reason to exclude them from the trial.

Informed consent should be obtained from each patient 16 years and over before enrolment into the study. However, if the patient lacks capacity to give consent due to the severity of their medical condition (e.g. acute respiratory failure or need for immediate ventilation) or prior disease, then consent may be obtained from a relative acting as the patient’s legally designated representative or independent doctor. Further consent will then be sought with the patient if they recover sufficiently. For children aged <16 years old consent will be sought from their parents or legal guardian. Where possible, children aged between 10-15 years old will also be asked for assent. Children aged ≥16 years old will be asked for consent as for adults. Witnessed consent may be obtained over the telephone or web video link if hospital visiting rules or parental infection mean a parent/guardian cannot be physically present. Where Local Clinical Centre's (LCC's) plan to recruit children the principal investigator will co-opt support from a local paediatrician and/or neonatologists to oversee the management of children and infants in the trial. PLEASE NOTE: Only data for participants 16+ at randomisation is being requested from NHSD.

By October 2022, the RECOVERY trial had successfully recruited over 48,000 participants, making it one of the largest trials assessing treatments for COVID-19 worldwide. This rapid recruitment and resulting large sample size has enabled the study to produce important results (see yielded benefits section). This remarkable participant recruitment is the result of a number of factors. Eligibility criteria are simple and trial processes (including paperwork) are minimised. Critically, the trial (and subsequent data collection) is designed to minimise the burden on front-line hospital staff working within an overstretched care system during a major pandemic.

The large sample size allows the focus of the COVID-19 Core Protocol to be the impact of candidate treatments on mortality, the key outcome which smaller studies are not able to assess.

For each pairwise comparison with the ‘no additional treatment’ arm, the primary objective is to provide reliable estimates of the effect of study treatments on all-cause mortality at 28 days after first randomisation (with subsidiary analyses of cause of death and of death at various timepoints following discharge).

The secondary objectives are to assess the effects of study treatments on duration of hospital stay; the use of (and duration of) ventilation; and, among patients not on ventilation at baseline, the composite endpoint of death or need for mechanical ventilation or ECMO (extracorporeal membrane oxygenation). Other objectives include the assessment of the effects of study treatments on the use of any ventilation (and duration of invasive mechanical ventilation), acute kidney injury, renal replacement therapy, thrombotic events, and infections. Study outcomes will be assessed based on data recorded up to 28 days and up to 6 months after the main randomisation.

Data from routinely-collected healthcare records (including linkage to medical databases held by organisations such as NHS Digital) will allow subsidiary analyses of the effect of the study treatments on particular non-fatal events (e.g. ascertained through linkage to Hospital Episode Statistics), the influence of pre-existing major co-morbidity (e.g. diabetes, heart disease, lung disease, hepatic insufficiency, severe depression, severe kidney impairment, immunosuppression), and longer-term outcomes (e.g. six month survival) as well as in particular sub-categories of patient.

Follow-up information is to be collected on all study participants, irrespective of whether or not they complete the scheduled course of allocated study treatment. Study staff will seek follow-up information through various means including medical staff, reviewing information from medical notes, routine healthcare systems, and registries.

**** VERSION 10 AMENDMENT - ADDITION OF A SUB-STUDY ****

A Trial Steering Committee approved sub-study aims to further explore the clinical and biological effects of immunomodulatory therapies within the RECOVERY trial in order to:

• Understand the full spectrum of host and pathogen risk factors which predict outcomes and modify the effects of immunomodulatory therapy in COVID-19

• Describe the long-term clinical effects of immunomodulatory therapy in COVID-19

• Identify potential causal pathways through which the beneficial effects of immunomodulatory treatment for COVID-19 are mediated.

The work will use data from the RECOVERY trial dataset; derived data from RECOVERY’s linked national routine healthcare datasets, including those obtained through NHS Digital; and from the ISARIC4C external research consortium (https://isaric4c.net/).

The data for ISARIC4C consists of clinical, biological and genetic data on recruited participants to this research consortium and sits within the Scottish National Safe Haven Trusted Research Environment (Public Health Scotland, https://www.isdscotland.org/products-and-services/edris/use-of-the-national-safe-haven/).

The RECOVERY team will combine data from datasets held by Nuffield Department of Public Health (NDPH) to create a project-specific analysis dataset (“The RECOVERY extended analysis dataset” - contents described in Section 5b Processing). This new dataset will be linked with the ISARIC4C datasets by Public Health Scotland’s electronic Data Research and Innovation Service (eDRIS) (https://www.isdscotland.org/Products-and-Services/eDRIS/).

This amendment (version 10) adds Public Health Scotland as a data processor so they can receive the data for this purpose. Access to NHS Digital data will be restricted to named researchers approved by the University of Oxford RECOVERY team, who are all substantive employees of the University of Oxford, and system administrators only. The University of Oxford will remain the sole Data Controller. An appropriate controller-processor agreement is in place between the University of Oxford and Public Health Scotland prior to any data transfer.

**********************************************************************

The study team require the following information from NHS Digital at the following frequencies (where an end date is given the frequency will be re-evaluated at that time unless otherwise stated):

- SUS (Secondary Use Services) Admitted Patient Care (APC) on hospital discharge. Under v8 this is provided on a monthly basis from September 2021 to September 2022. From v9 this will be provided on a Quarterly basis until March 2023 when it will cease flowing.

- HES (Hospital Episode Statistics) APC data. Under v8 this is provided on a Quarterly basis from September 2021 to September 2022. From v9 this will continue to be provided on a Quarterly basis until March 2023 when it will become Annual.

- HES Critical Care data Set (CC). Under v8 this is provided on a Quarterly basis from September 2021 to September 2022. From v9 this will continue to be provided on a Quarterly basis until March 2023 when it will become Annual.

Emergency Care Data Set (ECDS). Under v8 this is provided on a Quarterly basis from September 2021 to September 2022. From v9 this will continue to be provided on a Quarterly basis until March 2023 when it will become Annual.

- Civil Registrations (Deaths) data (death certificate information). Under v8 this is provided on a monthly basis from September 2021 to September 2022. From v9 there will be three ad-hoc data drops between Sept 2022 and March 2023 when it will become Annual.

- CHESS (COVID-19 Hospitalization in England Surveillance System) Data. Under v8 this is provided on a monthly basis from September 2021 to September 2022. This ceased flowing after September 2022.

- SGSS (COVID-19 Second Generation Surveillance System) Data. Under v8 this is provided on a monthly basis from September 2021 to September 2022. From v9 there will be three ad-hoc data drops between Sept 2022 and March 2023 when it will cease flowing.

- GDPPR Data (General Practice Extraction Service Data for Pandemic Planning and Research). Under v8 this is provided on a monthly basis from September 2021 to September 2022. From v9 there will be three ad-hoc data drops between Sept 2022 and March 2023 when it will become Annual.

- Cancer Registration Data. Under v8 this is provided on an annual basis from September 2021 to September 2022. This will continue to be provided on an Annual basis from v9.

- Medicines dispensed in Primary Care NHS Business Services Authority data (NHS BSA). Under v8 this is provided on a monthly basis from September 2021 to September 2022. From v9 there will be three ad-hoc data drops between Sept 2022 and March 2023 when it will become Annual.

- COVID-19 Vaccination Status Data. Under v8 this is provided on a Quarterly basis from September 2021 to September 2022. From v9 this will continue to be provided on a Quarterly basis until March 2023 when it will cease flowing.

- COVID-19 UK Non-hospital Antigen Testing Results (Pillar 2). Under v8 this is provided on a Monthly basis from September 2021 to September 2022. From v9 this will continue to be provided on a Quarterly basis until March 2023 when it will cease flowing.

- Electronic Prescribing and Medicines Administration (EPMA) data in Secondary Care for COVID-19. Under v8 this is provided on a Quarterly basis from September 2021 to September 2022. From v9 this will continue to be provided on a Quarterly basis until March 2023 when it will cease flowing.

- One off Demographics linkage (Agreed and received under v7).

- Demographics data set (added under version 9 of this agreement) - On a Quarterly basis until March 2023 when it will become Annual.

JUSTIFICATION FOR DATASETS

The GDPPR and Primary Care Medicines data sets are requested for two key reasons. First they will be used to assess whether the effects of the treatments being tested in RECOVERY are different in people with certain medical conditions (e.g. lung disease or diabetes) or who were receiving particular medications (e.g. immunosuppression) prior to joining the trial. To avoid burdening front-line staff, only limited data about other medical conditions is collected from the site staff at the point of enrolment and information about usual medications is not collected. Therefore these data sets will be invaluable in identifying relevant subgroups and will supplement information derived from the HES data regarding prior hospitalisations.

Second, information in these data sets after randomization will be used to determine the effects of the treatments on outcomes after discharge from hospital. For example, effects on the development of long term lung problems, diabetes (particularly relevant to the steroid comparison), liver disease (particularly relevant to the lopinavir/ritonavir comparison) or on infections (particularly relevant to the tocilizumab comparison). It is understood that the GDPPR data is available for a limited time period only but will allow analysis of these and other outcomes following discharge from hospital.

Identifiable data is necessary for the purpose of data linkage and quality assurance of that linkage – the Data Sharing Agreement (DSA) requires measures are taken to ensure subsequent uses of the data will involve only data that has been appropriately pseudonymised. Data requested includes a range of detailed information that is necessary for analysis and sub-group analysis of the effectiveness of the treatments – in particular to enable accounting for a range of biases and confounding factors relevant to the primary and secondary outcomes, specifically:

• The trial has taken a pragmatic approach to recruitment and as a result may have included some patients who do not have confirmed Covid 19. For sub-group analysis and interpretation of the trial results, they need to know which patients have a confirmed positive Covid 19 test.

• Disease onset, dates for admission into hospital, ICU (Intensive Care Unit) and HDU (High Dependency Units) are important to track progression of the disease to understand the efficacy of treatment.

• These data will allow the trial to assess the impact of trial drugs on the precise care and treatment delivered e.g. the need for mechanical ventilation and any complications such as secondary infections.

• Risk factors and underlying conditions enable sub-group analysis of other factors that may have a bearing on the treatments being tested.

The Cancer Registrations dataset will be important in identifying long-term outcomes in the cohort and will help to reliably assess the long-term safety of the treatments tested in RECOVERY. This is particularly important as treatment arms now include medicines which have not previously been assessed in large trials with long follow-up (e.g. tocilizumab) or are currently unlicensed (e.g. REGN-COV2). Cancer Registration data will be provided on an annual basis for the cohort.

The BSA Medicines dataset will be used to further characterise the cohort at baseline, allowing analysis of the effects of the treatments in different types of patient. In addition, refreshes of these data are requested as part of the long-term follow-up as they help to identify new diagnoses (e.g. diabetes). This will help the team identify long-term outcomes in the cohort and assess the effects of the treatments on these outcomes. The processing of this data and use of the BSA Medicines dataset for RECOVERY is in line with the Direction covering that dataset where the purpose is “…to drive the linkage of medicines data with other data sets to provide intelligence about the safety and effectives of medicines”.

The HES Critical Care Dataset is requested to allow comparison with the more rapidly available SUS data. The SUS Critical Care Dataset is used to identify use of organ support following randomisation including use of invasive mechanical ventilation (the secondary outcome for the study). SUS is timely, allowing rapid generation of practice changing data but its quality is not assured to the same extend as HES. Checking for consistencies between SUS Critical Care Dataset and HES Critical Care Dataset would enable further analyses from SUS to be used with confidence and provide reassurance to regulators (e.g. the MHRA) on the validity of the trial results.

The COVID-19 vaccination data will be particularly relevant to the convalescent plasma and Regeneron monoclonal antibody results since these treatments may be less effective in people with prior vaccination. Obtaining linked vaccination data will help understand which patients are likely to benefit from the Regeneron monoclonal antibody.

Linkage to the Covid-19 UK Non-hospital Antigen Testing Results (pillar 2) testing data will enable the study to reliably distinguish those with laboratory proven COVID-19 from those with suspected COVID-19. By the time patients present to hospital PCR testing may be negative and therefore the study may currently underestimate the proportion of participants who have proven SARS-CoV-2 infection. This is important for the interpretation of the study results and acceptance of the findings by regulators.

The Emergency Care Dataset is requested principally to identify safety outcomes occurring in the study participants after discharge from hospital. Most relevant would be infections presenting to A&E. A number of immunosuppressant treatments have been and continue to be tested in RECOVERY and understanding the effects of these treatments on infections following the COVID-19 admission will help doctors to make informed treatment choices for their patients with COVID-19.

The Electronic Prescribing and Medicines Administration (EPMA) data in Secondary Care for COVID-19 dataset is requested for two purposes. First treatments given in hospital will be used to verify or identify complications occurring in the study participants (e.g. infections requiring intravenous antibiotics). Second, these data will be used to assess adherence to the treatments allocated in the study, which is important in the interpretation of the study results.

A one off Personal Demographics Service matching service is requested. Among the 40,000 participants recruited to date around 400 were recruited in England but have a missing or invalid NHS number recorded by the site staff. In order to allow data to be collected on these consented participants going forward, the RECOVERY team are requesting a one off patient matching service. The study team will provide date of birth, sex and name.

Quarterly Demographics data is being requested for the purpose of identifying individuals during follow-up who no longer have a registered GP. This is because if a study participant no longer has a GP then no further follow-up data can be obtained, so the date of removal of the study participant is needed to calculate the follow-up time for each participant. The variable fields Reason_for_removal, and Date_of_removal are required for this.

RECOVERY PPIE activities:

When RECOVERY was initially being set up in March 2020, there was public involvement via the mechanism of Oxford Population Health’s existing public advisory panel (Nuffield Department of Public Health (NDPH) public advisory panel). They undertook the following tasks:

• Reviewed patient information sheets – 6 NDPH public advisory panel members reviewed

• Reviewed patient consent form – as above

• Reviewed the storyboard of the RECOVERY patient-facing animation – 4 members of NDPH public advisory panel

• Reviewed the first newsletters to trial participants/parents and carers of child participants – 11 people reviewed and the newsletters were revised considerably as a result, splitting out the most scientific detail into a separate section.

During 2021 the trial set up a ‘bespoke’ public advisory group, comprised of patient volunteers who had been participants in RECOVERY. This group has met, in general, every couple of months. The group consists of 10 females and six males, who range in age from late 40s to early 80s. Prior to RECOVERY, none had previous experience of patient and public involvement.

This advisory group has been consulted and reviewed further newsletters to participants, an animation on RECOVERY-related data sharing and has also been involved in the development of a proposal to adapt the RECOVERY platform to trial treatment for flu.

The group reviewed the adapted patient information sheet and consent form that will be used for RECOVERY flu. This resulted in these being simplified as far as was possible. (NB: The study team don’t yet know if RECOVERY flu is going to be funded)

There has also been public engagement around RECOVERY. This includes:

• Media outreach including liaison with participants for case studies/media interviews, 16,100 items of coverage.

• Social media activity (>19,000 uses of #RECOVERYtrial in 2020)

• Development of case studies, e.g. for Understanding Patient Data

• Public-facing webinar with NIHR (participant, husband and clinician in discussion)

• Swindon Science Festival materials (Martin Landray panel event, and online ‘Anatomy of clinical trials’ resource for schools)

• NIHR Oxford Biomedical Research Centre (BRC) Health Open Day – RECOVERY Trial featured in engagement activities with school age children at this small event in Oxford.

NDPH will continue to engage with the RECOVERY PPI Panel and the public throughout the course of the trial.

LEGAL BASIS FOR THE PROCESSING OF DATA

The University of Oxford, as the Data Controller who is also processing the data will process Personal Data under GDPR Article 6 (1) (e) - Processing is necessary for the performance of a task carried out in the public interest or in the exercise of official authority vested in the controller. As a higher education establishment, the University of Oxford conducts research to improve health care and services, and the data requested is necessary for the performance of a task carried out in the public interest. The University of Oxford hopes that the RECOVERY methodology will to improve the speed, scale and quality of trials in COVID-19 and other diseases, providing reliable information on potential treatments for COVID-19 and potentially changing the standard of care that the NHS offers which could improve outcomes for many thousands of patients in the UK and around the world. The RECOVERY extended analysis of immunomodulatory therapies sub-study is anticipated further to lead of the causal biological pathways through which immunomodulatory therapies benefit patients with COVID-19; aiding research into future therapeutics for COVID-19 and therefore contribute to improvements in risk assessment and treatment of NHS patients in the UK.

Additionally, under GDPR Article 9(2)(j) processing of Special Category Personal Data (of which Health data is one) is necessary for archiving for research purposes. Data minimisation processes are being followed and only data that is specifically required for the purposes of this study have been requested, to protect the rights of the data subjects. The data are required for research purposes in the public interest – meeting the conditions in the DPA 2018 Schedule 1 Part 1 (4) – which GDPR Recital 52(2) determines is an appropriate derogation from the prohibition on processing special categories of personal data.

In accordance with GDPR Article 89(1) processing is subject to appropriate safeguards. These include:

- The data recipient’s technical and organisational measures to safeguard the data have been assessed and meet NHS Digital’s acceptance criteria (see sections 2 and 5b of this application for further details;

- The requested data have been assessed as proportionate to the aim pursued (see section 5a of this application for further details);

Controls, data retention and processing activities have been assessed to ensure respect to the essence of the right to data protection (see sections 5a, 5b and 8a of this application for further details);

- Measures to protect the rights and freedoms of data subjects have been assessed including transparency (fair processing) publishing subject’s rights to withdraw consent and /or have their data erased or rectified, etc.

Onward sharing of data (ADaM and SDTM) for manufacturers / regulators:

The University of Oxford has designed two types of data outputs which the RECOVERY trial needs to onwardly share:

• Analysis Data Model (ADaM) output

• Study Data Tabulation Model (SDTM) output

These trial analysis datasets, are developed using data provided by NHS Digital under this Data Sharing Agreement along with data from other sources, and are subject to a range of processing in line with standards published by the Clinical Data Interchange Standards Consortium (or CDISC).

The ADaM and SDTM outputs for the RECOVERY trial have undergone technical assessment and IG assessment by NHS Digital and NHS Digital has determined that, subject to being shared under strict conditions (specified below), the datasets are sufficiently low risk and can be onward shared by the University of Oxford with the manufacturer or regulator where necessary.

ADaM output:

Based on a desk review and further inform

COVID-19 is an emerging pathogen which presents a significant threat to the population in terms of increased morbidity and mortality, particularly among vulnerable groups such as those with pre-existing disease.

The primary objective is to provide reliable estimates of the effect of study treatments on in-hospital death (with subsidiary analyses of cause of death and death at various timepoints following discharge).

The secondary objectives are to assess the effects of study treatments on duration of hospital stay; the need for (and duration of) ventilation; and the need for renal replacement therapy.

The interim trial results will be monitored by an independent Data Monitoring Committee (DMC). The most important task for the DMC will be to assess whether the randomised comparisons in the study have provided evidence on mortality that is strong enough (with a range of uncertainty around the results that is narrow enough) to affect national and global treatment strategies. In such a circumstance, the DMC will inform the Trial Steering Committee who will make the results available to the public and amend the trial arms accordingly.

The data requested will be used to evaluate the efficacy and safety of the study treatments and will help shape the public health response. The rapid feed will be used to ensure that the trial Data Monitoring Committee have complete, up-to-date information on the major outcomes in the trial on which to base their decisions. If they find compelling evidence of efficacy or safety then that arm may be stopped and – if effective – added to standard care across the NHS.

All outputs produced will be in the form of aggregated reports with small number suppression applied.

Outputs delivered to date include the following results publications

RECOVERY Collaborative Group, Horby P, Lim WS, Emberson JR, Mafham M, Bell JL, Linsell L, et al. Dexamethasone in Hospitalized Patients with Covid-19. N Engl J Med 2021; 384:693-704

RECOVERY Collaborative Group. Effect of Hydroxychloroquine in Hospitalized Patients with Covid-19. N Engl J Med. 2020;383(21):2030-2040

RECOVERY Collaborative Group. Lopinavir–ritonavir in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. The Lancet, 2020;396(10259): 1345-1352

RECOVERY Collaborative Group. Azithromycin in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. Lancet. 2021;397(10274):605-612

RECOVERY Collaborative Group. Tocilizumab in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. Lancet. 2021 May 1;397(10285):1637-1645.

RECOVERY Collaborative Group. Convalescent plasma in patients admitted to hospital with COVID-19 (RECOVERY): a randomised controlled, open-label, platform trial. Lancet. 2021 May 29;397(10289):2049-2059

• RECOVERY Collaborative Group. Colchicine in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. Lancet Respir. Med. 2021 Dec 1;9(12):1419-1426 EPub 2021 Oct 18

• RECOVERY Collaborative Group. Aspirin in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. Lancet. 2022 Jan 8;399(10320):143-151 EPub 2021 Nov 17

• RECOVERY collaborative Group. Casirivimab and imdevimab in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. Lancet 2022 Feb 12;399(10325):665-676

• RECOVERY Collaborative Group. Baricitinib in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial and updated meta-analysis. Lancet 2022 July 30;400(10349):359-368

• RECOVERY Collaborative Group. Dimethyl fumarate in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. medRxiv 2022.09.23.22280285

By the end of July 2021, the RECOVERY trial had successfully established over 175 sites across the UK and recruited over 40,000 participants treated in hospital for COVID-19. The data linkage already established to received SUS+ and other data is providing important information to the Data Monitoring Committee on a regular basis about patients' recovery (i.e. discharge from hospital), in-hospital death and procedures required. Provision of complete and reliable data to the DMC throughout the trial is critical to allow robust assessment of the effects of the trial treatments with a major contribution to these data expected from analysis of the routine health care data requested under this agreement.

On Thursday 4 June 2020, in response to a request from the UK Medicines and Healthcare Products Regulatory Agency (MHRA), the independent Data Monitoring Committee conducted a further review of the data. The DMC recommended the chief investigators review the unblinded data on the hydroxychloroquine arm of the trial. The results included a total of 1542 patients randomised to hydroxychloroquine compared with 3132 patients randomised to usual care alone. There was no significant difference in the primary endpoint of 28-day mortality (25.7% hydroxychloroquine vs. 23.5% usual care; hazard ratio 1.11 [95% confidence interval 0.98-1.26]; p=0.10). There was also no evidence of beneficial effects on hospital stay duration or other outcomes. These results were released to the public and on the 15 July 2020, the U.S. Food and Drug Administration (FDA) revoked the emergency use authorization (EUA) that allowed for chloroquine phosphate and hydroxychloroquine sulfate donated to the Strategic National Stockpile to be used to treat certain hospitalized patients with COVID-19 when a clinical trial was unavailable, or participation in a clinical trial was not feasible. This result has major implications for countries around the world who were planning to scale manufacturing of these drugs in order to treat COVID patients.

On 8 June 2020, recruitment to the dexamethasone arm was halted since, in the view of the trial Steering Committee, sufficient patients had been enrolled to establish whether or not the drug had a meaningful benefit. On 16 June 2020, preliminary results were released. A total of 2104 patients were randomised to receive dexamethasone 6 mg once per day (either by mouth or by intravenous injection) for ten days and were compared with 4321 patients randomised to usual care alone. Among the patients who received usual care alone, 28-day mortality was highest in those who required ventilation (41%), intermediate in those patients who required oxygen only (25%), and lowest among those who did not require any respiratory intervention (13%). Dexamethasone reduced deaths by one-third in ventilated patients (rate ratio 0.65 [95% confidence interval 0.48 to 0.88]; p=0.0003) and by one fifth in other patients receiving oxygen only (0.80 [0.67 to 0.96]; p=0.0021). There was no benefit among those patients who did not require respiratory support (1.22 [0.86 to 1.75]; p=0.14). Based on these results, 1 death would be prevented by treatment of around 8 ventilated patients or around 25 patients requiring oxygen alone.

On the 16 June 2020, on the basis of these results, the MHRA issued an alert to health care providers in the UK recommending the use of dexamethasone in hospitalised patients with COVID who require oxygen or ventilation. In addition dexamethasone has also been added to the government’s parallel export list, which bans companies from buying medicines meant for UK patients and selling them on for a higher price in another country.

An external analysis estimated that treatment with Dexamethasone may have saved 22,000 lives in the UK and 1 million lives worldwide up to March 2021 (most recent estimates available at: ((https://www.england.nhs.uk/2021/03/covid-treatment-developed-in-the-nhs-saves-a-million-lives/))

Without the results from RECOVERY the effectiveness of this treatment would not be known.

On 11 February 2021, analyses of the RECOVERY trial showed that, among hospitalised COVID-19 patients who had low oxygen levels and elevated markers of inflammation in the blood, the rheumatoid arthritis drug Tocilizumab significantly reduced deaths: 596 (29%) of the patients in the tocilizumab group died within 28 days compared with 694 (33%) patients in the usual care group (rate ratio 0·86; [95% confidence interval [CI] 0·77 to 0·96]; p=0·007), an absolute difference of 4%. This means that for every 25 patients treated with tocilizumab, one additional life would be saved. Tocilizumab also increased the probability of discharge alive within 28 days from 47% to 54% (rate ratio 1·23, [95% CI 1·12 to 1·34], p<0·0001). These benefits were seen in all patient subgroups, including those requiring oxygen via a simple face mask through to those requiring mechanical ventilators in an intensive care unit and those already receiving steroid treatment. Tocilizumab is now standard of care within the NHS. In December 2022, the US Food and Drug Administration approved a new indication for tocilizumab for the treatment of patients with COVID-19 based on data from trials including RECOVERY FDA Roundup: December 23, 2022 | FDA.

On 16 June 2021, analysis of 9785 patients hospitalised with COVID-19 randomly allocated to receive usual care plus the antibody combination treatment (casirivimab 4g with imdevimab 4g by intravenous infusion) or usual care alone produced practice changing results. Among the one-third of participants who were seronegative at baseline (ie they had not mounted a natural antibody response of their own) the monoclonal antibody combination significantly reduced the primary outcome of 28-day mortality by one-fifth compared with usual care alone (24% of patients in the antibody combination group died vs 30% of patients in the usual care group; rate ratio 0·80; 95% confidence interval 0·70–0·91; p=0·001). Thus, for every 100 such patients treated with the antibody combination, there would be six fewer deaths.

On 3 March 2022, analyses of 8,156 patients hospitalized with COVID-19 randomly allocated to receive usual care plus baricitinib, or usual care alone produced positive results. Treatment with baricitinib significantly reduced deaths: 513 (12%) of the patients in the baricitinib group died within 28 days compared with 546 (14%) patients in the usual care group, a reduction of 13% (age-adjusted rate ratio 0.87, 95% confidence interval [CI] 0.77 to 0.98; p= 0.026). The benefit of baricitinib was consistent regardless of which other COVID-19 treatments the patients were also receiving, including corticosteroids, tocilizumab, or remdesivir. This result provides new evidence of the additional beneficial of baricitinib on top of other immunomodulatory treatments. The MHRA have issued an alert to health care providers in the UK recommending the use of baricitinib as an additional treatment option for patients admitted to hospital with COVID-19.

RECOVERY has also shown that a number of widely used treatments are not effective in COVID-19 including aspirin, colchicine, lopinavir-ritonavir and Convalescent plasma. These findings are important since they avoid unnecessary exposure to potentially hazardous treatments and allow healthcare systems and researchers to focus resources on treatments which are effective.

The University of Oxford requires access to data for a study entitled Randomised Evaluation of COVid-19 thERapY (RECOVERY).

In 2019 a novel coronavirus-induced disease (COVID-19) emerged in Wuhan, China. A month later the Chinese Center for Disease Control and Prevention identified a new betacoronavirus (SARS (Severe Acute Respiratory Syndrome) coronavirus 2, or SARS-CoV-2) as the aetiological (causing or contributing to the development of a disease or condition) agent. The clinical manifestations of COVID-19 range from asymptomatic infection or mild, transient symptoms to severe viral pneumonia with respiratory failure. It is estimated that by December 2021, 18 million people worldwide had died from COVID-19. RECOVERY was set up in March 2020 as the UK’s national platform trial assessing potential treatments for patients hospitalised with COVID-19. Since then it has delivered over 11 practice changing results (see benefits section) with further results expected.

This study is supported by a grant to the University of Oxford from UK Research and Innovation/National Institute for Health Research (NIHR) and by core funding provided by NIHR Oxford Biomedical Research Centre, the Wellcome Trust, the Bill and Melinda Gates Foundation, Health Data Research UK (HDRUK), and the Medical Research Council Population Health Research Unit, and NIHR Clinical Trials Unit Support Funding. The new trial has been classed as an Urgent Public Health Research Study. It is one of a round of projects to receive £10.5 million as part of the £20 million rapid research response funded by UK Research and Innovation, and by the Department of Health and Social Care through the National Institute for Health Research.

This study aims to compare several different treatments that may be useful for patients with COVID-19. These treatments have been recommended by the expert panel that advises the Chief Medical Officer (CMO) in England.

The study protocol allows reliable assessment of the effects of multiple different treatments (including re-purposed and novel drugs) on major outcomes in COVID-19. All patients will receive usual care for the participating hospital. From version 6.0 of the protocol, a factorial design is being used such that eligible and consenting participants may be randomised simultaneously to one or more of the study treatment arms (depending on location and infection).

Randomisation part A: Eligible patients may be randomly allocated between the following treatment arms (although not all arms may be available at any one time):

• No additional treatment:

• Corticosteroids (children ≤44 weeks gestational age, or >44 weeks gestational age with PIMS-TS only): Children >44 weeks gestational age with SARS-CoV-2 infection without PIMS-TS will be excluded from this arm (as it is standard practice to treat with low-dose corticosteroids).

This comparison has now reported results for adults (see benefits section).

• Azithromycin: Azithromycin is a macrolide antibiotic. Azithromycin has immunomodulatory properties that has shown benefit in inflammatory lung disease.

This comparison has now reported preliminary results (see benefits section).

• Intravenous immunoglobulin (children >44 weeks gestational age with PIMSTS only): a 2 g/kg single dose has been shown to be effective in immunomodulation and preventing cardiovascular complications in Kawasaki disease, an inflammatory condition with overlapping clinical features with PIMS-TS.

• Colchicine (adults ≥18 years old only)

This comparison has now reported results (see benefits section).

• Dimethyl fumarate (UK adults ≥18 years old only; early phase assessment)

This comparison has now closed. Preliminary results have been reported.

Randomisation part B (UK only): Simultaneously, eligible patients will be randomly allocated between the following treatment arms (provided there are no contraindications and the appropriate consent has been given):

• Convalescent plasma: Plasma from patients who have recovered from SARS-CoV2 infection may contain antibodies that can bind to and neutralise the virus.

This comparison has now reported results (see benefits section).

• Synthetic neutralising antibodies: Synthetic human monoclonal antibodies have been developed to bind to and neutralise the virus.

This comparison has now reported results (see benefits section).

Randomisation part C (adults ≥18 years old only): Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• Aspirin

This comparison has now reported results (see benefits section).

Randomisation part D (adults, and children ≥2 years old with COVID-19 pneumonia [UK only]): Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• Baricitinib

This comparison has now reported results (see benefits section).

Randomisation Part E (adults ≥18 years old with hypoxia only). Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• High-dose corticosteroids

Randomisation Part F (adults ≥18 years): Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• Empagliflozin

****** NEW FOR VERSION 9********

Randomisation Part J (UK only, patients ≥12 years old): Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• Sotrovimab (a monoclonal antibody treatment against the spike protein)

Randomisation Part K (UK only, patients ≥18 years old): Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• Molnupiravir (an antiviral treatment)

Randomisation Part L (UK only, patients ≥18 years old): Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• Paxlovid (an antiviral treatment)

********************************

Second randomisation for patients with progressive COVID-19

Participants with progressive COVID-19 (as evidenced by hypoxia and an inflammatory state) may undergo an optional second randomisation between the following treatment arms:

• No additional treatment

• Tocilizumab: (an immunosuppressive drug, mainly for the treatment of rheumatoid arthritis) or control (both in addition to the treatment assigned at the first randomisation).

This comparison has now reported results (see benefits section).

Second randomisation for children with PIMS-TS

Children (at least 1 year old) with PIMS-TS (as evidenced by an exaggerated inflammatory state) may undergo an optional second randomisation between the following treatment arms:

• No additional treatment

• Tocilizumab (children ≥1 <18 years old only)

• Anakinra (children ≥1 <18 years old only)

Modifications to the number of treatment arms: Other arms can be added to the first or second randomisation if evidence emerges that there are suitable candidate therapeutics. Appendix 1 of the trial protocol sets out the rational for each individual treatment comparison currently being evaluated ((https://www.recoverytrial.net) and previous versions of the protocol do the same for treatment comparisons which have been completed.

Conversely, in some patient populations, not all trial arms are appropriate (e.g. due to contraindications based on co-morbid conditions or concomitant medication); in some hospitals or countries, not all treatment arms will be available (e.g. due to manufacturing and supply shortages); and at some times, not all treatment arms will be active (e.g. due to lack of relevant approvals and contractual agreements). The Trial Steering Committee may elect to pause one or more of the arms in order to increase trial efficiency during a fluctuating epidemic. In any of these situations, randomisation will be between fewer arms. Depending on the availability and suitability of treatments, it may be allowed for participants to be randomised in only one part (A or B) of the main randomisation. This will not impact or change the level of data that is required from NHS Digital - as the cohort in the trial remains the same, regardless of what drug is being trialled.

Data from the trial will be regularly reviewed so that any effective treatment can be identified quickly and made available to all patients. The RECOVERY Trial team will constantly review information on new drugs and include promising ones in the trial.

Patients are eligible for the study if all of the following are true:

(i) Hospitalised

(ii) SARS-CoV-2 infection (clinically suspected or laboratory confirmed)

(iii) No medical history that might, in the opinion of the attending clinician, put the patient at significant risk if he/she were to participate in the trial

In addition, if the attending clinician believes that there is a specific contra-indication to one of the active drug treatment arms or that the patient should definitely be receiving one of the active drug treatment arms then that arm will not be available for randomisation for that patient. For patients who lack capacity, an advanced directive or behaviour that clearly indicates that they would not wish to participate in the trial would be considered sufficient reason to exclude them from the trial.

Informed consent should be obtained from each patient 16 years and over before enrolment into the study. However, if the patient lacks capacity to give consent due to the severity of their medical condition (e.g. acute respiratory failure or need for immediate ventilation) or prior disease, then consent may be obtained from a relative acting as the patient’s legally designated representative or independent doctor. Further consent will then be sought with the patient if they recover sufficiently. For children aged <16 years old consent will be sought from their parents or legal guardian. Where possible, children aged between 10-15 years old will also be asked for assent. Children aged ≥16 years old will be asked for consent as for adults. Witnessed consent may be obtained over the telephone or web video link if hospital visiting rules or parental infection mean a parent/guardian cannot be physically present. Where Local Clinical Centre's (LCC's) plan to recruit children the principal investigator will co-opt support from a local paediatrician and/or neonatologists to oversee the management of children and infants in the trial. PLEASE NOTE: Only data for participants 16+ at randomisation is being requested from NHSD.

By October 2022, the RECOVERY trial had successfully recruited over 48,000 participants, making it one of the largest trials assessing treatments for COVID-19 worldwide. This rapid recruitment and resulting large sample size has enabled the study to produce important results (see yielded benefits section). This remarkable participant recruitment is the result of a number of factors. Eligibility criteria are simple and trial processes (including paperwork) are minimised. Critically, the trial (and subsequent data collection) is designed to minimise the burden on front-line hospital staff working within an overstretched care system during a major pandemic.

The large sample size allows the focus of the COVID-19 Core Protocol to be the impact of candidate treatments on mortality, the key outcome which smaller studies are not able to assess.

For each pairwise comparison with the ‘no additional treatment’ arm, the primary objective is to provide reliable estimates of the effect of study treatments on all-cause mortality at 28 days after first randomisation (with subsidiary analyses of cause of death and of death at various timepoints following discharge).

The secondary objectives are to assess the effects of study treatments on duration of hospital stay; the use of (and duration of) ventilation; and, among patients not on ventilation at baseline, the composite endpoint of death or need for mechanical ventilation or ECMO (extracorporeal membrane oxygenation). Other objectives include the assessment of the effects of study treatments on the use of any ventilation (and duration of invasive mechanical ventilation), acute kidney injury, renal replacement therapy, thrombotic events, and infections. Study outcomes will be assessed based on data recorded up to 28 days and up to 6 months after the main randomisation.

Data from routinely-collected healthcare records (including linkage to medical databases held by organisations such as NHS Digital) will allow subsidiary analyses of the effect of the study treatments on particular non-fatal events (e.g. ascertained through linkage to Hospital Episode Statistics), the influence of pre-existing major co-morbidity (e.g. diabetes, heart disease, lung disease, hepatic insufficiency, severe depression, severe kidney impairment, immunosuppression), and longer-term outcomes (e.g. six month survival) as well as in particular sub-categories of patient.

Follow-up information is to be collected on all study participants, irrespective of whether or not they complete the scheduled course of allocated study treatment. Study staff will seek follow-up information through various means including medical staff, reviewing information from medical notes, routine healthcare systems, and registries.

The study team require the following information from NHS Digital at the following frequencies (where an end date is given the frequency will be re-evaluated at that time unless otherwise stated):

**** UPDATES TO DATA FREQUENCY FOR V9 AS BELOW ****

- SUS (Secondary Use Services) Admitted Patient Care (APC) on hospital discharge. Under v8 this is provided on a monthly basis from September 2021 to September 2022. From v9 this will be provided on a Quarterly basis until March 2023 when it will cease flowing.

- HES (Hospital Episode Statistics) APC data. Under v8 this is provided on a Quarterly basis from September 2021 to September 2022. From v9 this will continue to be provided on a Quarterly basis until March 2023 when it will become Annual.

- HES Critical Care data Set (CC). Under v8 this is provided on a Quarterly basis from September 2021 to September 2022. From v9 this will continue to be provided on a Quarterly basis until March 2023 when it will become Annual.

Emergency Care Data Set (ECDS). Under v8 this is provided on a Quarterly basis from September 2021 to September 2022. From v9 this will continue to be provided on a Quarterly basis until March 2023 when it will become Annual.

- Civil Registrations (Deaths) data (death certificate information). Under v8 this is provided on a monthly basis from September 2021 to September 2022. From v9 there will be three ad-hoc data drops between Sept 2022 and March 2023 when it will become Annual.

- CHESS (COVID-19 Hospitalization in England Surveillance System) Data. Under v8 this is provided on a monthly basis from September 2021 to September 2022. This will cease flowing after September 2022.

- SGSS (COVID-19 Second Generation Surveillance System) Data. Under v8 this is provided on a monthly basis from September 2021 to September 2022. From v9 there will be three ad-hoc data drops between Sept 2022 and March 2023 when it will cease flowing.

- GDPPR Data (General Practice Extraction Service Data for Pandemic Planning and Research). Under v8 this is provided on a monthly basis from September 2021 to September 2022. From v9 there will be three ad-hoc data drops between Sept 2022 and March 2023 when it will become Annual.

- Cancer Registration Data. Under v8 this is provided on an annual basis from September 2021 to September 2022. This will continue to be provided on an Annual basis.

- Medicines dispensed in Primary Care NHS Business Services Authority data (NHS BSA). Under v8 this is provided on a monthly basis from September 2021 to September 2022. From v9 there will be three ad-hoc data drops between Sept 2022 and March 2023 when it will become Annual.

- COVID-19 Vaccination Status Data. Under v8 this is provided on a Quarterly basis from September 2021 to September 2022. From v9 this will continue to be provided on a Quarterly basis until March 2023 when it will cease flowing.

- COVID-19 UK Non-hospital Antigen Testing Results (Pillar 2). Under v8 this is provided on a Monthly basis from September 2021 to September 2022. From v9 this will continue to be provided on a Quarterly basis until March 2023 when it will cease flowing.

- Electronic Prescribing and Medicines Administration (EPMA) data in Secondary Care for COVID-19. Under v8 this is provided on a Quarterly basis from September 2021 to September 2022. From v9 this will continue to be provided on a Quarterly basis until March 2023 when it will cease flowing.

- One off Demographics linkage (Agreed and received under v7).

- *** NEW for v9*** Demographics data set. On a Quarterly basis until March 2023 when it will become Annual.

JUSTIFICATION FOR DATASETS

The GDPPR and Primary Care Medicines data sets are requested for two key reasons. First they will be used to assess whether the effects of the treatments being tested in RECOVERY are different in people with certain medical conditions (e.g. lung disease or diabetes) or who were receiving particular medications (e.g. immunosuppression) prior to joining the trial. To avoid burdening front-line staff, only limited data about other medical conditions is collected from the site staff at the point of enrolment and information about usual medications is not collected. Therefore these data sets will be invaluable in identifying relevant subgroups and will supplement information derived from the HES data regarding prior hospitalisations.

Second, information in these data sets after randomization will be used to determine the effects of the treatments on outcomes after discharge from hospital. For example, effects on the development of long term lung problems, diabetes (particularly relevant to the steroid comparison), liver disease (particularly relevant to the lopinavir/ritonavir comparison) or on infections (particularly relevant to the tocilizumab comparison). It is understood that the GDPPR data is available for a limited time period only but will allow analysis of these and other outcomes following discharge from hospital.

Identifiable data is necessary for the purpose of data linkage and quality assurance of that linkage – the Data Sharing Agreement (DSA) requires measures are taken to ensure subsequent uses of the data will involve only data that has been appropriately pseudonymised. Data requested includes a range of detailed information that is necessary for analysis and sub-group analysis of the effectiveness of the treatments – in particular to enable accounting for a range of biases and confounding factors relevant to the primary and secondary outcomes, specifically:

• The trial has taken a pragmatic approach to recruitment and as a result may have included some patients who do not have confirmed Covid 19. For sub-group analysis and interpretation of the trial results, they need to know which patients have a confirmed positive Covid 19 test.

• Disease onset, dates for admission into hospital, ICU (Intensive Care Unit) and HDU (High Dependency Units) are important to track progression of the disease to understand the efficacy of treatment.

• These data will allow the trial to assess the impact of trial drugs on the precise care and treatment delivered e.g. the need for mechanical ventilation and any complications such as secondary infections.

• Risk factors and underlying conditions enable sub-group analysis of other factors that may have a bearing on the treatments being tested.

The Cancer Registrations dataset will be important in identifying long-term outcomes in the cohort and will help to reliably assess the long-term safety of the treatments tested in RECOVERY. This is particularly important as treatment arms now include medicines which have not previously been assessed in large trials with long follow-up (e.g. tocilizumab) or are currently unlicensed (e.g. REGN-COV2). Cancer Registration data will be provided on an annual basis for the cohort.

The BSA Medicines dataset will be used to further characterise the cohort at baseline, allowing analysis of the effects of the treatments in different types of patient. In addition, refreshes of these data are requested as part of the long-term follow-up as they help to identify new diagnoses (e.g. diabetes). This will help the team identify long-term outcomes in the cohort and assess the effects of the treatments on these outcomes. The processing of this data and use of the BSA Medicines dataset for RECOVERY is in line with the Direction covering that dataset where the purpose is “…to drive the linkage of medicines data with other data sets to provide intelligence about the safety and effectives of medicines”.

The HES Critical Care Dataset is requested to allow comparison with the more rapidly available SUS data. The SUS Critical Care Dataset is used to identify use of organ support following randomisation including use of invasive mechanical ventilation (the secondary outcome for the study). SUS is timely, allowing rapid generation of practice changing data but its quality is not assured to the same extend as HES. Checking for consistencies between SUS Critical Care Dataset and HES Critical Care Dataset would enable further analyses from SUS to be used with confidence and provide reassurance to regulators (e.g. the MHRA) on the validity of the trial results.

The COVID-19 vaccination data will be particularly relevant to the convalescent plasma and Regeneron monoclonal antibody results since these treatments may be less effective in people with prior vaccination. Obtaining linked vaccination data will help understand which patients are likely to benefit from the Regeneron monoclonal antibody.

Linkage to the Covid-19 UK Non-hospital Antigen Testing Results (pillar 2) testing data will enable the study to reliably distinguish those with laboratory proven COVID-19 from those with suspected COVID-19. By the time patients present to hospital PCR testing may be negative and therefore the study may currently underestimate the proportion of participants who have proven SARS-CoV-2 infection. This is important for the interpretation of the study results and acceptance of the findings by regulators.

The Emergency Care Dataset is requested principally to identify safety outcomes occurring in the study participants after discharge from hospital. Most relevant would be infections presenting to A&E. A number of immunosuppressant treatments have been and continue to be tested in RECOVERY and understanding the effects of these treatments on infections following the COVID-19 admission will help doctors to make informed treatment choices for their patients with COVID-19.

The Electronic Prescribing and Medicines Administration (EPMA) data in Secondary Care for COVID-19 dataset is requested for two purposes. First treatments given in hospital will be used to verify or identify complications occurring in the study participants (e.g. infections requiring intravenous antibiotics). Second, these data will be used to assess adherence to the treatments allocated in the study, which is important in the interpretation of the study results.

A one off Personal Demographics Service matching service is requested. Among the 40,000 participants recruited to date around 400 were recruited in England but have a missing or invalid NHS number recorded by the site staff. In order to allow data to be collected on these consented participants going forward, the RECOVERY team are requesting a one off patient matching service. The study team will provide date of birth, sex and name.

*********NEW FOR VERSION 9***********

- Additional dataset requested (Amendment v9)

Quarterly Demographics data is being requested for the purpose of identifying individuals during follow-up who no longer have a registered GP. This is because if a study participant no longer has a GP then no further follow-up data can be obtained, so the date of removal of the study participant is needed to calculate the follow-up time for each participant. The variable fields Reason_for_removal, and Date_of_removal are required for this.

RECOVERY PPIE activities:

When RECOVERY was initially being set up in March 2020, there was public involvement via the mechanism of Oxford Population Health’s existing public advisory panel (Nuffield Department of Public Health (NDPH) public advisory panel). They undertook the following tasks:

• Reviewed patient information sheets – 6 NDPH public advisory panel members reviewed

• Reviewed patient consent form – as above

• Reviewed the storyboard of the RECOVERY patient-facing animation – 4 members of NDPH public advisory panel

• Reviewed the first newsletters to trial participants/parents and carers of child participants – 11 people reviewed and the newsletters were revised considerably as a result, splitting out the most scientific detail into a separate section.

During 2021 the trial set up a ‘bespoke’ public advisory group, comprised of patient volunteers who had been participants in RECOVERY. This group has met, in general, every couple of months. The group consists of 10 females and six males, who range in age from late 40s to early 80s. Prior to RECOVERY, none had previous experience of patient and public involvement.

This advisory group has been consulted and reviewed further newsletters to participants, an animation on RECOVERY-related data sharing and has also been involved in the development of a proposal to adapt the RECOVERY platform to trial treatment for flu.

The group reviewed the adapted patient information sheet and consent form that will be used for RECOVERY flu. This resulted in these being simplified as far as was possible. (NB: The study team don’t yet know if RECOVERY flu is going to be funded)

There has also been public engagement around RECOVERY. This includes:

• Media outreach including liaison with participants for case studies/media interviews, 16,100 items of coverage.

• Social media activity (>19,000 uses of #RECOVERYtrial in 2020)

• Development of case studies, e.g. for Understanding Patient Data

• Public-facing webinar with NIHR (participant, husband and clinician in discussion)

• Swindon Science Festival materials (Martin Landray panel event, and online ‘Anatomy of clinical trials’ resource for schools)

• NIHR Oxford Biomedical Research Centre (BRC) Health Open Day – RECOVERY Trial featured in engagement activities with school age children at this small event in Oxford.

NDPH will continue to engage with the RECOVERY PPI Panel and the public throughout the course of the trial.

***************************

LEGAL BASIS FOR THE PROCESSING OF DATA

The University of Oxford, as the Data Controller who is also processing the data will process Personal Data under GDPR Article 6 (1) (e) - Processing is necessary for the performance of a task carried out in the public interest or in the exercise of official authority vested in the controller. As a higher education establishment, the University of Oxford conducts research to improve health care and services, and the data requested is necessary for the performance of a task carried out in the public interest. The University of Oxford hopes that the RECOVERY methodology will to improve the speed, scale and quality of trials in COVID-19 and other diseases, providing reliable information on potential treatments for COVID-19 and potentially changing the standard of care that the NHS offers which could improve outcomes for many thousands of patients in the UK and around the world. The RECOVERY extended analysis of immunomodulatory therapies sub-study is anticipated further to lead of the causal biological pathways through which immunomodulatory therapies benefit patients with COVID-19; aiding research into future therapeutics for COVID-19 and therefore contribute to improvements in risk assessment and treatment of NHS patients in the UK.

Additionally, under GDPR Article 9(2)(j) processing of Special Category Personal Data (of which Health data is one) is necessary for archiving for research purposes. Data minimisation processes are being followed and only data that is specifically required for the purposes of this study have been requested, to protect the rights of the data subjects. The data are required for research purposes in the public interest – meeting the conditions in the DPA 2018 Schedule 1 Part 1 (4) – which GDPR Recital 52(2) determines is an appropriate derogation from the prohibition on processing special categories of personal data.

In accordance with GDPR Article 89(1) processing is subject to appropriate safeguards. These include:

- The data recipient’s technical and organisational measures to safeguard the data have been assessed and meet NHS Digital’s acceptance criteria (see sections 2 and 5b of this application for further details;

- The requested data have been assessed as proportionate to the aim pursued (see section 5a of this application for further details);

Controls, data retention and processing activities have been assessed to ensure respect to the essence of the right to data protection (see sections 5a, 5b and 8a of this application for further details);

- Measures to protect the rights and freedoms of data subjects have been assessed including transparency (fair processing) publishing subject’s rights to withdraw consent and /or have their data erased or rectified, etc.

Onward sharing of data (ADaM and SDTM) for manufacturers / regulators:

The University of Oxford has designed two types of data outputs which the RECOVERY trial needs to onwardly share:

• Analysis Data Model (ADaM) output

• Study Data Tabulation Model (SDTM) output

These trial analysis datasets, are developed using data provided by NHS Digital under this Data Sharing Agreement along with data from other sources, and are subject to a range of processing in line with standards published by the Clinical Data Interchange Standards Consortium (or CDISC).

The ADaM and SDTM outputs for the RECOVERY trial have undergone technical assessment and IG assessment by NHS Digital and NHS Digital has determined that, subject to being shared under strict conditions (specified below), the datasets are sufficiently low risk and can be onward shared by the University of Oxford with the manufacturer or regulator where necessary.

ADaM output:

Based on a desk review and further information from the RECOVERY team at the University of Oxford in response to specific questions, NHS Digital has concluded that none of the fields examined in the ADaM output can be used on their own or with other information in the onwardly transferred file or in the public domain to reconstitute fully fields in the data provided to the University of Oxford under this Agreement.

SDTM output:

Based on a desk review, responses to specific questions from the University of Oxford and a general self-certification from the University of Oxford, NHS Digital has concluded that none of the fields examined in the SDTM output can be used on their own or with other information in the onwardly transferred file or in the public domain to reconstitute fully fields in the data provided to the University of Oxford. A highly motivated and technically literate intruder could attempt to do this for a subset of the information in a specific field but this would not be the information in the field in its entirety and accuracy in doing this could not be proven or guaranteed.

Based on the above assessments, the SDTM and ADaM datasets can be onwardly shared under the following conditions:

1. The dataset may be shared with the manufacturer of the treatment(s) evaluated in the RECOVERY trial for the purpose of the manufacturer using the data in ongoing research in the development programme for the treatment (for example, evaluating additional beneficial or harmful effects of the treatment or evaluating the effects of the treatment in particular patient populations). Any such uses will comply with the following controls.

2. The dataset may be shared by either the University of Oxford or the manufacturer with a competent authority (regulator) in circumstances where the RECOVERY trial has a legal obligation to disclose data with the regulator to allow trial results to be evaluated to demonstrate the safety and effectiveness of the treatment(s) tested. Regulators include the Medicines & Healthcare products Regulatory Agency (MHRA) in the UK, the Food and Drug Administration (F

COVID-19 is an emerging pathogen which presents a significant threat to the population in terms of increased morbidity and mortality, particularly among vulnerable groups such as those with pre-existing disease.

The primary objective is to provide reliable estimates of the effect of study treatments on in-hospital death (with subsidiary analyses of cause of death and death at various timepoints following discharge).

The secondary objectives are to assess the effects of study treatments on duration of hospital stay; the need for (and duration of) ventilation; and the need for renal replacement therapy.

The interim trial results will be monitored by an independent Data Monitoring Committee (DMC). The most important task for the DMC will be to assess whether the randomised comparisons in the study have provided evidence on mortality that is strong enough (with a range of uncertainty around the results that is narrow enough) to affect national and global treatment strategies. In such a circumstance, the DMC will inform the Trial Steering Committee who will make the results available to the public and amend the trial arms accordingly.

The data requested will be used to evaluate the efficacy and safety of the study treatments and will help shape the public health response. The rapid feed will be used to ensure that the trial Data Monitoring Committee have complete, up-to-date information on the major outcomes in the trial on which to base their decisions. If they find compelling evidence of efficacy or safety then that arm may be stopped and – if effective – added to standard care across the NHS.

All outputs produced will be in the form of aggregated reports with small number suppression applied.

Outputs delivered to date include the following results publications

RECOVERY Collaborative Group, Horby P, Lim WS, Emberson JR, Mafham M, Bell JL, Linsell L, et al. Dexamethasone in Hospitalized Patients with Covid-19. N Engl J Med 2021; 384:693-704

RECOVERY Collaborative Group. Effect of Hydroxychloroquine in Hospitalized Patients with Covid-19. N Engl J Med. 2020;383(21):2030-2040

RECOVERY Collaborative Group. Lopinavir–ritonavir in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. The Lancet, 2020;396(10259): 1345-1352

RECOVERY Collaborative Group. Azithromycin in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. Lancet. 2021;397(10274):605-612

RECOVERY Collaborative Group. Tocilizumab in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. Lancet. 2021 May 1;397(10285):1637-1645.

RECOVERY Collaborative Group. Convalescent plasma in patients admitted to hospital with COVID-19 (RECOVERY): a randomised controlled, open-label, platform trial. Lancet. 2021 May 29;397(10289):2049-2059

• RECOVERY Collaborative Group. Colchicine in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. Lancet Respir. Med. 2021 Dec 1;9(12):1419-1426 EPub 2021 Oct 18

• RECOVERY Collaborative Group. Aspirin in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. Lancet. 2022 Jan 8;399(10320):143-151 EPub 2021 Nov 17

• RECOVERY collaborative Group. Casirivimab and imdevimab in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. Lancet 2022 Feb 12;399(10325):665-676

• RECOVERY Collaborative Group. Baricitinib in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial and updated meta-analysis. Lancet 2022 July 30;400(10349):359-368

• RECOVERY Collaborative Group. Dimethyl fumarate in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. medRxiv 2022.09.23.22280285

By the end of July 2021, the RECOVERY trial had successfully established over 175 sites across the UK and recruited over 40,000 participants treated in hospital for COVID-19. The data linkage already established to received SUS+ and other data is providing important information to the Data Monitoring Committee on a regular basis about patients' recovery (i.e. discharge from hospital), in-hospital death and procedures required. Provision of complete and reliable data to the DMC throughout the trial is critical to allow robust assessment of the effects of the trial treatments with a major contribution to these data expected from analysis of the routine health care data requested under this agreement.

On Thursday 4 June 2020, in response to a request from the UK Medicines and Healthcare Products Regulatory Agency (MHRA), the independent Data Monitoring Committee conducted a further review of the data. The DMC recommended the chief investigators review the unblinded data on the hydroxychloroquine arm of the trial. The results included a total of 1542 patients randomised to hydroxychloroquine compared with 3132 patients randomised to usual care alone. There was no significant difference in the primary endpoint of 28-day mortality (25.7% hydroxychloroquine vs. 23.5% usual care; hazard ratio 1.11 [95% confidence interval 0.98-1.26]; p=0.10). There was also no evidence of beneficial effects on hospital stay duration or other outcomes. These results were released to the public and on the 15 July 2020, the U.S. Food and Drug Administration (FDA) revoked the emergency use authorization (EUA) that allowed for chloroquine phosphate and hydroxychloroquine sulfate donated to the Strategic National Stockpile to be used to treat certain hospitalized patients with COVID-19 when a clinical trial was unavailable, or participation in a clinical trial was not feasible. This result has major implications for countries around the world who were planning to scale manufacturing of these drugs in order to treat COVID patients.

On 8 June 2020, recruitment to the dexamethasone arm was halted since, in the view of the trial Steering Committee, sufficient patients had been enrolled to establish whether or not the drug had a meaningful benefit. On 16 June 2020, preliminary results were released. A total of 2104 patients were randomised to receive dexamethasone 6 mg once per day (either by mouth or by intravenous injection) for ten days and were compared with 4321 patients randomised to usual care alone. Among the patients who received usual care alone, 28-day mortality was highest in those who required ventilation (41%), intermediate in those patients who required oxygen only (25%), and lowest among those who did not require any respiratory intervention (13%). Dexamethasone reduced deaths by one-third in ventilated patients (rate ratio 0.65 [95% confidence interval 0.48 to 0.88]; p=0.0003) and by one fifth in other patients receiving oxygen only (0.80 [0.67 to 0.96]; p=0.0021). There was no benefit among those patients who did not require respiratory support (1.22 [0.86 to 1.75]; p=0.14). Based on these results, 1 death would be prevented by treatment of around 8 ventilated patients or around 25 patients requiring oxygen alone.

On the 16 June 2020, on the basis of these results, the MHRA issued an alert to health care providers in the UK recommending the use of dexamethasone in hospitalised patients with COVID who require oxygen or ventilation. In addition dexamethasone has also been added to the government’s parallel export list, which bans companies from buying medicines meant for UK patients and selling them on for a higher price in another country.

A recent analysis estimates that treatment with Dexamethasone may save between 4,000 and 27,000 lives in the UK by January 20201 (depending on the number of COVID-19 cases during Q3-4 2020) and potentially over half a million lives worldwide (https://pubmed.ncbi.nlm.nih.gov/32678530/). Without the results from RECOVERY the effectiveness of this treatment would not be known.

On 11 February 2021, analyses of the RECOVERY trial showed that, among hospitalised COVID-19 patients who had low oxygen levels and elevated markers of inflammation in the blood, the rheumatoid arthritis drug Tocilizumab significantly reduced deaths: 596 (29%) of the patients in the tocilizumab group died within 28 days compared with 694 (33%) patients in the usual care group (rate ratio 0·86; [95% confidence interval [CI] 0·77 to 0·96]; p=0·007), an absolute difference of 4%. This means that for every 25 patients treated with tocilizumab, one additional life would be saved. Tocilizumab also increased the probability of discharge alive within 28 days from 47% to 54% (rate ratio 1·23, [95% CI 1·12 to 1·34], p<0·0001). These benefits were seen in all patient subgroups, including those requiring oxygen via a simple face mask through to those requiring mechanical ventilators in an intensive care unit and those already receiving steroid treatment. Tocilizumab is now standard of care within the NHS.

On 16 June 2021, analysis of 9785 patients hospitalised with COVID-19 randomly allocated to receive usual care plus the antibody combination treatment (casirivimab 4g with imdevimab 4g by intravenous infusion) or usual care alone produced practice changing results. Among the one-third of participants who were seronegative at baseline (ie they had not mounted a natural antibody response of their own) the monoclonal antibody combination significantly reduced the primary outcome of 28-day mortality by one-fifth compared with usual care alone (24% of patients in the antibody combination group died vs 30% of patients in the usual care group; rate ratio 0·80; 95% confidence interval 0·70–0·91; p=0·001). Thus, for every 100 such patients treated with the antibody combination, there would be six fewer deaths.

**************NEW FOR VERSION 9***********

On 3 March 2022, analyses of 8,156 patients hospitalized with COVID-19 randomly allocated to receive usual care plus baricitinib, or usual care alone produced positive results. Treatment with baricitinib significantly reduced deaths: 513 (12%) of the patients in the baricitinib group died within 28 days compared with 546 (14%) patients in the usual care group, a reduction of 13% (age-adjusted rate ratio 0.87, 95% confidence interval [CI] 0.77 to 0.98; p= 0.026). The benefit of baricitinib was consistent regardless of which other COVID-19 treatments the patients were also receiving, including corticosteroids, tocilizumab, or remdesivir. This result provides new evidence of the additional beneficial of baricitinib on top of other immunomodulatory treatments.

RECOVERY has also shown that a number of widely used treatments are not effective in COVID-19 including aspirin, colchicine, lopinavir-ritonavir and Convalescent plasma. These findings are important since they avoid unnecessary exposure to potentially hazardous treatments and allow healthcare systems and researchers to focus resources on treatments which are effective.

The University of Oxford requires access to data for a study entitled Randomised Evaluation of COVid-19 thERapY (RECOVERY).

In 2019 a novel coronavirus-induced disease (COVID-19) emerged in Wuhan, China. A month later the Chinese Center for Disease Control and Prevention identified a new betacoronavirus (SARS (Severe Acute Respiratory Syndrome) coronavirus 2, or SARS-CoV-2) as the aetiological (causing or contributing to the development of a disease or condition) agent. The clinical manifestations of COVID-19 range from asymptomatic infection or mild, transient symptoms to severe viral pneumonia with respiratory failure. As many patients do not progress to severe disease the overall case fatality rate per infected individual is low, but hospitals in areas with significant community transmission experienced a major increase in the number of hospitalized pneumonia patients, and the frequency of severe disease in hospitalised patients was recorded as high as 30%. The progression from prodrome (an early symptom indicating the onset of a disease or illness - in this case usually fever, fatigue and cough) to severe pneumonia requiring oxygen support or mechanical ventilation often takes one to two weeks after the onset of symptoms. The kinetics of viral replication in the respiratory tract are not well characterized, but this relatively slow progression provides a potential time window in which antiviral therapies could influence the course of disease.

This study aims to compare several different treatments that may be useful for patients with COVID-19. These treatments have been recommended by the expert panel that advises the Chief Medical Officer (CMO) in England.

The protocol allows reliable assessment of the effects of multiple different treatments (including re-purposed and novel drugs) on major outcomes in COVID-19. All patients will receive usual care for the participating hospital. From version 6.0 of the protocol, a factorial design was used such that eligible and consenting participants may be randomised to one of the treatment arms in Randomisation A and, simultaneously, to one of the treatment arms in other Randomisations.

Randomisation part A: Eligible patients may be randomly allocated between the following treatment arms (although not all arms may be available at any one time):

• No additional treatment:

• Corticosteroids (children ≤44 weeks gestational age, or >44 weeks gestational age with PIMS-TS only): Children >44 weeks gestational age with SARS-CoV-2 infection without PIMS-TS will be excluded from this arm (as it is standard practice to treat with low-dose corticosteroids).

This comparison has now reported preliminary results for adults (see benefits section).

• Azithromycin: Azithromycin is a macrolide antibiotic. Azithromycin has immunomodulatory properties that has shown benefit in inflammatory lung disease.

This comparison has now reported preliminary results (see benefits section).

• Intravenous immunoglobulin (children >44 weeks gestational age with PIMSTS only): a 2 g/kg single dose has been shown to be effective in immunomodulation and preventing cardiovascular complications in Kawasaki disease, an inflammatory condition with overlapping clinical features with PIMS-TS.

• Colchicine (adults ≥18 years old only)

This comparison has now reported preliminary results (see benefits section).

• Dimethyl fumarate (UK adults ≥18 years old only; early phase assessment)

Randomisation part B (UK only): Simultaneously, eligible patients will be randomly allocated between the following treatment arms (provided there are no contraindications and the appropriate consent has been given):

• Convalescent plasma: Plasma from patients who have recovered from SARS-CoV2 infection may contain antibodies that can bind to and neutralise the virus.

This comparison has now reported preliminary results (see benefits section).

• Synthetic neutralising antibodies: Synthetic human monoclonal antibodies have been developed to bind to and neutralise the virus.

This comparison has now reported preliminary results (see benefits section).

Randomisation part C (adults ≥18 years old only): Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• Aspirin

This comparison has now reported preliminary results (see benefits section).

Randomisation part D (adults, and children ≥2 years old with COVID-19 pneumonia [UK only]): Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• Baricitinib

Randomisation Part E (non-UK countries: adults ≥18 years old with hypoxia only). Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• High-dose corticosteroids

Note that no NHS Digital data is used in this randomisation or any other international arms of the RECOVERY Trial.

Randomisation Part F (adults ≥18 years): Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• Empagliflozin

Second randomisation for patients with progressive COVID-19

Participants with progressive COVID-19 (as evidenced by hypoxia and an inflammatory state) may undergo an optional second randomisation between the following treatment arms:

• No additional treatment

• Tocilizumab: (an immunosuppressive drug, mainly for the treatment of rheumatoid arthritis) or control (both in addition to the treatment assigned at the first randomisation).

This comparison has now reported preliminary results (see benefits section).

Second randomisation for children with PIMS-TS

Children (at least 1 year old) with PIMS-TS (as evidenced by an exaggerated inflammatory state) may undergo an optional second randomisation between the following treatment arms:

• No additional treatment

• Tocilizumab (children ≥1 <18 years old only)

• Anakinra (children ≥1 <18 years old only)

Modifications to the number of treatment arms: Other arms can be added to the first or second randomisation if evidence emerges that there are suitable candidate therapeutics. Appendix 1 of the trial protocol sets out the rational for each individual treatment comparison currently being evaluated (https://www.recoverytrial.net/files/recovery-protocol-v17-1-2021-08-10-1.pdf) and previous versions of the protocol do the same for treatment comparisons which have been completed.

Conversely, in some patient populations, not all trial arms are appropriate (e.g. due to contraindications based on co-morbid conditions or concomitant medication); in some hospitals or countries, not all treatment arms will be available (e.g. due to manufacturing and supply shortages); and at some times, not all treatment arms will be active (e.g. due to lack of relevant approvals and contractual agreements). The Trial Steering Committee may elect to pause one or more of the arms in order to increase trial efficiency during a fluctuating epidemic. In any of these situations, randomisation will be between fewer arms. Depending on the availability and suitability of treatments, it may be allowed for participants to be randomised in only one part (A or B) of the main randomisation. This will not impact or change the level of data that is required from NHS Digital - as the cohort in the trial remains the same, regardless of what drug is being trialled.

Data from the trial will be regularly reviewed so that any effective treatment can be identified quickly and made available to all patients. The RECOVERY Trial team will constantly review information on new drugs and include promising ones in the trial.

Patients are eligible for the study if all of the following are true:

(i) Hospitalised

(ii) SARS-CoV-2 infection (clinically suspected or laboratory confirmed)

(iii) No medical history that might, in the opinion of the attending clinician, put the patient at significant risk if he/she were to participate in the trial

In addition, if the attending clinician believes that there is a specific contra-indication to one of the active drug treatment arms or that the patient should definitely be receiving one of the active drug treatment arms then that arm will not be available for randomisation for that patient. For patients who lack capacity, an advanced directive or behaviour that clearly indicates that they would not wish to participate in the trial would be considered sufficient reason to exclude them from the trial.

Informed consent should be obtained from each patient 16 years and over before enrolment into the study. However, if the patient lacks capacity to give consent due to the severity of their medical condition (e.g. acute respiratory failure or need for immediate ventilation) or prior disease, then consent may be obtained from a relative acting as the patient’s legally designated representative or independent doctor. Further consent will then be sought with the patient if they recover sufficiently. For children aged <16 years old consent will be sought from their parents or legal guardian. Where possible, children aged between 10-15 years old will also be asked for assent. Children aged ≥16 years old will be asked for consent as for adults. Witnessed consent may be obtained over the telephone or web video link if hospital visiting rules or parental infection mean a parent/guardian cannot be physically present. Where Local Clinical Centre's (LCC's) plan to recruit children the principal investigator will co-opt support from a local paediatrician and/or neonatologists to oversee the management of children and infants in the trial.

By July 2021, the RECOVERY trial had successfully recruited over 40,000 participants, making it one of the largest trials assessing treatments for COVID-19 worldwide. Already this rapid recruitment and resulting large sample size has enabled the study to produce important results (see yielded benefits section). This remarkable participant recruitment is the result of a number of factors. Eligibility criteria are simple and trial processes (including paperwork) are minimised. Critically, the trial (and subsequent data collection) is designed to minimise the burden on front-line hospital staff working within an overstretched care system during a major pandemic. Given the high mortality rate among patients hospitalised with acute COVID-19, even treatments with only a moderate impact on survival or on hospital resources could be worthwhile.

The large sample size allows the focus of the COVID-19 Core Protocol to be the impact of candidate treatments on mortality, the key outcome which smaller studies are not able to assess.

For each pairwise comparison with the ‘no additional treatment’ arm, the primary objective is to provide reliable estimates of the effect of study treatments on all-cause mortality at 28 days after first randomisation (with subsidiary analyses of cause of death and of death at various timepoints following discharge).

The secondary objectives are to assess the effects of study treatments on duration of hospital stay; the need for (and duration of) ventilation; and, among patients not on ventilation at baseline, the composite endpoint of death or need for mechanical ventilation or ECMO (extracorporeal membrane oxygenation). Other objectives include the assessment of the effects of study treatments on the need for renal replacement therapy and new major cardiac arrhythmias. Study outcomes will be assessed based on data recorded up to 28 days and up to 6 months after the main randomization.

Data from routinely-collected healthcare records (including linkage to medical databases held by organisations such as NHS Digital) and from relevant research studies (such as UK Biobank and Genomics England (possibly in the future)) will allow subsidiary analyses of the effect of the study treatments on particular non-fatal events (e.g. ascertained through linkage to Hospital Episode Statistics), the influence of pre-existing major co-morbidity (e.g. diabetes, heart disease, lung disease, hepatic insufficiency, severe depression, severe kidney impairment, immunosuppression), and longer-term outcomes (e.g. six month survival) as well as in particular sub-categories of patient (e.g. by genotype). For clarification - this agreement does not permit the linkage of NHS Digital data to data held by UK Biobank or Genomics England, and a subsequent amendment to this agreement will be put in place should this occur.

Follow-up information is to be collected on all study participants, irrespective of whether or not they complete the scheduled course of allocated study treatment. Study staff will seek follow-up information through various means including medical staff, reviewing information from medical notes, routine healthcare systems, and registries.

The study team require the following information from NHS Digital at the following frequencies (where an end date is given the frequency will be re-evaluated at that time unless otherwise stated):

- Prompt feed from SUS (Secondary Use Services) on hospital discharge. This will be provided on a monthly basis from September 2021 to September 2022.

- Quarterly feed on HES (Hospital Episode Statistics) data (until September 2022, then switch to annually).

- Monthly feeds of civil registration data (death certificate information) (until September 2022).

- CHESS (COVID-19 Hospitalization in England Surveillance System) Data. To be provided monthly from September 2021 to September 2022.

- SGSS (COVID-19 Second Generation Surveillance System) Data. To be provided monthly from September 2021 to September 2022.

- GDPPR Data (General Practice Extraction Service Data for Pandemic Planning and Research) on a monthly basis until September 2022.

- Cancer Registration Data (annually from September 2021)

- Prescribing Data (monthly until September 2022)

- HES CCDS (to be provided quarterly until September 2022 and then annually).

- COVID-19 vaccination (quarterly)

- Covid-19 UK Non-hospital Antigen Testing Results (pillar 2) (monthly until September 2022)

- Emergency Care Data Set (ECDS) (quarterly)

- Electronic Prescribing and Medicines Administration (EPMA) data in Secondary Care for COVID-19 (quarterly)

- One off Demographics linkage

The GDPPR and Primary Care Medicines data sets are requested for two key reasons. First they will be used to assess whether the effects of the treatments being tested in RECOVERY are different in people with certain medical conditions (e.g. lung disease or diabetes) or who were receiving particular medications (e.g. immunosuppression) prior to joining the trial. To avoid burdening front-line staff, only limited data about other medical conditions is collected from the site staff at the point of enrolment and information about usual medications is not collected. Therefore these data sets will be invaluable in identifying relevant subgroups and will supplement information derived from the HES data regarding prior hospitalisations.

Second, information in these data sets after randomization will be used to determine the effects of the treatments on outcomes after discharge from hospital. For example, effects on the development of long term lung problems, diabetes (particularly relevant to the steroid comparison), liver disease (particularly relevant to the lopinavir/ritonavir comparison) or on infections (particularly relevant to the tocilizumab comparison). It is understood that the GDPPR data is available for a limited time period only but will allow analysis of these and other outcomes following discharge from hospital.

Identifiable data is necessary for the purpose of data linkage and quality assurance of that linkage – the Data Sharing Agreement (DSA) requires measures are taken to ensure subsequent uses of the data will involve only data that has been appropriately pseudonymised. Data requested includes a range of detailed information that is necessary for analysis and sub-group analysis of the effectiveness of the treatments – in particular to enable accounting for a range of biases and confounding factors relevant to the primary and secondary outcomes, specifically:

• The trial has taken a pragmatic approach to recruitment and as a result may have included some patients who do not have confirmed Covid 19. For sub-group analysis and interpretation of the trial results, they need to know which patients have a confirmed positive Covid 19 test.

• Disease onset, dates for admission into hospital, ICU (Intensive Care Unit) and HDU (High Dependency Units) are important to track progression of the disease to understand the efficacy of treatment.

• These data will allow the trial to assess the impact of trial drugs on the precise care and treatment delivered e.g. the need for mechanical ventilation and any complications such as secondary infections.

• Risk factors and underlying conditions enable sub-group analysis of other factors that may have a bearing on the treatments being tested.

This study is supported by a grant to the University of Oxford from UK Research and Innovation/National Institute for Health Research (NIHR) and by core funding provided by NIHR Oxford Biomedical Research Centre, the Wellcome Trust, the Bill and Melinda Gates Foundation, Health Data Research UK (HDRUK), and the Medical Research Council Population Health Research Unit, and NIHR Clinical Trials Unit Support Funding. The new trial has been classed as an Urgent Public Health Research Study. It is one of a round of projects to receive £10.5 million as part of the £20 million rapid research response funded by UK Research and Innovation, and by the Department of Health and Social Care through the National Institute for Health Research.

- Additional datasets requested, October 2020:

The Cancer Registrations dataset will be important in identifying long-term outcomes in the cohort and will help to reliably assess the long-term safety of the treatments tested in RECOVERY. This is particularly important as treatment arms now include medicines which have not previously been assessed in large trials with long follow-up (e.g. tocilizumab) or are currently unlicensed (e.g. REGN-COV2). Cancer Registration data will be provided on an annual basis for the cohort.

The BSA Medicines dataset will be used to further characterise the cohort at baseline, allowing analysis of the effects of the treatments in different types of patient. In addition, refreshes of these data are requested as part of the long-term follow-up as they help to identify new diagnoses (e.g. diabetes). This will help the team identify long-term outcomes in the cohort and assess the effects of the treatments on these outcomes. The processing of this data and use of the BSA Medicines dataset for RECOVERY is in line with the Direction covering that dataset where the purpose is “…to drive the linkage of medicines data with other data sets to provide intelligence about the safety and effectives of medicines”.

- Additional datasets requested, August 2021:

The HES Critical Care Dataset is requested to allow comparison with the more rapidly available SUS data. The SUS Critical Care Dataset is used to identify use of organ support following randomisation including use of invasive mechanical ventilation (the secondary outcome for the study). SUS is timely, allowing rapid generation of practice changing data but its quality is not assured to the same extend as HES. Checking for consistencies between SUS Critical Care Dataset and HES Critical Care Dataset would enable further analyses from SUS to be used with confidence and provide reassurance to regulators (e.g. the MHRA) on the validity of the trial results.

The COVID-19 vaccination data will be particularly relevant to the convalescent plasma and Regeneron monoclonal antibody results since these treatments may be less effective in people with prior vaccination. Obtaining linked vaccination data will help understand which patients are likely to benefit from the Regeneron monoclonal antibody.

Linkage to the Covid-19 UK Non-hospital Antigen Testing Results (pillar 2) testing data will enable the study to reliably distinguish those with laboratory proven COVID-19 from those with suspected COVID-19. By the time patients present to hospital PCR testing may be negative and therefore the study may currently underestimate the proportion of participants who have proven SARS-CoV-2 infection. This is important for the interpretation of the study results and acceptance of the findings by regulators.

The Emergency Care Dataset is requested principally to identify safety outcomes occurring in the study participants after discharge from hospital. Most relevant would be infections presenting to A&E. A number of immunosuppressant treatments have been and continue to be tested in RECOVERY and understanding the effects of these treatments on infections following the COVID-19 admission will help doctors to make informed treatment choices for their patients with COVID-19.

The Electronic Prescribing and Medicines Administration (EPMA) data in Secondary Care for COVID-19 dataset is requested for two purposes. First treatments given in hospital will be used to verify or identify complications occurring in the study participants (e.g. infections requiring intravenous antibiotics). Second, these data will be used to assess adherence to the treatments allocated in the study, which is important in the interpretation of the study results.

A one off Personal Demographics Service matching service is requested. Among the 40,000 participants recruited to date around 400 were recruited in England but have a missing or invalid NHS number recorded by the site staff. In order to allow data to be collected on these consented participants going forward, the REOCVERY team are requesting a one off patient matching service. The study team will provide date of birth, sex and name.

Onward sharing of data (ADaM and SDTM) for manufacturers / regulators:

The University of Oxford has designed two types of data outputs which the RECOVERY trial needs to onwardly share:

• Analysis Data Model (ADaM) output

• Study Data Tabulation Model (SDTM) output

These trial analysis datasets, are developed using data provided by NHS Digital under this Data Sharing Agreement along with data from other sources, and are subject to a range of processing in line with standards published by the Clinical Data Interchange Standards Consortium (or CDISC).

The ADaM and SDTM outputs for the RECOVERY trial have undergone technical assessment and IG assessment by NHS Digital and NHS Digital has determined that, subject to being shared under strict conditions (specified below), the datasets are sufficiently low risk and can be onward shared by the University of Oxford with the manufacturer or regulator where necessary.

ADaM output:

Based on a desk review and further information from the RECOVERY team at the University of Oxford in response to specific questions, NHS Digital has concluded that none of the fields examined in the ADaM output can be used on their own or with other information in the onwardly transferred file or in the public domain to reconstitute fully fields in the data provided to the University of Oxford under this Agreement.

SDTM output:

Based on a desk review, responses to specific questions from the University of Oxford and a general self-certification from the University of Oxford, NHS Digital has concluded that none of the fields examined in the SDTM output can be used on their own or with other information in the onwardly transferred file or in the public domain to reconstitute fully fields in the data provided to the University of Oxford. A highly motivated and technically literate intruder could attempt to do this for a subset of the information in a specific field but this would not be the information in the field in its entirety and accuracy in doing this could not be proven or guaranteed.

Based on the above assessments, the SDTM and ADaM datasets can be onwardly shared under the following conditions:

1. The dataset may be shared with the manufacturer of the treatment(s) evaluated in the RECOVERY trial for the purpose of the manufacturer using the data in ongoing research in the development programme for the treatment (for example, evaluating additional beneficial or harmful effects of the treatment or evaluating the effects of the treatment in particular patient populations). Any such uses will comply with the following controls.

2. The dataset may be shared by either the University of Oxford or the manufacturer with a competent authority (regulator) in circumstances where the RECOVERY trial has a legal obligation to disclose data with the regulator to allow trial results to be evaluated to demonstrate the safety and effectiveness of the treatment(s) tested. Regulators include the Medicines & Healthcare products Regulatory Agency (MHRA) in the UK, the Food and Drug Administration (FDA) in the United States, and any competent authority in the European Medicines Agency (EMA).

3. NHS Digital permits the Infectious Diseases Data Observatory (IDDO) to onwards share ADaM & SDTM datasets subject to the special conditions outlined in the Sub-licensing section.

4. The dataset shared must conform to the SDTM and ADaM outputs agreed with NHS Digital and may not contain additional variables.

5. The dataset must not be shared alongside any source data provided under this Agreement.

6. The dataset may only be shared under the conditions that the recipient(s):

i. must not combine it with other datasets which could potentially increase the risk of reidentification for individuals in the dataset;

ii. must not attempt to re-identify individuals in the dataset;

iii. must not onwardly share the dataset other than permitted sharing between the manufacturer and the regulatory authority as required;

iv. must not use the dataset for any purpose other than those defined above, and

v. must not publish the data sourced under this Agreement.

7. University of Oxford has contractual controls in place with manufacturers, which will prevent re-identification of trial participants and limit the purposes for use of the data.

The datasets to be onwardly shared contain no dates nor the order of randomisation.

The SDTM and ADaM datasets has its own set of unique identifiers. It is reversible in the event of a query from a regulatory authority but only by the team in Oxford.

NHS Digital source data files are not disclosed.

NHS Digital require the right to be informed of data shared for the above purpose.

It is currently under review as to whether NHS Digital will waive the right to audit the regulator or manufacturer of treatment(s) evaluated in the RECOVERY trial.

Sublicense:

This agreement includes a sublicence, which permits the following:

• The University of Oxford to onward share the RECOVERY trial SDTM (Study Data Tabulation Model) and ADaM (Analysis Data Model) datasets with the Infectious Diseases Data Observatory (IDDO) COVID-19 platform;

• IDDO to use their Data Access Committee to enable researchers to submit a request to access analysis outputs from the RECOVERY trial.

The SDTM and ADaM datasets include data provided by NHS Digital under this Data Sharing Agreement, and other data from other sources. Data has been manipulated, including combination with other datasets not supplied under this agreement, to create the analysis datasets for onward sharing, and relates to a specific cohort of consented participants. It is therefore accepted that the data shared under a sub-license agreement could not be supplied by NHS Digital directly. A technical and Information Governance (IG) assessment of the analysis datasets outputs (SDTM / AdAM) concluded that the ADaM and SDTM outputs for the RECOVERY trial, subject to being shared under strict conditions, are sufficiently low risk.

What is the IDDO?

The IDDO is a data repository based at the University of Oxford, and has developed a COVID-19 clinical data platform with a governance and ethical framework to ensure the equitable sharing of data. The RECOVERY trial will use this platform for onward sharing of data. The University of Oxford is the legal entity for the platform.

What data, and how will data be shared with IDDO?

NHS Digital shares Data with the University of Oxford for participants recruited to the RECOVERY trial. The Data, and other trial data collected as part of the RECOVERY trial which NHS Digital does not hold, is managed and analysed using international Clinical Data Interchange Standards Consortium (CDISC) Standards, to generate datasets for the purposes of analysis, Data Sharing and Regulatory Submission.

Data is transformed into the SDTM dataset, which provides a standard way of comprehensively representing the collected data, and is then transformed into ADaM datasets for analysis. Throughout the SDTM and ADaM data, participants are identified by their unique subject identifier. This is internal to the RECOVERY trial and independent of any external identifiers, such as NHS number, which may have been used to obtain linkage data. Neither these external identifiers nor other direct identifiers such as names and addresses are included in datasets for analysis.

Under the sub-license model, a RECOVERY data analyst will create the SDTM and AdAM outputs containing all the variables for each of the drugs evaluated in RECOVERY. These datasets will be approved by the information asset owner and will be submitted to the platform via the secure portal. No raw data will be submitted.

The RECOVERY trial, as a contributor of data to the platform, will agree a Terms of Submission with the IDDO prior to submission, which outlines the rights of the contributor, their obligations, the terms of processing and how the data will be used.

How researchers can request to access data?

Potential applicants are able to explore the availability of data via a Data Inventory, accessible via the IDDO website.

Data Access Application form:

The Data Requestor completes a Data Access Application Form. This contains details of the following: researcher; organisation, and team outlining role in analysis; research plan, including objectives, analysis, ethical considerations, publication, funding and any relevant scientific review, and data variables. The completed application form is submitted to the IDDO.

Review Considerations:

Upon receipt of a request, the Data Access Requests will be sent to the RECOVERY trial Principal Investigators to confirm that the request is in line with the expectations of the trial participants (based on the consent materials) and will maintain the scientific value and validity of the trial.

Data Access Request, and feedback from the RECOVERY Principal Investigators, will be considered by the IDDO Data Access Committee. This will take into account the following:

a) The researcher is working in a field relevant to COVID-19, and has a formal affiliation to a health, research, humanitarian, government, inter-government or academic institution with legal status;

b) The researcher is suitably qualified, either an academic record consistent with execution of the proposed analysis, or the support of a supervising co-applicant with appropriate expertise;

c) Availability of sufficient funding;

d) No prior breach of Data Transfer Agreements;

e) No concerns with respect to the ethical aspects of the application. This includes necessary legal, scientific and/or ethics approvals from their institution.;

f) No conflict of interest;

g) Impact in terms of collaboration and knowledge sharing

Approvals and contracts:

All required scientific and ethical reviews and/or approvals must be obtained before any Data are disclosed.

Prior to release a Data Transfer Agreement (DTA) will put in place between the University of Oxford and the recipient. This agreement shall limit purpose, preventing re-identification, and permit audit and destruction of data upon expiry or termination of the agreement. Additionally, if the recipient is in breach of any obligations, it may lead to termination of agreement. To enable harmonisation across all IDDO agreements, the term of the DTA is one year, in line with the NHS Digital standard for Length of Data Sharing Agreement. The DTA constitutes the Sublicense Agreement.

Data transfer:

The likelihood of re-identification of RECOVERY trial participants, accounting for all means reasonably likely, is mitigated by the use of CDISC standards, removal of external and direct identifiers, and use of a legally binding contract to prevent re-identi

COVID-19 is an emerging pathogen which presents a significant threat to the population in terms of increased morbidity and mortality, particularly among vulnerable groups such as those with pre-existing disease.

The primary objective is to provide reliable estimates of the effect of study treatments on in-hospital death (with subsidiary analyses of cause of death and death at various timepoints following discharge).

The secondary objectives are to assess the effects of study treatments on duration of hospital stay; the need for (and duration of) ventilation; and the need for renal replacement therapy.

The interim trial results will be monitored by an independent Data Monitoring Committee (DMC). The most important task for the DMC will be to assess whether the randomised comparisons in the study have provided evidence on mortality that is strong enough (with a range of uncertainty around the results that is narrow enough) to affect national and global treatment strategies. In such a circumstance, the DMC will inform the Trial Steering Committee who will make the results available to the public and amend the trial arms accordingly.

The data requested will be used to evaluate the efficacy and safety of the study treatments and will help shape the public health response. The rapid feed will be used to ensure that the trial Data Monitoring Committee have complete, up-to-date information on the major outcomes in the trial on which to base their decisions. If they find compelling evidence of efficacy or safety then that arm may be stopped and – if effective – added to standard care across the NHS.

All outputs produced will be in the form of aggregated reports with small number suppression applied.

Outputs delivered to date include the following results publications

RECOVERY Collaborative Group, Horby P, Lim WS, Emberson JR, Mafham M, Bell JL, Linsell L, et al. Dexamethasone in Hospitalized Patients with Covid-19. N Engl J Med 2021; 384:693-704

RECOVERY Collaborative Group. Effect of Hydroxychloroquine in Hospitalized Patients with Covid-19. N Engl J Med. 2020;383(21):2030-2040

RECOVERY Collaborative Group. Lopinavir–ritonavir in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. The Lancet, 2020;396(10259): 1345-1352

RECOVERY Collaborative Group. Azithromycin in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. Lancet. 2021;397(10274):605-612

RECOVERY Collaborative Group. Tocilizumab in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. Lancet. 2021 May 1;397(10285):1637-1645.

RECOVERY Collaborative Group. Convalescent plasma in patients admitted to hospital with COVID-19 (RECOVERY): a randomised controlled, open-label, platform trial. Lancet. 2021 May 29;397(10289):2049-2059

RECOVERY Collaborative Group. Aspirin in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. medRxiv 2021.06.08.21258132

RECOVERY Collaborative Group. Colchicine in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. medRxiv 2021.05.18.21257267

RECOVERY collaborative Group. Casirivimab and imdevimab in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. medRxiv 2021.06.15.21258542

By the end of July 2021, the RECOVERY trial had successfully established over 175 sites across the UK and recruited over 40,000 participants treated in hospital for COVID-19. The data linkage already established to received SUS+ and other data is providing important information to the Data Monitoring Committee on a regular basis about patients' recovery (i.e. discharge from hospital), in-hospital death and procedures required. Provision of complete and reliable data to the DMC throughout the trial is critical to allow robust assessment of the effects of the trial treatments with a major contribution to these data expected from analysis of the routine health care data requested under this agreement.

On Thursday 4 June 2020, in response to a request from the UK Medicines and Healthcare Products Regulatory Agency (MHRA), the independent Data Monitoring Committee conducted a further review of the data. The DMC recommended the chief investigators review the unblinded data on the hydroxychloroquine arm of the trial. The results included a total of 1542 patients randomised to hydroxychloroquine compared with 3132 patients randomised to usual care alone. There was no significant difference in the primary endpoint of 28-day mortality (25.7% hydroxychloroquine vs. 23.5% usual care; hazard ratio 1.11 [95% confidence interval 0.98-1.26]; p=0.10). There was also no evidence of beneficial effects on hospital stay duration or other outcomes. These results were released to the public and on the 15 July 2020, the U.S. Food and Drug Administration (FDA) revoked the emergency use authorization (EUA) that allowed for chloroquine phosphate and hydroxychloroquine sulfate donated to the Strategic National Stockpile to be used to treat certain hospitalized patients with COVID-19 when a clinical trial was unavailable, or participation in a clinical trial was not feasible. This result has major implications for countries around the world who were planning to scale manufacturing of these drugs in order to treat COVID patients.

On 8 June 2020, recruitment to the dexamethasone arm was halted since, in the view of the trial Steering Committee, sufficient patients had been enrolled to establish whether or not the drug had a meaningful benefit. On 16 June 2020, preliminary results were released. A total of 2104 patients were randomised to receive dexamethasone 6 mg once per day (either by mouth or by intravenous injection) for ten days and were compared with 4321 patients randomised to usual care alone. Among the patients who received usual care alone, 28-day mortality was highest in those who required ventilation (41%), intermediate in those patients who required oxygen only (25%), and lowest among those who did not require any respiratory intervention (13%). Dexamethasone reduced deaths by one-third in ventilated patients (rate ratio 0.65 [95% confidence interval 0.48 to 0.88]; p=0.0003) and by one fifth in other patients receiving oxygen only (0.80 [0.67 to 0.96]; p=0.0021). There was no benefit among those patients who did not require respiratory support (1.22 [0.86 to 1.75]; p=0.14). Based on these results, 1 death would be prevented by treatment of around 8 ventilated patients or around 25 patients requiring oxygen alone.

On the 16 June 2020, on the basis of these results, the MHRA issued an alert to health care providers in the UK recommending the use of dexamethasone in hospitalised patients with COVID who require oxygen or ventilation. In addition dexamethasone has also been added to the government’s parallel export list, which bans companies from buying medicines meant for UK patients and selling them on for a higher price in another country.

A recent analysis estimates that treatment with Dexamethasone may save between 4,000 and 27,000 lives in the UK by January 20201 (depending on the number of COVID-19 cases during Q3-4 2020) and potentially over half a million lives worldwide (https://www.medrxiv.org/content/10.1101/2020.07.29.20164269v1). Without the results from RECOVERY the effectiveness of this treatment would not be known.

On 11 February 2021, analyses of the RECOVERY trial showed that, among hospitalised COVID-19 patients who had low oxygen levels and elevated markers of inflammation in the blood, the rheumatoid arthritis drug Tocilizumab significantly reduced deaths: 596 (29%) of the patients in the tocilizumab group died within 28 days compared with 694 (33%) patients in the usual care group (rate ratio 0·86; [95% confidence interval [CI] 0·77 to 0·96]; p=0·007), an absolute difference of 4%. This means that for every 25 patients treated with tocilizumab, one additional life would be saved. Tocilizumab also increased the probability of discharge alive within 28 days from 47% to 54% (rate ratio 1·23, [95% CI 1·12 to 1·34], p<0·0001). These benefits were seen in all patient subgroups, including those requiring oxygen via a simple face mask through to those requiring mechanical ventilators in an intensive care unit and those already receiving steroid treatment. Tocilizumab is now standard of care within the NHS.

On 16 June 2021, analysis of 9785 patients hospitalised with COVID-19 randomly allocated to receive usual care plus the antibody combination treatment (casirivimab 4g with imdevimab 4g by intravenous infusion) or usual care alone produced practice changing results. Among the one-third of participants who were seronegative at baseline (ie they had not mounted a natural antibody response of their own) the monoclonal antibody combination significantly reduced the primary outcome of 28-day mortality by one-fifth compared with usual care alone (24% of patients in the antibody combination group died vs 30% of patients in the usual care group; rate ratio 0·80; 95% confidence interval 0·70–0·91; p=0·001). Thus, for every 100 such patients treated with the antibody combination, there would be six fewer deaths.

RECOVERY has also shown that a number of widely used treatments are not effective in COVID-19 including aspirin, colchicine, lopinavir-ritonavir and Convalescent plasma. These findings are important since they avoid unnecessary exposure to potentially hazardous treatments and allow healthcare systems and researchers to focus resources on treatments which are effective.

The University of Oxford requires access to data for a study entitled Randomised Evaluation of COVid-19 thERapY (RECOVERY).

In 2019 a novel coronavirus-induced disease (COVID-19) emerged in Wuhan, China. A month later the Chinese Center for Disease Control and Prevention identified a new betacoronavirus (SARS (Severe Acute Respiratory Syndrome) coronavirus 2, or SARS-CoV-2) as the aetiological (causing or contributing to the development of a disease or condition) agent. The clinical manifestations of COVID-19 range from asymptomatic infection or mild, transient symptoms to severe viral pneumonia with respiratory failure. As many patients do not progress to severe disease the overall case fatality rate per infected individual is low, but hospitals in areas with significant community transmission experienced a major increase in the number of hospitalized pneumonia patients, and the frequency of severe disease in hospitalised patients was recorded as high as 30%. The progression from prodrome (an early symptom indicating the onset of a disease or illness - in this case usually fever, fatigue and cough) to severe pneumonia requiring oxygen support or mechanical ventilation often takes one to two weeks after the onset of symptoms. The kinetics of viral replication in the respiratory tract are not well characterized, but this relatively slow progression provides a potential time window in which antiviral therapies could influence the course of disease.

This study aims to compare several different treatments that may be useful for patients with COVID-19. These treatments have been recommended by the expert panel that advises the Chief Medical Officer (CMO) in England.

The protocol allows reliable assessment of the effects of multiple different treatments (including re-purposed and novel drugs) on major outcomes in COVID-19. All patients will receive usual care for the participating hospital. From version 6.0 of the protocol, a factorial design was used such that eligible and consenting participants may be randomised to one of the treatment arms in Randomisation A and, simultaneously, to one of the treatment arms in other Randomisations.

Randomisation part A: Eligible patients may be randomly allocated between the following treatment arms (although not all arms may be available at any one time):

• No additional treatment:

• Corticosteroids (children ≤44 weeks gestational age, or >44 weeks gestational age with PIMS-TS only): Children >44 weeks gestational age with SARS-CoV-2 infection without PIMS-TS will be excluded from this arm (as it is standard practice to treat with low-dose corticosteroids).

This comparison has now reported preliminary results for adults (see benefits section).

• Azithromycin: Azithromycin is a macrolide antibiotic. Azithromycin has immunomodulatory properties that has shown benefit in inflammatory lung disease.

This comparison has now reported preliminary results (see benefits section).

• Intravenous immunoglobulin (children >44 weeks gestational age with PIMSTS only): a 2 g/kg single dose has been shown to be effective in immunomodulation and preventing cardiovascular complications in Kawasaki disease, an inflammatory condition with overlapping clinical features with PIMS-TS.

• Colchicine (adults ≥18 years old only)

This comparison has now reported preliminary results (see benefits section).

• Dimethyl fumarate (UK adults ≥18 years old only; early phase assessment)

Randomisation part B (UK only): Simultaneously, eligible patients will be randomly allocated between the following treatment arms (provided there are no contraindications and the appropriate consent has been given):

• Convalescent plasma: Plasma from patients who have recovered from SARS-CoV2 infection may contain antibodies that can bind to and neutralise the virus.

This comparison has now reported preliminary results (see benefits section).

• Synthetic neutralising antibodies: Synthetic human monoclonal antibodies have been developed to bind to and neutralise the virus.

This comparison has now reported preliminary results (see benefits section).

Randomisation part C (adults ≥18 years old only): Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• Aspirin

This comparison has now reported preliminary results (see benefits section).

Randomisation part D (adults, and children ≥2 years old with COVID-19 pneumonia [UK only]): Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• Baricitinib

Randomisation Part E (non-UK countries: adults ≥18 years old with hypoxia only). Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• High-dose corticosteroids

Note that no NHS Digital data is used in this randomisation or any other international arms of the RECOVERY Trial.

Randomisation Part F (adults ≥18 years): Simultaneously, eligible patients will be randomly allocated between the following treatment arms:

• No additional treatment

• Empagliflozin

Second randomisation for patients with progressive COVID-19

Participants with progressive COVID-19 (as evidenced by hypoxia and an inflammatory state) may undergo an optional second randomisation between the following treatment arms:

• No additional treatment

• Tocilizumab: (an immunosuppressive drug, mainly for the treatment of rheumatoid arthritis) or control (both in addition to the treatment assigned at the first randomisation).

This comparison has now reported preliminary results (see benefits section).

Second randomisation for children with PIMS-TS

Children (at least 1 year old) with PIMS-TS (as evidenced by an exaggerated inflammatory state) may undergo an optional second randomisation between the following treatment arms:

• No additional treatment

• Tocilizumab (children ≥1 <18 years old only)

• Anakinra (children ≥1 <18 years old only)

Modifications to the number of treatment arms: Other arms can be added to the first or second randomisation if evidence emerges that there are suitable candidate therapeutics. Appendix 1 of the trial protocol sets out the rational for each individual treatment comparison currently being evaluated (https://www.recoverytrial.net/files/recovery-protocol-v17-1-2021-08-10-1.pdf) and previous versions of the protocol do the same for treatment comparisons which have been completed.

Conversely, in some patient populations, not all trial arms are appropriate (e.g. due to contraindications based on co-morbid conditions or concomitant medication); in some hospitals or countries, not all treatment arms will be available (e.g. due to manufacturing and supply shortages); and at some times, not all treatment arms will be active (e.g. due to lack of relevant approvals and contractual agreements). The Trial Steering Committee may elect to pause one or more of the arms in order to increase trial efficiency during a fluctuating epidemic. In any of these situations, randomisation will be between fewer arms. Depending on the availability and suitability of treatments, it may be allowed for participants to be randomised in only one part (A or B) of the main randomisation. This will not impact or change the level of data that is required from NHS Digital - as the cohort in the trial remains the same, regardless of what drug is being trialled.

Data from the trial will be regularly reviewed so that any effective treatment can be identified quickly and made available to all patients. The RECOVERY Trial team will constantly review information on new drugs and include promising ones in the trial.

Patients are eligible for the study if all of the following are true:

(i) Hospitalised

(ii) SARS-CoV-2 infection (clinically suspected or laboratory confirmed)

(iii) No medical history that might, in the opinion of the attending clinician, put the patient at significant risk if he/she were to participate in the trial

In addition, if the attending clinician believes that there is a specific contra-indication to one of the active drug treatment arms or that the patient should definitely be receiving one of the active drug treatment arms then that arm will not be available for randomisation for that patient. For patients who lack capacity, an advanced directive or behaviour that clearly indicates that they would not wish to participate in the trial would be considered sufficient reason to exclude them from the trial.

Informed consent should be obtained from each patient 16 years and over before enrolment into the study. However, if the patient lacks capacity to give consent due to the severity of their medical condition (e.g. acute respiratory failure or need for immediate ventilation) or prior disease, then consent may be obtained from a relative acting as the patient’s legally designated representative or independent doctor. Further consent will then be sought with the patient if they recover sufficiently. For children aged <16 years old consent will be sought from their parents or legal guardian. Where possible, children aged between 10-15 years old will also be asked for assent. Children aged ≥16 years old will be asked for consent as for adults. Witnessed consent may be obtained over the telephone or web video link if hospital visiting rules or parental infection mean a parent/guardian cannot be physically present. Where Local Clinical Centre's (LCC's) plan to recruit children the principal investigator will co-opt support from a local paediatrician and/or neonatologists to oversee the management of children and infants in the trial.

By July 2021, the RECOVERY trial had successfully recruited over 40,000 participants, making it one of the largest trials assessing treatments for COVID-19 worldwide. Already this rapid recruitment and resulting large sample size has enabled the study to produce important results (see yielded benefits section). This remarkable participant recruitment is the result of a number of factors. Eligibility criteria are simple and trial processes (including paperwork) are minimised. Critically, the trial (and subsequent data collection) is designed to minimise the burden on front-line hospital staff working within an overstretched care system during a major pandemic. Given the high mortality rate among patients hospitalised with acute COVID-19, even treatments with only a moderate impact on survival or on hospital resources could be worthwhile.

The large sample size allows the focus of the COVID-19 Core Protocol to be the impact of candidate treatments on mortality, the key outcome which smaller studies are not able to assess.

For each pairwise comparison with the ‘no additional treatment’ arm, the primary objective is to provide reliable estimates of the effect of study treatments on all-cause mortality at 28 days after first randomisation (with subsidiary analyses of cause of death and of death at various timepoints following discharge).

The secondary objectives are to assess the effects of study treatments on duration of hospital stay; the need for (and duration of) ventilation; and, among patients not on ventilation at baseline, the composite endpoint of death or need for mechanical ventilation or ECMO (extracorporeal membrane oxygenation). Other objectives include the assessment of the effects of study treatments on the need for renal replacement therapy and new major cardiac arrhythmias. Study outcomes will be assessed based on data recorded up to 28 days and up to 6 months after the main randomization.

Data from routinely-collected healthcare records (including linkage to medical databases held by organisations such as NHS Digital) and from relevant research studies (such as UK Biobank and Genomics England (possibly in the future)) will allow subsidiary analyses of the effect of the study treatments on particular non-fatal events (e.g. ascertained through linkage to Hospital Episode Statistics), the influence of pre-existing major co-morbidity (e.g. diabetes, heart disease, lung disease, hepatic insufficiency, severe depression, severe kidney impairment, immunosuppression), and longer-term outcomes (e.g. six month survival) as well as in particular sub-categories of patient (e.g. by genotype). For clarification - this agreement does not permit the linkage of NHS Digital data to data held by UK Biobank or Genomics England, and a subsequent amendment to this agreement will be put in place should this occur.

Follow-up information is to be collected on all study participants, irrespective of whether or not they complete the scheduled course of allocated study treatment. Study staff will seek follow-up information through various means including medical staff, reviewing information from medical notes, routine healthcare systems, and registries.

The study team require the following information from NHS Digital at the following frequencies (where an end date is given the frequency will be re-evaluated at that time unless otherwise stated):

- Prompt feed from SUS (Secondary Use Services) on hospital discharge. This will be provided on a monthly basis from September 2021 to September 2022.

- Quarterly feed on HES (Hospital Episode Statistics) data (until September 2022, then switch to annually).

- Monthly feeds of civil registration data (death certificate information) (until September 2022).

- CHESS (COVID-19 Hospitalization in England Surveillance System) Data. To be provided monthly from September 2021 to September 2022.

- SGSS (COVID-19 Second Generation Surveillance System) Data. To be provided monthly from September 2021 to September 2022.

- GDPPR Data (General Practice Extraction Service Data for Pandemic Planning and Research) on a monthly basis until September 2022.

- Cancer Registration Data (annually from September 2021)

- Prescribing Data (monthly until September 2022)

- HES CCDS (to be provided quarterly until September 2022 and then annually).

- COVID-19 vaccination (quarterly)

- Covid-19 UK Non-hospital Antigen Testing Results (pillar 2) (monthly until September 2022)

- Emergency Care Data Set (ECDS) (quarterly)

- Electronic Prescribing and Medicines Administration (EPMA) data in Secondary Care for COVID-19 (quarterly)

- One off Demographics linkage

The GDPPR and Primary Care Medicines data sets are requested for two key reasons. First they will be used to assess whether the effects of the treatments being tested in RECOVERY are different in people with certain medical conditions (e.g. lung disease or diabetes) or who were receiving particular medications (e.g. immunosuppression) prior to joining the trial. To avoid burdening front-line staff, only limited data about other medical conditions is collected from the site staff at the point of enrolment and information about usual medications is not collected. Therefore these data sets will be invaluable in identifying relevant subgroups and will supplement information derived from the HES data regarding prior hospitalisations.

Second, information in these data sets after randomization will be used to determine the effects of the treatments on outcomes after discharge from hospital. For example, effects on the development of long term lung problems, diabetes (particularly relevant to the steroid comparison), liver disease (particularly relevant to the lopinavir/ritonavir comparison) or on infections (particularly relevant to the tocilizumab comparison). It is understood that the GDPPR data is available for a limited time period only but will allow analysis of these and other outcomes following discharge from hospital.

Identifiable data is necessary for the purpose of data linkage and quality assurance of that linkage – the Data Sharing Agreement (DSA) requires measures are taken to ensure subsequent uses of the data will involve only data that has been appropriately pseudonymised. Data requested includes a range of detailed information that is necessary for analysis and sub-group analysis of the effectiveness of the treatments – in particular to enable accounting for a range of biases and confounding factors relevant to the primary and secondary outcomes, specifically:

• The trial has taken a pragmatic approach to recruitment and as a result may have included some patients who do not have confirmed Covid 19. For sub-group analysis and interpretation of the trial results, they need to know which patients have a confirmed positive Covid 19 test.

• Disease onset, dates for admission into hospital, ICU (Intensive Care Unit) and HDU (High Dependency Units) are important to track progression of the disease to understand the efficacy of treatment.

• These data will allow the trial to assess the impact of trial drugs on the precise care and treatment delivered e.g. the need for mechanical ventilation and any complications such as secondary infections.

• Risk factors and underlying conditions enable sub-group analysis of other factors that may have a bearing on the treatments being tested.

This study is supported by a grant to the University of Oxford from UK Research and Innovation/National Institute for Health Research (NIHR) and by core funding provided by NIHR Oxford Biomedical Research Centre, the Wellcome Trust, the Bill and Melinda Gates Foundation, Health Data Research UK (HDRUK), and the Medical Research Council Population Health Research Unit, and NIHR Clinical Trials Unit Support Funding. The new trial has been classed as an Urgent Public Health Research Study. It is one of a round of projects to receive £10.5 million as part of the £20 million rapid research response funded by UK Research and Innovation, and by the Department of Health and Social Care through the National Institute for Health Research.

- Additional datasets requested, October 2020:

The Cancer Registrations dataset will be important in identifying long-term outcomes in the cohort and will help to reliably assess the long-term safety of the treatments tested in RECOVERY. This is particularly important as treatment arms now include medicines which have not previously been assessed in large trials with long follow-up (e.g. tocilizumab) or are currently unlicensed (e.g. REGN-COV2). Cancer Registration data will be provided on an annual basis for the cohort.

The BSA Medicines dataset will be used to further characterise the cohort at baseline, allowing analysis of the effects of the treatments in different types of patient. In addition, refreshes of these data are requested as part of the long-term follow-up as they help to identify new diagnoses (e.g. diabetes). This will help the team identify long-term outcomes in the cohort and assess the effects of the treatments on these outcomes. The processing of this data and use of the BSA Medicines dataset for RECOVERY is in line with the Direction covering that dataset where the purpose is “…to drive the linkage of medicines data with other data sets to provide intelligence about the safety and effectives of medicines”.

- Additional datasets requested, August 2021:

The HES Critical Care Dataset is requested to allow comparison with the more rapidly available SUS data. The SUS Critical Care Dataset is used to identify use of organ support following randomisation including use of invasive mechanical ventilation (the secondary outcome for the study). SUS is timely, allowing rapid generation of practice changing data but its quality is not assured to the same extend as HES. Checking for consistencies between SUS Critical Care Dataset and HES Critical Care Dataset would enable further analyses from SUS to be used with confidence and provide reassurance to regulators (e.g. the MHRA) on the validity of the trial results.

The COVID-19 vaccination data will be particularly relevant to the convalescent plasma and Regeneron monoclonal antibody results since these treatments may be less effective in people with prior vaccination. Obtaining linked vaccination data will help understand which patients are likely to benefit from the Regeneron monoclonal antibody.

Linkage to the Covid-19 UK Non-hospital Antigen Testing Results (pillar 2) testing data will enable the study to reliably distinguish those with laboratory proven COVID-19 from those with suspected COVID-19. By the time patients present to hospital PCR testing may be negative and therefore the study may currently underestimate the proportion of participants who have proven SARS-CoV-2 infection. This is important for the interpretation of the study results and acceptance of the findings by regulators.

The Emergency Care Dataset is requested principally to identify safety outcomes occurring in the study participants after discharge from hospital. Most relevant would be infections presenting to A&E. A number of immunosuppressant treatments have been and continue to be tested in RECOVERY and understanding the effects of these treatments on infections following the COVID-19 admission will help doctors to make informed treatment choices for their patients with COVID-19.

The Electronic Prescribing and Medicines Administration (EPMA) data in Secondary Care for COVID-19 dataset is requested for two purposes. First treatments given in hospital will be used to verify or identify complications occurring in the study participants (e.g. infections requiring intravenous antibiotics). Second, these data will be used to assess adherence to the treatments allocated in the study, which is important in the interpretation of the study results.

A one off Personal Demographics Service matching service is requested. Among the 40,000 participants recruited to date around 400 were recruited in England but have a missing or invalid NHS number recorded by the site staff. In order to allow data to be collected on these consented participants going forward, the REOCVERY team are requesting a one off patient matching service. The study team will provide date of birth, sex and name.

Onward sharing of data (ADaM and SDTM) for manufacturers / regulators:

The University of Oxford has designed two types of data outputs which the RECOVERY trial needs to onwardly share:

• Analysis Data Model (ADaM) output

• Study Data Tabulation Model (SDTM) output

These trial analysis datasets, are developed using data provided by NHS Digital under this Data Sharing Agreement along with data from other sources, and are subject to a range of processing in line with standards published by the Clinical Data Interchange Standards Consortium (or CDISC).

The ADaM and SDTM outputs for the RECOVERY trial have undergone technical assessment and IG assessment by NHS Digital and NHS Digital has determined that, subject to being shared under strict conditions (specified below), the datasets are sufficiently low risk and can be onward shared by the University of Oxford with the manufacturer or regulator where necessary.

ADaM output:

Based on a desk review and further information from the RECOVERY team at the University of Oxford in response to specific questions, NHS Digital has concluded that none of the fields examined in the ADaM output can be used on their own or with other information in the onwardly transferred file or in the public domain to reconstitute fully fields in the data provided to the University of Oxford under this Agreement.

SDTM output:

Based on a desk review, responses to specific questions from the University of Oxford and a general self-certification from the University of Oxford, NHS Digital has concluded that none of the fields examined in the SDTM output can be used on their own or with other information in the onwardly transferred file or in the public domain to reconstitute fully fields in the data provided to the University of Oxford. A highly motivated and technically literate intruder could attempt to do this for a subset of the information in a specific field but this would not be the information in the field in its entirety and accuracy in doing this could not be proven or guaranteed.

Based on the above assessments, the SDTM and ADaM datasets can be onwardly shared under the following conditions:

1. The dataset may be shared with the manufacturer of the treatment(s) evaluated in the RECOVERY trial for the purpose of the manufacturer using the data in ongoing research in the development programme for the treatment (for example, evaluating additional beneficial or harmful effects of the treatment or evaluating the effects of the treatment in particular patient populations). Any such uses will comply with the following controls.

2. The dataset may be shared by either the University of Oxford or the manufacturer with a competent authority (regulator) in circumstances where the RECOVERY trial has a legal obligation to disclose data with the regulator to allow trial results to be evaluated to demonstrate the safety and effectiveness of the treatment(s) tested. Regulators include the Medicines & Healthcare products Regulatory Agency (MHRA) in the UK, the Food and Drug Administration (FDA) in the United States, and any competent authority in the European Medicines Agency (EMA).

3. NHS Digital permits the Infectious Diseases Data Observatory (IDDO) to onwards share ADaM & SDTM datasets subject to the special conditions outlined in the Sub-licensing section.

4. The dataset shared must conform to the SDTM and ADaM outputs agreed with NHS Digital and may not contain additional variables.

5. The dataset must not be shared alongside any source data provided under this Agreement.

6. The dataset may only be shared under the conditions that the recipient(s):

i. must not combine it with other datasets which could potentially increase the risk of reidentification for individuals in the dataset;

ii. must not attempt to re-identify individuals in the dataset;

iii. must not onwardly share the dataset other than permitted sharing between the manufacturer and the regulatory authority as required;

iv. must not use the dataset for any purpose other than those defined above, and

v. must not publish the data sourced under this Agreement.

7. University of Oxford has contractual controls in place with manufacturers, which will prevent re-identification of trial participants and limit the purposes for use of the data.

The datasets to be onwardly shared contain no dates nor the order of randomisation.

The SDTM and ADaM datasets has its own set of unique identifiers. It is reversible in the event of a query from a regulatory authority but only by the team in Oxford.

NHS Digital source data files are not disclosed.

NHS Digital require the right to be informed of data shared for the above purpose.

It is currently under review as to whether NHS Digital will waive the right to audit the regulator or manufacturer of treatment(s) evaluated in the RECOVERY trial.

Sublicense:

This agreement includes a sublicence, which permits the following:

• The University of Oxford to onward share the RECOVERY trial SDTM (Study Data Tabulation Model) and ADaM (Analysis Data Model) datasets with the Infectious Diseases Data Observatory (IDDO) COVID-19 platform;

• IDDO to use their Data Access Committee to enable researchers to submit a request to access analysis outputs from the RECOVERY trial.

The SDTM and ADaM datasets include data provided by NHS Digital under this Data Sharing Agreement, and other data from other sources. Data has been manipulated, including combination with other datasets not supplied under this agreement, to create the analysis datasets for onward sharing, and relates to a specific cohort of consented participants. It is therefore accepted that the data shared under a sub-license agreement could not be supplied by NHS Digital directly. A technical and Information Governance (IG) assessment of the analysis datasets outputs (SDTM / AdAM) concluded that the ADaM and SDTM outputs for the RECOVERY trial, subject to being shared under strict conditions, are sufficiently low risk.

What is the IDDO?

The IDDO is a data repository based at the University of Oxford, and has developed a COVID-19 clinical data platform with a governance and ethical framework to ensure the equitable sharing of data. The RECOVERY trial will use this platform for onward sharing of data. The University of Oxford is the legal entity for the platform.

What data, and how will data be shared with IDDO?

NHS Digital shares Data with the University of Oxford for participants recruited to the RECOVERY trial. The Data, and other trial data collected as part of the RECOVERY trial which NHS Digital does not hold, is managed and analysed using international Clinical Data Interchange Standards Consortium (CDISC) Standards, to generate datasets for the purposes of analysis, Data Sharing and Regulatory Submission.

Data is transformed into the SDTM dataset, which provides a standard way of comprehensively representing the collected data, and is then transformed into ADaM datasets for analysis. Throughout the SDTM and ADaM data, participants are identified by their unique subject identifier. This is internal to the RECOVERY trial and independent of any external identifiers, such as NHS number, which may have been used to obtain linkage data. Neither these external identifiers nor other direct identifiers such as names and addresses are included in datasets for analysis.

Under the sub-license model, a RECOVERY data analyst will create the SDTM and AdAM outputs containing all the variables for each of the drugs evaluated in RECOVERY. These datasets will be approved by the information asset owner and will be submitted to the platform via the secure portal. No raw data will be submitted.

The RECOVERY trial, as a contributor of data to the platform, will agree a Terms of Submission with the IDDO prior to submission, which outlines the rights of the contributor, their obligations, the terms of processing and how the data will be used.

How researchers can request to access data?

Potential applicants are able to explore the availability of data via a Data Inventory, accessible via the IDDO website.

Data Access Application form:

The Data Requestor completes a Data Access Application Form. This contains details of the following: researcher; organisation, and team outlining role in analysis; research plan, including objectives, analysis, ethical considerations, publication, funding and any relevant scientific review, and data variables. The completed application form is submitted to the IDDO.

Review Considerations:

Upon receipt of a request, the Data Access Requests will be sent to the RECOVERY trial Principal Investigators to confirm that the request is in line with the expectations of the trial participants (based on the consent materials) and will maintain the scientific value and validity of the trial.

Data Access Request, and feedback from the RECOVERY Principal Investigators, will be considered by the IDDO Data Access Committee. This will take into account the following:

a) The researcher is working in a field relevant to COVID-19, and has a formal affiliation to a health, research, humanitarian, government, inter-government or academic institution with legal status;

b) The researcher is suitably qualified, either an academic record consistent with execution of the proposed analysis, or the support of a supervising co-applicant with appropriate expertise;

c) Availability of sufficient funding;

d) No prior breach of Data Transfer Agreements;

e) No concerns with respect to the ethical aspects of the application. This includes necessary legal, scientific and/or ethics approvals from their institution.;

f) No conflict of interest;

g) Impact in terms of collaboration and knowledge sharing

Approvals and contracts:

All required scientific and ethical reviews and/or approvals must be obtained before any Data are disclosed.

Prior to release a Data Transfer Agreement (DTA) will put in place between the University of Oxford and the recipient. This agreement shall limit purpose, preventing re-identification, and permit audit and destruction of data upon expiry or termination of the agreement. Additionally, if the recipient is in breach of any obligations, it may lead to termination of agreement. To enable harmonisation across all IDDO agreements, the term of the DTA is one year, in line with the NHS Digital standard for Length of Data Sharing Agreement. The DTA constitutes the Sublicense Agreement.

Data transfer:

The likelihood of re-identification of RECOVERY trial participants, accounting for all means reasonably likely, is mitigated by the use of CDISC standards, removal of external and direct identifiers, and use of a legally binding contract to prevent re-iden

COVID-19 is an emerging pathogen which presents a significant threat to the population in terms of increased morbidity and mortality, particularly among vulnerable groups such as those with pre-existing disease.

The primary objective is to provide reliable estimates of the effect of study treatments on in-hospital death (with subsidiary analyses of cause of death and death at various timepoints following discharge).

The secondary objectives are to assess the effects of study treatments on duration of hospital stay; the need for (and duration of) ventilation; and the need for renal replacement therapy.

The interim trial results will be monitored by an independent Data Monitoring Committee (DMC). The most important task for the DMC will be to assess whether the randomised comparisons in the study have provided evidence on mortality that is strong enough (with a range of uncertainty around the results that is narrow enough) to affect national and global treatment strategies. In such a circumstance, the DMC will inform the Trial Steering Committee who will make the results available to the public and amend the trial arms accordingly.

The data requested will be used to evaluate the efficacy and safety of the study treatments and will help shape the public health response. The rapid feed will be used to ensure that the trial Data Monitoring Committee have complete, up-to-date information on the major outcomes in the trial on which to base their decisions. If they find compelling evidence of efficacy or safety then that arm may be stopped and – if effective – added to standard care across the NHS.

All outputs produced will be in the form of aggregated reports with small number suppression applied.

Outputs delivered to date include the following results publications

RECOVERY Collaborative Group, Horby P, Lim WS, Emberson JR, Mafham M, Bell JL, Linsell L, et al. Dexamethasone in Hospitalized Patients with Covid-19. N Engl J Med 2021; 384:693-704

RECOVERY Collaborative Group. Effect of Hydroxychloroquine in Hospitalized Patients with Covid-19. N Engl J Med. 2020;383(21):2030-2040

RECOVERY Collaborative Group. Lopinavir–ritonavir in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. The Lancet, 2020;396(10259): 1345-1352

RECOVERY Collaborative Group. Azithromycin in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. Lancet. 2021;397(10274):605-612

RECOVERY Collaborative Group. Tocilizumab in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. Lancet. 2021 May 1;397(10285):1637-1645.

RECOVERY Collaborative Group. Convalescent plasma in patients admitted to hospital with COVID-19 (RECOVERY): a randomised controlled, open-label, platform trial. Lancet. 2021 May 29;397(10289):2049-2059

RECOVERY Collaborative Group. Aspirin in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. medRxiv 2021.06.08.21258132

RECOVERY Collaborative Group. Colchicine in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. medRxiv 2021.05.18.21257267

RECOVERY collaborative Group. Casirivimab and imdevimab in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. medRxiv 2021.06.15.21258542

By the end of July 2021, the RECOVERY trial had successfully established over 175 sites across the UK and recruited over 40,000 participants treated in hospital for COVID-19. The data linkage already established to received SUS+ and other data is providing important information to the Data Monitoring Committee on a regular basis about patients' recovery (i.e. discharge from hospital), in-hospital death and procedures required. Provision of complete and reliable data to the DMC throughout the trial is critical to allow robust assessment of the effects of the trial treatments with a major contribution to these data expected from analysis of the routine health care data requested under this agreement.

On Thursday 4 June 2020, in response to a request from the UK Medicines and Healthcare Products Regulatory Agency (MHRA), the independent Data Monitoring Committee conducted a further review of the data. The DMC recommended the chief investigators review the unblinded data on the hydroxychloroquine arm of the trial. The results included a total of 1542 patients randomised to hydroxychloroquine compared with 3132 patients randomised to usual care alone. There was no significant difference in the primary endpoint of 28-day mortality (25.7% hydroxychloroquine vs. 23.5% usual care; hazard ratio 1.11 [95% confidence interval 0.98-1.26]; p=0.10). There was also no evidence of beneficial effects on hospital stay duration or other outcomes. These results were released to the public and on the 15 July 2020, the U.S. Food and Drug Administration (FDA) revoked the emergency use authorization (EUA) that allowed for chloroquine phosphate and hydroxychloroquine sulfate donated to the Strategic National Stockpile to be used to treat certain hospitalized patients with COVID-19 when a clinical trial was unavailable, or participation in a clinical trial was not feasible. This result has major implications for countries around the world who were planning to scale manufacturing of these drugs in order to treat COVID patients.

On 8 June 2020, recruitment to the dexamethasone arm was halted since, in the view of the trial Steering Committee, sufficient patients had been enrolled to establish whether or not the drug had a meaningful benefit. On 16 June 2020, preliminary results were released. A total of 2104 patients were randomised to receive dexamethasone 6 mg once per day (either by mouth or by intravenous injection) for ten days and were compared with 4321 patients randomised to usual care alone. Among the patients who received usual care alone, 28-day mortality was highest in those who required ventilation (41%), intermediate in those patients who required oxygen only (25%), and lowest among those who did not require any respiratory intervention (13%). Dexamethasone reduced deaths by one-third in ventilated patients (rate ratio 0.65 [95% confidence interval 0.48 to 0.88]; p=0.0003) and by one fifth in other patients receiving oxygen only (0.80 [0.67 to 0.96]; p=0.0021). There was no benefit among those patients who did not require respiratory support (1.22 [0.86 to 1.75]; p=0.14). Based on these results, 1 death would be prevented by treatment of around 8 ventilated patients or around 25 patients requiring oxygen alone.

On the 16 June 2020, on the basis of these results, the MHRA issued an alert to health care providers in the UK recommending the use of dexamethasone in hospitalised patients with COVID who require oxygen or ventilation. In addition dexamethasone has also been added to the government’s parallel export list, which bans companies from buying medicines meant for UK patients and selling them on for a higher price in another country.

A recent analysis estimates that treatment with Dexamethasone may save between 4,000 and 27,000 lives in the UK by January 20201 (depending on the number of COVID-19 cases during Q3-4 2020) and potentially over half a million lives worldwide (https://www.medrxiv.org/content/10.1101/2020.07.29.20164269v1). Without the results from RECOVERY the effectiveness of this treatment would not be known.

On 11 February 2021, analyses of the RECOVERY trial showed that, among hospitalised COVID-19 patients who had low oxygen levels and elevated markers of inflammation in the blood, the rheumatoid arthritis drug Tocilizumab significantly reduced deaths: 596 (29%) of the patients in the tocilizumab group died within 28 days compared with 694 (33%) patients in the usual care group (rate ratio 0·86; [95% confidence interval [CI] 0·77 to 0·96]; p=0·007), an absolute difference of 4%. This means that for every 25 patients treated with tocilizumab, one additional life would be saved. Tocilizumab also increased the probability of discharge alive within 28 days from 47% to 54% (rate ratio 1·23, [95% CI 1·12 to 1·34], p<0·0001). These benefits were seen in all patient subgroups, including those requiring oxygen via a simple face mask through to those requiring mechanical ventilators in an intensive care unit and those already receiving steroid treatment. Tocilizumab is now standard of care within the NHS.

On 16 June 2021, analysis of 9785 patients hospitalised with COVID-19 randomly allocated to receive usual care plus the antibody combination treatment (casirivimab 4g with imdevimab 4g by intravenous infusion) or usual care alone produced practice changing results. Among the one-third of participants who were seronegative at baseline (ie they had not mounted a natural antibody response of their own) the monoclonal antibody combination significantly reduced the primary outcome of 28-day mortality by one-fifth compared with usual care alone (24% of patients in the antibody combination group died vs 30% of patients in the usual care group; rate ratio 0·80; 95% confidence interval 0·70–0·91; p=0·001). Thus, for every 100 such patients treated with the antibody combination, there would be six fewer deaths.

RECOVERY has also shown that a number of widely used treatments are not effective in COVID-19 including aspirin, colchicine, lopinavir-ritonavir and Convalescent plasma. These findings are important since they avoid unnecessary exposure to potentially hazardous treatments and allow healthcare systems and researchers to focus resources on treatments which are effective.

The University of Oxford requires access to data for a study entitled Randomised Evaluation of COVid-19 thERapY (RECOVERY).

In 2019 a novel coronavirus-induced disease (COVID-19) emerged in Wuhan, China. A month later the Chinese Center for Disease Control and Prevention identified a new betacoronavirus (SARS [Severe Acute Respiratory Syndrome] coronavirus 2, or SARS-CoV-2) as the aetiological (causing or contributing to the development of a disease or condition) agent. The clinical manifestations of COVID-19 range from asymptomatic infection or mild, transient symptoms to severe viral pneumonia with respiratory failure. As many patients do not progress to severe disease the overall case fatality rate per infected individual is low, but hospitals in areas with significant community transmission experienced a major increase in the number of hospitalized pneumonia patients, and the frequency of severe disease in hospitalised patients was recorded as high as 30%. The progression from prodrome (an early symptom indicating the onset of a disease or illness - in this case usually fever, fatigue and cough) to severe pneumonia requiring oxygen support or mechanical ventilation often takes one to two weeks after the onset of symptoms. The kinetics of viral replication in the respiratory tract are not well characterized, but this relatively slow progression provides a potential time window in which antiviral therapies could influence the course of disease.

This study aims to compare several different treatments that may be useful for patients with COVID-19. These treatments have been recommended by the expert panel that advises the Chief Medical Officer (CMO) in England.

The protocol allows reliable assessment of the effects of multiple different treatments (including re-purposed and novel drugs) on major outcomes in COVID-19. All patients will receive usual care for the participating hospital. From version 6.0 of the protocol, a factorial design was used such that eligible and consenting participants may be randomised to one of the treatment arms in Randomisation A and, simultaneously, to one of the treatment arms in Randomisation B.

Randomisation part A: Eligible patients may be randomly allocated between the following treatment arms (although not all arms may be available at any one time):

• No additional treatment:

• Corticosteroids (children ≤44 weeks gestational age, or >44 weeks gestational age with PIMS-TS only): Favourable immune response modulation by

corticosteroids might help treat severe acute respiratory coronavirus infections (including COVID-19, SARS and MERS) and prevent deterioration and the development of cardiovascular adverse events in PIMS-TS. Children >44 weeks gestational age with SARS-CoV-2 infection without PIMS-TS will be excluded from

this arm (as it is standard practice to treat with low-dose corticosteroids).

This comparison has now reported preliminary results (see benefits section).

• Azithromycin: Azithromycin is a macrolide antibiotic. The macrolides inhibit the growth of bacteria and are often prescribed to treat rather common bacterial infections. Azithromycin has immunomodulatory properties that has shown benefit in inflammatory lung disease.

• Intravenous immunoglobulin (children >44 weeks gestational age with PIMSTS only): a 2 g/kg single dose has been shown to be effective in immunomodulation

and preventing cardiovascular complications in Kawasaki disease, an inflammatory condition with overlapping clinical features with PIMS-TS.

Randomisation part B [UK only]: Simultaneously, eligible patients will be randomly allocated between the following treatment arms (provided there are no contraindications and the appropriate consent has been given):

• No additional treatment

• Convalescent plasma: Plasma from patients who have recovered from SARS-CoV2 infection may contain antibodies that can bind to and neutralise the virus. Infusion of convalescent plasma containing high concentrations of neutralising antibody may accelerate clearance of the virus and clinical improvement.

Convalescent plasma therapy had been given to at least 245 COVID-19 patients by the end of February 2020, and, according to a Chinese health official, 91 cases had shown improvement in clinical indicators and symptoms (http://www.xinhuanet.com/english/2020-02/28/c_138828177.htm). Five small case series (26 patients in total) have been published that report the use of convalescent plasma in people with COVID-19 infection.59-63 These studies have reported clinical and radiological improvements after treatment with convalescent plasma. However, these small uncontrolled studies have significant flaws and the reported effects are unreliable. Convalescent plasma is currently being tested in the REMAP-CAP trial (a separate trial) among patients on intensive care units.

• Synthetic neutralising antibodies: Synthetic human monoclonal antibodies have been developed to bind to and neutralise the virus. Such antibodies may accelerate clearance of the virus and clinical improvement.).

1.2.2 Second randomisation for patients with progressive COVID-19

Severe COVID-19 is associated with release of pro-inflammatory cytokines, such as IL-1,IL-6 and TNFα, and other markers of systemic inflammation including ferritin and C-reactive protein.3,6,7 There is a possibility that this response may cause or exacerbate lung injury, leading to life-threatening disease.

Participants with progressive COVID-19 (as evidenced by hypoxia and an inflammatory state) may undergo an optional second randomisation between the following treatment arms:

• No additional treatment: There are currently no approved immunomodulatory or other host-directed treatments to prevent the progression of COVID-19.

• Tocilizumab: (an immunosuppressive drug, mainly for the treatment of rheumatoid arthritis) or control (both in addition to the treatment assigned at the first randomisation). Tocilizumab is an interleukin-6 (IL-6) receptor antibody, which blocks a component of the immune response that may drive progression to Acute Respiratory Distress Syndrome. and the most severe forms of PIMS-TS.

Modifications to the number of treatment arms: Other arms can be added to the first or second randomisation if evidence emerges that there are suitable candidate therapeutics.

Conversely, in some patient populations, not all trial arms are appropriate (e.g. due to contraindications based on co-morbid conditions or concomitant medication); in some hospitals or countries, not all treatment arms will be available (e.g. due to manufacturing and supply shortages); and at some times, not all treatment arms will be active (e.g. due to lack of relevant approvals and contractual agreements). The Trial Steering Committee may elect to pause one or more of the arms in order to increase trial efficiency during a fluctuating epidemic. In any of these situations, randomisation will be between fewer arms. Depending on the availability and suitability of treatments, it may be allowed for participants to be randomised in only one part (A or B) of the main randomisation. This will not impact or change the level of data that is required from NHS Digital - as the cohort in the trial remains the same, regardless of what drug is being trialed.

Data from the trial will be regularly reviewed so that any effective treatment can be identified quickly and made available to all patients. The RECOVERY Trial team will constantly review information on new drugs and include promising ones in the trial.

Patients are eligible for the study if all of the following are true:

(i) Hospitalised

(ii) SARS-CoV-2 infection (clinically suspected or laboratory confirmed)

(iii) No medical history that might, in the opinion of the attending clinician, put the patient at significant risk if he/she were to participate in the trial

In addition, if the attending clinician believes that there is a specific contra-indication to one of the active drug treatment arms or that the patient should definitely be receiving one of the active drug treatment arms then that arm will not be available for randomisation for that patient. For patients who lack capacity, an advanced directive or behaviour that clearly indicates that they would not wish to participate in the trial would be considered sufficient reason to exclude them from the trial.

Informed consent should be obtained from each patient 16 years and over before enrolment into the study. However, if the patient lacks capacity to give consent due to the severity of their medical condition (e.g. acute respiratory failure or need for immediate ventilation) or prior disease, then consent may be obtained from a relative acting as the patient’s legally designated representative or independent doctor. Further consent will then be sought with the patient if they recover sufficiently. For children aged <16 years old consent will be sought from their parents or legal guardian. Where possible, children aged between 10-15 years old will also be asked for assent. Children aged ≥16 years old will be asked for consent as for adults. Witnessed consent may be obtained over the telephone or web video link if hospital visiting rules or parental infection mean a parent/guardian cannot be physically present.

Where Local Clinical Centre's (LCC's) plan to recruit children the principal investigator will co-opt support from a local paediatrician and/or neonatologists to oversee the management of children and infants in the trial.

By July 2020, the RECOVERY trial had successfully recruited over 12,000 participants, making it one of the largest trials assessing treatments for COVID-19 worldwide. Already this rapid recruitment and resulting large sample size has enabled the study to produce important results (see yielded benefits section). This remarkable participant recruitment is the result of a number of factors. Eligibility criteria are simple and trial processes (including paperwork) are minimised. Critically, the trial (and subsequent data collection) is designed to minimise the burden on front-line hospital staff working within an overstretched care system during a major pandemic. Given the high mortality rate among patients hospitalised with acute COVID-19, even treatments with only a moderate impact on survival or on hospital resources could be worthwhile.

The large sample size allows the focus of the COVID-19 Core Protocol to be the impact of candidate treatments on mortality, the key outcome which smaller studies are not able to assess.

For each pairwise comparison with the ‘no additional treatment’ arm, the primary objective is to provide reliable estimates of the effect of study treatments on all-cause mortality at 28 days after first randomisation (with subsidiary analyses of cause of death and of death at various timepoints following discharge).

The secondary objectives are to assess the effects of study treatments on duration of hospital stay; the need for (and duration of) ventilation; and, among patients not on ventilation at baseline, the composite endpoint of death or need for mechanical ventilation or ECMO (extracorporeal membrane oxygenation)

Other objectives include the assessment of the effects of study treatments on the need for renal replacement therapy and new major cardiac arrhythmias.

Study outcomes will be assessed based on data recorded up to 28 days and up to 6 months after the main randomization.

Data from routine healthcare records (including linkage to medical databases held by organisations such as NHS Digital) and from relevant research studies (such as UK Biobank and Genomics England (possibly in the future) will allow subsidiary analyses of the effect of the study treatments on particular non-fatal events (e.g. ascertained through linkage to Hospital Episode Statistics), the influence of pre-existing major co-morbidity (e.g. diabetes, heart disease, lung disease, hepatic insufficiency, severe depression, severe kidney impairment, immunosuppression), and longer-term outcomes (e.g. 6 month survival) as well as in particular sub-categories of patient (e.g. by genotype). For clarification - this agreement does not permit the linkage of NHS Digital data to data held by UK Biobank or Genomics England, and a subsequent amendment to this agreement will be put in place should this occur.

Follow-up information is to be collected on all study participants, irrespective of whether or not they complete the scheduled course of allocated study treatment. Study staff will seek Follow-up information through various means including medical staff, reviewing information from medical notes, routine healthcare systems, and registries.

The study team require the following information from NHS Digital:

- Prompt feed from SUS (Secondary Use Services) on hospital discharge. This will be provided on a fortnightly basis going forward.

- Quarterly feed on HES (Hospital Episode Statistics) data. This will change to a bi-annual feed from September 2021

- Monthly feeds on civil registration data (death certificate information). This will change to quarterly feeds from July 2021.

- CHESS (COVID-19 Hospitalization in England Surveillance System) Data. This will be provided on a fortnightly basis going forward.

- SGSS (COVID-19 Second Generation Surveillance System) Data. This will be provided on a fortnightly basis going forward.

- GDPPR Data (General Practice Extraction Service [GPES] Data for Pandemic Planning and Research) on a monthly basis. This will change to quarterly feeds from July 2021

- Cancer Registration Data

- Prescribing Data

The GDPPR and Primary Care Medicines data sets are requested for two key reasons. First they will be used to assess whether the effects of the treatments being tested in Recovery are different in people with certain medical conditions (e.g. lung disease or diabetes) or who were receiving particular medications (e.g. immunosupression) prior to joining the trial. To avoid burdening front-line staff, only limited data about other medical conditions is collected from the site staff at the point of enrollment and information about usual medications is not collected. Therefore these data sets will be invaluable in identifying relevant subgroups and will supplement information derived from the HES data regarding prior hospitalisations.

Second, information in these data sets after randomization will be used to determine the effects of the treatments on outcomes after discharge from hospital. For example, effects on the development on the development of long term lung problems, diabetes (particularly relevant to the steroid comparison), liver disease (particularly relevant to the lopinavir/ritonavir comparison) or on infections (particularly relevant to the tocolizumab comparison). It is understood that the GDPPR data is available for a limited time period only but will allow analysis of these and other outcomes following discharge from hospital.

Identifiable data is necessary for the purpose of data linkage and quality assurance of that linkage – the Data Sharing Agreement (DSA) requires measures are taken to ensure subsequent uses of the data will involve only data that has been appropriately pseudonymised.

Data requested includes a range of detailed information that is necessary for analysis and sub-group analysis of the effectiveness of the treatments – in particular to enable accounting for a range of biases and confounding factors relevant to the primary and secondary outcomes. Specifically:

• The trial has taken a pragmatic approach to recruitment and as a result may have included some patients who do not have confirmed Covid 19. For sub-group analysis and interpretation of the trial results, they need to know which patients have a confirmed positive Covid 19 test.

• Disease onset, dates for admission into hospital, ICU [Intensive Care Unit] and HDU [High Dependency Units] are important to track progression of the disease to understand the efficacy of treatment especially for those which inhibit viral replication.

• These data will allow the trial to assess the impact of trial drugs on the precise care and treatment delivered e.g the need for mechanical ventilation and any complications such as secondary infections.

• Risk factors and underlying conditions enable sub-group analysis of other factors that may have a bearing on the treatments being tested.

Additional datasets requested October 2020:

The RECOVERY team are now requesting cancer registrations and BSA Medicines datasets. The cancer registrations will be important in identifying long-term outcomes in the cohort and will help to reliably assess the long-term safety of the treatments tested in RECOVERY. This is particularly important as treatment arms now include medicines which have not previously been assessed in large trials with long follow-up (e.g. tocilizumab) or are currently unlicensed (e.g. REGN-COV2).

Cancer Registration data will be provided on an annual basis for the cohort.

The BSA Medicines dataset will be used to further characterise the cohort at baseline, allowing analysis of the effects of the treatments in different types of patient. In addition, refreshes of these data are requested as part of the long-term follow-up as they help to identify new diagnoses (e.g. diabetes). This will help the team identify long-term outcomes in the cohort and assess the effects of the treatments on these outcomes.

This study is supported by a grant to the University of Oxford from UK Research and Innovation/National Institute for Health Research (NIHR) and by core funding provided by NIHR Oxford Biomedical Research Centre, the Wellcome Trust, the Bill and Melinda Gates Foundation, Health Data Research UK (HDRUK), and the Medical Research Council Population Health Research Unit, and NIHR Clinical Trials Unit Support Funding.

The new trial has been classed as an Urgent Public Health Research Study. It is one of a round of projects to receive £10.5 million as part of the £20 million rapid research response funded by UK Research and Innovation, and by the Department of Health and Social Care through the National Institute for Health Research.

Independent External Validation of Results:

The RECOVERY trial is generating high profile results which change practice and there will be considerable external scrutiny of the trial processes. In order to provide an independent assessment of the validity of the research findings from the RECOVERY trial, the researchers at the University of Oxford have asked an independent statistician and associated team at the University of Bristol to reproduce the key findings using a copy of the trial analysis data-set.

This dataset does not include the full files provided to the RECOVERY team under this agreement but does include individual participant data from those files, such as the date on which a particular medical procedure took place for a participant. In line with data minimisation principles, only those fields required to replicate the analyses will be provided and direct identifiers (such as name, date of birth and NHS number) will be removed. The data will be transferred securely using password protection and encryption. This amendment adds the University of Bristol as a data processor so they can receive the data for this purpose. An appropriate controller processor agreement will be in place between the University of Oxford and the University of Bristol prior to any data transfer.

The University of Oxford are unable to provide external researchers with secure remote access to the data within the University of Oxford servers for technical reasons. This is something that will be considered in the future but has to be weighed against other demands on the department. The analysis team have shared individual participant data with external researchers before (outside of this agreement in terms of other research projects) and have a process for securely transferring encrypted files. A controller - processor agreement with Bristol University will be in place prior to data transfer. The duration of processing will be 12 months to enable the Professor to validate the results of future comparisons as needed.

The University of Bristol will access the data via remote connection to the University of Bristol Server - using University issued encrypted lap tops from home addresses. Data will be downloaded to these encrypted lap tops also.

The University of Oxford invited this Leading Independent Statistician to undertake this external validation due to previous academic and clinical experience. The statistician is leading the World Health Organization’s meta-analysis of dexamethasone trials and as such is an expert in the clinical context and methodological issues in relation to a trial such as RECOVERY. The Lead Statistician (from the team at Bristol) is a Professor of Medical Statistics and Epidemiology at University of Bristol and also the Director of Health Data Research UK South-West.

The task is sufficiently well defined to make the independent validation possible whist Oxford remain the data controller, and Bristol the data processor. The University of Oxford is asking the University of Bristol to validate the published results and the type of analyses to be done are already pre-specified (as they always would be in this type of trial) in the Statistical Analysis Plan (published by the RECOVERY team on the website prior to analysing the results https://www.recoverytrial.net/results).

The University of Oxford has designed two types of data outputs which the RECOVERY trial needs to onwardly share:

• Analysis Data Model (ADaM) output

• Study Data Tabulation Model (SDTM) output

These trial analysis datasets, are developed using data sourced from Data provided by NHS Digital under this Data Sharing Agreement along with data from other sources, and are subject to a range of processing in line with standards published by the Clinical Data Interchange Standards Consortium (or CDISC).

The ADaM and SDTM outputs for the RECOVERY trial undergone technical assessment and IG assessment by NHS Digital and NHS Digital has determined that, subject to being shared under strict conditions (specified below), the datasets are sufficiently low risk and can be onward shared by the University of Oxford with the manufacturer or regulator where necessary.

SDTM output:

Based on a desk review, responses to specific questions from the University of Oxford and a general self-certification from the University of Oxford , NHS Digital has concluded that none of the fields examined in the SDTM output can be used on their own or with other information in the onwardly transferred file or in the public domain to reconstitute fully fields in the data provided to the University of Oxford.

A highly motivated and technically literate intruder could attempt to do this for a subset of the information in a specific field but this would not be the information in the field in its entirety and accuracy in doing this could not be proven or guaranteed.

ADaM output:

Based on a desk review and further information from the RECOVERY team at the University of Oxford in response to specific questions, NHS Digital has concluded that none of the fields examined in the ADaM output can be used on their own or with other information in the onwardly transferred file or in the public domain to reconstitute fully fields in the data provided to the University of Oxford under this Agreement.

Based on the above assessments, the SDTM and ADaM datasets can be onwardly shared under the following conditions:

1. The dataset may be shared with the manufacturer of the treatment(s) evaluated in the RECOVERY trial for the purpose of the manufacturer using the data in ongoing research in the development programme for the treatment (for example, evaluating additional beneficial or harmful effects of the treatment or evaluating the effects of the treatment in particular patient populations). Any such uses will comply with the following controls.

2. The dataset may be shared by either the University of Oxford or the manufacturer with a competent authority (regulator) in circumstances where the RECOVERY trial has a legal obligation to disclose data with the regulator to allow trial results to be evaluated to demonstrate the safety and effectiveness of the treatment(s) tested. Regulators include the Medicines & Healthcare products Regulatory Agency (MHRA) in the UK, the Food and Drug Administration (FDA) in the United States, and any competent authority in the European Medicines Agency (EMA).

3. The dataset shared must conform to the SDTM and ADaM outputs agreed with NHS Digital and may not contain additional variables.

4. The dataset must not be shared alongside any source data provided under this Agreement.

5. The dataset may only be shared under the conditions that the recipient(s):

i. must not combine it with other datasets which could potentially increase the risk of reidentification for individuals in the dataset;

ii. must not attempt to re-identify individuals in the dataset;

iii. must not onwardly share the dataset other than permitted sharing between the manufacturer and the regulatory authority as required;

iv. must use the dataset for any purpose other than those defined above, and

v. must not publish the data sourced under this Agreement.

6. University of Oxford has contractual controls in place with manufacturers, which will prevent re-identification of trial participants and limit the purposes for use of the data.

The datasets to be onwardly shared contain no dates nor the order of randomisation.

The SDTM and ADaM datasets has its own set of unique identifiers. It is reversible in the event of a query from a regulatory authority but only by the team in Oxford.

NHS Digital source data files are not disclosed.

NHS Digital require the right to be informed of data shared for the above purpose.

NHS Digital waive the right to audit the regulator or manufacturer of treatment(s) evaluated in the RECOVERY trial.

COVID-19 is an emerging pathogen which presents a significant threat to the population in terms of increased morbidity and mortality, particularly among vulnerable groups such as those with pre-existing disease.

The primary objective is to provide reliable estimates of the effect of study treatments on in-hospital death (with subsidiary analyses of cause of death and death at various timepoints following discharge).

The secondary objectives are to assess the effects of study treatments on duration of hospital stay; the need for (and duration of) ventilation; and the need for renal replacement therapy.

The interim trial results will be monitored by an independent Data Monitoring Committee (DMC). The most important task for the DMC will be to assess whether the randomised comparisons in the study have provided evidence on mortality that is strong enough (with a range of uncertainty around the results that is narrow enough) to affect national and global treatment strategies. In such a circumstance, the DMC will inform the Trial Steering Committee who will make the results available to the public and amend the trial arms accordingly.

The data requested will be used to evaluate the efficacy and safety of the study treatments and will help shape the public health response. The rapid feed will be used to ensure that the trial Data Monitoring Committee have complete, up-to-date information on the major outcomes in the trial on which to base their decisions. If they find compelling evidence of efficacy or safety then that arm may be stopped and – if effective – added to standard care across the NHS.

All outputs produced will be in the form of aggregated reports with small number suppression applied.

A paper was recently published in the New England Journal of Medicine entitled "Dexamethasone in Hospitalized Patients with Covid-19 — Preliminary Report" (17/07/2020) detailing the results.

By the end of July 2020, the RECOVERY trial had successfully established over 175 sites across the UK and recruited over 11,000 participants treated in hospital for COVID-19. The data linkage already established to received SUS+ and other data is providing important information to the Data Monitoring Committee on a weekly basis about patients' recovery (i.e. discharge from hospital), in-hospital death and procedures required. Provision of complete and reliable data to the DMC through May and early June 2020 is critical to allow robust assessment of the effects of the trial treatments with a major contribution to these data expected from analysis of the routine health care data requested under this agreement.

On Thursday 4 June, in response to a request from the UK Medicines and Healthcare Products Regulatory Agency (MHRA), the independent Data Monitoring Committee conducted a further review of the data. The DMC recommended the chief investigators review the unblinded data on the hydroxychloroquine arm of the trial. The results included a total of 1542 patients randomised to hydroxychloroquine compared with 3132 patients randomised to usual care alone. There was no significant difference in the primary endpoint of 28-day mortality (25.7% hydroxychloroquine vs. 23.5% usual care; hazard ratio 1.11 [95% confidence interval 0.98-1.26]; p=0.10). There was also no evidence of beneficial effects on hospital stay duration or other outcomes. These results were released to the public and on the 15 July 2020, the U.S. Food and Drug Administration (FDA) revoked the emergency use authorization (EUA) that allowed for chloroquine phosphate and hydroxychloroquine sulfate donated to the Strategic National Stockpile to be used to treat certain hospitalized patients with COVID-19 when a clinical trial was unavailable, or participation in a clinical trial was not feasible. This result has major implications for countries around the world who were planning to scale manufacturing of these drugs in order to treat COVID patients.

On 8 June 2020, recruitment to the dexamethasone arm was halted since, in the view of the trial Steering Committee, sufficient patients had been enrolled to establish whether or not the drug had a meaningful benefit. On 16 June 2020, preliminary results were released. A total of 2104 patients were randomised to receive dexamethasone 6 mg once per day (either by mouth or by intravenous injection) for ten days and were compared with 4321 patients randomised to usual care alone. Among the patients who received usual care alone, 28-day mortality was highest in those who required ventilation (41%), intermediate in those patients who required oxygen only (25%), and lowest among those who did not require any respiratory intervention (13%). Dexamethasone reduced deaths by one-third in ventilated patients (rate ratio 0.65 [95% confidence interval 0.48 to 0.88]; p=0.0003) and by one fifth in other patients receiving oxygen only (0.80 [0.67 to 0.96]; p=0.0021). There was no benefit among those patients who did not require respiratory support (1.22 [0.86 to 1.75]; p=0.14). Based on these results, 1 death would be prevented by treatment of around 8 ventilated patients or around 25 patients requiring oxygen alone.

On the 16 June 2020, on the basis of these results, the MHRA issued an alert to health care providers in the UK recommending the use of dexamethasone in hospitalised patients with COVID who require oxygen or ventilation. In addition dexamethasone has also been added to the government’s parallel export list, which bans companies from buying medicines meant for UK patients and selling them on for a higher price in another country.

A recent analysis estimates that treatment with Dexamethasone may save between 4,000 and 27,000 lives in the UK by January 20201 (depending on the number of COVID-19 cases during Q3-4 2020) and potentially over half a million lives worldwide (https://www.medrxiv.org/content/10.1101/2020.07.29.20164269v1). Without the results from RECOVERY the effectiveness of this treatment would not be known.

A paper was recently published in the New England Journal of Medicine entitled "Dexamethasone in Hospitalized Patients with Covid-19 — Preliminary Report" (17/07/2020) detailing the results.

University of Oxford are requesting to use the NHS Digital Clinical Trials Service for access to data for a study entitled Randomised Evaluation of COVid-19 thERapY (RECOVERY).

In 2019 a novel coronavirus-induced disease (COVID-19) emerged in Wuhan, China. A month later the Chinese Center for Disease Control and Prevention identified a new betacoronavirus (SARS [Severe Acute Respiratory Syndrome] coronavirus 2, or SARS-CoV-2) as the aetiological (causing or contributing to the development of a disease or condition) agent. The clinical manifestations of COVID-19 range from asymptomatic infection or mild, transient symptoms to severe viral pneumonia with respiratory failure. As many patients do not progress to severe disease the overall case fatality rate per infected individual is low, but hospitals in areas with significant community transmission experienced a major increase in the number of hospitalized pneumonia patients, and the frequency of severe disease in hospitalised patients was recorded as high as 30%. The progression from prodrome (an early symptom indicating the onset of a disease or illness - in this case usually fever, fatigue and cough) to severe pneumonia requiring oxygen support or mechanical ventilation often takes one to two weeks after the onset of symptoms. The kinetics of viral replication in the respiratory tract are not well characterized, but this relatively slow progression provides a potential time window in which antiviral therapies could influence the course of disease.

This study aims to compare several different treatments that may be useful for patients with COVID-19. These treatments have been recommended by the expert panel that advises the Chief Medical Officer (CMO) in England.

The protocol allows reliable assessment of the effects of multiple different treatments (including re-purposed and novel drugs) on major outcomes in COVID-19. All patients will receive usual care for the participating hospital. From version 6.0 of the protocol, a factorial design was used such that eligible and

consenting participants may be randomised to one of the treatment arms in Randomisation A and, simultaneously, to one of the treatment arms in Randomisation B.

Randomisation part A: Eligible patients may be randomly allocated between the following treatment arms (although not all arms may be available at any one time):

 No additional treatment:

 Corticosteroids (children ≤44 weeks gestational age, or >44 weeks gestational age with PIMS-TS only): Favourable immune response modulation by

corticosteroids might help treat severe acute respiratory coronavirus infections (including COVID-19, SARS and MERS) and prevent deterioration and the development of cardiovascular adverse events in PIMS-TS. Children >44 weeks gestational age with SARS-CoV-2 infection without PIMS-TS will be excluded from

this arm (as it is standard practice to treat with low-dose corticosteroids).

This comparison has now reported preliminary results (see benefits section).

 Azithromycin: Azithromycin is a macrolide antibiotic. The macrolides inhibit the growth of bacteria and are often prescribed to treat rather common bacterial infections. Azithromycin has immunomodulatory properties that has shown benefit in inflammatory lung disease.

 Intravenous immunoglobulin (children >44 weeks gestational age with PIMSTS only): a 2 g/kg single dose has been shown to be effective in immunomodulation

and preventing cardiovascular complications in Kawasaki disease, an inflammatory condition with overlapping clinical features with PIMS-TS.

Randomisation part B [UK only]: Simultaneously, eligible patients will be randomly allocated between the following treatment arms (provided there are no contraindications and the appropriate consent has been given):

 No additional treatment

 Convalescent plasma: Plasma from patients who have recovered from SARS-CoV2 infection may contain antibodies that can bind to and neutralise the virus. Infusion of convalescent plasma containing high concentrations of neutralising antibody may accelerate clearance of the virus and clinical improvement.

Convalescent plasma therapy had been given to at least 245 COVID-19 patients by the end of February 2020, and, according to a Chinese health official, 91 cases had shown improvement in clinical indicators and symptoms (http://www.xinhuanet.com/english/2020-02/28/c_138828177.htm). Five small case series (26 patients in total) have been published that report the use of convalescent plasma in people with COVID-19 infection.59-63 These studies have reported clinical and radiological improvements after treatment with convalescent plasma. However, these small uncontrolled studies have significant flaws and the reported effects are unreliable. Convalescent plasma is currently being tested in the REMAP-CAP trial (a seperate trial) among patients on intensive care units.

 Synthetic neutralising antibodies: Synthetic human monoclonal antibodies have been developed to bind to and neutralise the virus. Such antibodies may accelerate clearance of the virus and clinical improvement.).

1.2.2 Second randomisation for patients with progressive COVID-19

Severe COVID-19 is associated with release of pro-inflammatory cytokines, such as IL-1,IL-6 and TNFα, and other markers of systemic inflammation including ferritin and C-reactive protein.3,6,7 There is a possibility that this response may cause or exacerbate lung injury, leading to life-threatening disease.

Participants with progressive COVID-19 (as evidenced by hypoxia and an inflammatory state) may undergo an optional second randomisation between the following treatment arms:

 No additional treatment: There are currently no approved immunomodulatory or other host-directed treatments to prevent the progression of COVID-19.

 Tocilizumab: (an immunosuppressive drug, mainly for the treatment of rheumatoid arthritis) or control (both in addition to the treatment assigned at the first randomisation). Tocilizumab is an interleukin-6 (IL-6) receptor antibody, which blocks a component of the immune response that may drive progression to Acute Respiratory Distress Syndrome. and the most severe forms of PIMS-TS.

Modifications to the number of treatment arms: Other arms can be added to the first or second randomisation if evidence emerges that there are suitable candidate therapeutics.

Conversely, in some patient populations, not all trial arms are appropriate (e.g. due to contraindications based on co-morbid conditions or concomitant medication); in some hospitals or countries, not all treatment arms will be available (e.g. due to manufacturing and supply shortages); and at some times, not all treatment arms will be active (e.g. due to lack of relevant approvals and contractual agreements). The Trial Steering Committee may elect to pause one or more of the arms in order to increase trial efficiency during a fluctuating epidemic. In any of these situations, randomisation will be between fewer arms. Depending on the availability and suitability of treatments, it may be allowed for participants to be randomised in only one part (A or B) of the main randomisation. This will not impact or change the level of data that is required from NHS Digital - as the cohort in the trial remains the same, regardless of what drug is being trialed.

Data from the trial will be regularly reviewed so that any effective treatment can be identified quickly and made available to all patients. The RECOVERY Trial team will constantly review information on new drugs and include promising ones in the trial.

Patients are eligible for the study if all of the following are true:

(i) Hospitalised

(ii) SARS-CoV-2 infection (clinically suspected or laboratory confirmed)

(iii) No medical history that might, in the opinion of the attending clinician, put the patient at significant risk if he/she were to participate in the trial

In addition, if the attending clinician believes that there is a specific contra-indication to one of the active drug treatment arms or that the patient should definitely be receiving one of the active drug treatment arms then that arm will not be available for randomisation for that patient. For patients who lack capacity, an advanced directive or behaviour that clearly indicates that they would not wish to participate in the trial would be considered sufficient reason to exclude them from the trial.

Informed consent should be obtained from each patient 16 years and over before enrolment into the study. However, if the patient lacks capacity to give consent due to the severity of their medical condition (e.g. acute respiratory failure or need for immediate ventilation) or prior disease, then consent may be obtained from a relative acting as the patient’s legally designated representative or independent doctor. Further consent will then be sought with the patient if they recover sufficiently. For children aged <16 years old consent will be sought from their parents or legal guardian. Where possible, children aged between 10-15 years old will also be asked for assent. Children aged ≥16 years old will be asked for consent as for adults. Witnessed consent may be obtained over the telephone or web video link if hospital visiting rules or parental infection mean a parent/guardian cannot be physically present.

Where Local Clinical Centre's (LCC's) plan to recruit children the principal investigator will co-opt support from a local paediatrician and/or neonatologists to oversee the management of children and infants in the trial.

By July 2020, the RECOVERY trial had successfully recruited over 12,000 participants, making it one of the largest trials assessing treatments for COVID-19 worldwide. Already this rapid recruitment and resulting large sample size has enabled the study to produce important results (see yielded benefits section). This remarkable participant recruitment is the result of a number of factors. Eligibility criteria are simple and trial processes (including paperwork) are minimised. Critically, the trial (and subsequent data collection) is designed to minimise the burden on front-line hospital staff working within an overstretched care system during a major pandemic. Given the high mortality rate among patients hospitalised with acute COVID-19, even treatments with only a moderate impact on survival or on hospital resources could be worthwhile.

The large sample size allows the focus of the COVID-19 Core Protocol to be the impact of candidate treatments on mortality, the key outcome which smaller studies are not able to assess.

For each pairwise comparison with the ‘no additional treatment’ arm, the primary objective is to provide reliable estimates of the effect of study treatments on all-cause mortality at 28 days after first randomisation (with subsidiary analyses of cause of death and of death at various timepoints following discharge).

The secondary objectives are to assess the effects of study treatments on duration of hospital stay; the need for (and duration of) ventilation; and, among patients not on ventilation at baseline, the composite endpoint of death or need for mechanical ventilation or ECMO (extracorporeal membrane oxygenation)

Other objectives include the assessment of the effects of study treatments on the need for renal replacement therapy and new major cardiac arrhythmias.

Study outcomes will be assessed based on data recorded up to 28 days and up to 6 months after the main randomization.

Data from routine healthcare records (including linkage to medical databases held by organisations such as NHS Digital) and from relevant research studies (such as UK Biobank and Genomics England (possibly in the future) will allow subsidiary analyses of the effect of the study treatments on particular non-fatal events (e.g. ascertained through linkage to Hospital Episode Statistics), the influence of pre-existing major co-morbidity (e.g. diabetes, heart disease, lung disease, hepatic insufficiency, severe depression, severe kidney impairment, immunosuppression), and longer-term outcomes (e.g. 6 month survival) as well as in particular sub-categories of patient (e.g. by genotype). For clarification - this agreement does not permit the linkage of NHS Digital data to data held by UK Biobank or Genomics England, and a subsequent amendment to this agreement will be put in place should this occur.

Follow-up information is to be collected on all study participants, irrespective of whether or not they complete the scheduled course of allocated study treatment. Study staff will seek Follow-up information through various means including medical staff, reviewing information from medical notes, routine healthcare systems, and registries.

The study team require the following information from NHS Digital:

- Prompt feed from SUS (Secondary Use Services) on hospital discharge. This will be provided on a fortnightly basis going forward.

- Quarterly feed on HES (Hospital Episode Statistics) data. This will change to a bi-annual feed from September 2021

- Monthly feeds on civil registration data (death certificate information). This will change to quarterly feeds from July 2021.

- CHESS (COVID-19 Hospitalization in England Surveillance System) Data. This will be provided on a fortnightly basis going forward.

- SGSS (COVID-19 Second Generation Surveillance System) Data. This will be provided on a fortnightly basis going forward.

- GDPPR Data (General Practice Extraction Service [GPES] Data for Pandemic Planning and Research) on a monthly basis. This will change to quarterly feeds from July 2021

- Cancer Registration Data

- Prescribing Data

The GDPPR and Primary Care Medicines data sets are requested for two key reasons. First they will be used to assess whether the effects of the treatments being tested in Recovery are different in people with certain medical conditions (e.g. lung disease or diabetes) or who were receiving particular medications (e.g. immunosupression) prior to joining the trial. To avoid burdening front-line staff, only limited data about other medical conditions is collected from the site staff at the point of enrollment and information about usual medications is not collected. Therefore these data sets will be invaluable in identifying relevant subgroups and will supplement information derived from the HES data regarding prior hospitalisations.

Second, information in these data sets after randomization will be used to determine the effects of the treatments on outcomes after discharge from hospital. For example, effects on the development on the development of long term lung problems, diabetes (particularly relevant to the steroid comparison), liver disease (particularly relevant to the lopinavir/ritonavir comparison) or on infections (particularly relevant to the tocolizumab comparison). It is understood that the GDPPR data is available for a limited time period only but will allow analysis of these and other outcomes following discharge from hospital.

Identifiable data is necessary for the purpose of data linkage and quality assurance of that linkage – the Data Sharing Agreement (DSA) requires measures are taken to ensure subsequent uses of the data will involve only data that has been appropriately pseudonymised.

Data requested includes a range of detailed information that is necessary for analysis and sub-group analysis of the effectiveness of the treatments – in particular to enable accounting for a range of biases and confounding factors relevant to the primary and secondary outcomes. Specifically:

• The trial has taken a pragmatic approach to recruitment and as a result may have included some patients who do not have confirmed Covid 19. For sub-group analysis and interpretation of the trial results, they need to know which patients have a confirmed positive Covid 19 test.

• Disease onset, dates for admission into hospital, ICU [Intensive Care Unit] and HDU [High Dependency Units] are important to track progression of the disease to understand the efficacy of treatment esp for those which inhibit viral replication.

• These data will allow the trial to assess the impact of trial drugs on the precise care and treatment delivered e.g the need for mechanical ventilation and any complications such as secondary infections.

• Risk factors and underlying conditions enable sub-group analysis of other factors that may have a bearing on the treatments being tested.

Additional datasets requested October 2020

The RECOVERY team are now requesting cancer registrations and BSA Medicines datasets. The cancer registrations will be important in identifying long-term outcomes in the cohort and will help to reliably assess the long-term safety of the treatments tested in RECOVERY. This is particularly important as treatment arms now include medicines which have not previously been assessed in large trials with long follow-up (e.g. tocilizumab) or are currently unlicensed (e.g. REGN-COV2).

Cancer Registration data will be provided on an annual basis for the cohort.

The BSA Medicines dataset will be used to further characterise the cohort at baseline, allowing analysis of the effects of the treatments in different types of patient. In addition, refreshes of these data are requested as part of the long-term follow-up as they help to identify new diagnoses (e.g. diabetes). This will help the team identify long-term outcomes in the cohort and assess the effects of the treatments on these outcomes.

This study is supported by a grant to the University of Oxford from UK Research and Innovation/National Institute for Health Research (NIHR) and by core funding provided by NIHR Oxford Biomedical Research Centre, the Wellcome Trust, the Bill and Melinda Gates Foundation, Health Data Research UK (HDRUK), and the Medical Research Council Population Health Research Unit, and NIHR Clinical Trials Unit Support Funding.

The new trial has been classed as an Urgent Public Health Research Study. It is one of a round of projects to receive £10.5 million as part of the £20 million rapid research response funded by UK Research and Innovation, and by the Department of Health and Social Care through the National Institute for Health Research.

Independent External Validation of Results:

The RECOVERY trial is generating high profile results which change practice and there will be considerable external scrutiny of the trial processes. In order to provide an independent assessment of the validity of the research findings from the RECOVERY trial, the researchers at the University of Oxford have asked an independent statistician and associated team at the University of Bristol to reproduce the key findings using a copy of the trial analysis data-set.

This data-set does not include the full files provided to the RECOVERY team under this agreement, but does include individual participant data derived from those files, such as the date on which a particular medical procedure took place for a participant. In line with data minimization principles, only those fields required to replicate the analyses will be provided and direct identifiers (such as name, date of birth and NHS number) will be removed. The data will be transferred securely using password protection and encryption. This amendment adds the University of Bristol as a data processor so they can receive the data for this purpose. An appropriate controller processor agreement will be in place between the University of Oxford and the University of Bristol prior to any data transfer.

The University of Oxford are unable to provide external researchers with secure remote access to the data within the University of Oxford servers for technical reasons. This is something that will be considered in the future but has to be weighed against other demands on the department. The analysis team have shared individual participant data with external researchers before (outside of this agreement in terms of other research projects) and have a process for securely transferring encrypted files. A controller - processor agreement with Bristol University will be in place prior to data transfer. The duration of processing will be 12 months to enable the Professor to validate the results of future comparisons as needed.

The University of Bristol will access the data via remote connection to the University of Bristol Server - using University issued encrypted lap tops from home addresses. Data will be downloaded to these encrypted lap tops also.

The University of Oxford invited this Leading Independent Statistician to undertake this external validation due to previous academic and clinical experience. The statistician is leading the World Health Organization’s meta-analysis of dexamethasone trials and as such is an expert in the clinical context and methodological issues in relation to a trial such as RECOVERY. The Lead Statistician (from the team at Bristol) is a Professor of Medical Statistics and Epidemiology at University of Bristol and also the Director of Health Data Research UK South-West.

The task is sufficiently well defined to make the independent validation possible whist Oxford remain the data controller, and Bristol the data processor. The University of Oxford is asking the University of Bristol to validate the published results and the type of analyses to be done are already pre-specified (as they always would be in this type of trial) in the Statistical Analysis Plan (published by the RECOVERY team on the website prior to analyzing the results https://www.recoverytrial.net/results).

COVID-19 is an emerging pathogen which presents a significant threat to the population in terms of increased morbidity and mortality, particularly among vulnerable groups such as those with pre-existing disease.

The primary objective is to provide reliable estimates of the effect of study treatments on in-hospital death (with subsidiary analyses of cause of death and death at various timepoints following discharge).

The secondary objectives are to assess the effects of study treatments on duration of hospital stay; the need for (and duration of) ventilation; and the need for renal replacement therapy.

The interim trial results will be monitored by an independent Data Monitoring Committee (DMC). The most important task for the DMC will be to assess whether the randomised comparisons in the study have provided evidence on mortality that is strong enough (with a range of uncertainty around the results that is narrow enough) to affect national and global treatment strategies. In such a circumstance, the DMC will inform the Trial Steering Committee who will make the results available to the public and amend the trial arms accordingly.

The data requested will be used to evaluate the efficacy and safety of the study treatments and will help shape the public health response. The rapid feed will be used to ensure that the trial Data Monitoring Committee have complete, up-to-date information on the major outcomes in the trial on which to base their decisions. If they find compelling evidence of efficacy or safety then that arm may be stopped and – if effective – added to standard care across the NHS.

All outputs produced will be in the form of aggregated reports with small number suppression applied.

A paper was recently published in the New England Journal of Medicine entitled "Dexamethasone in Hospitalized Patients with Covid-19 — Preliminary Report" (17/07/2020) detailing the results.

By the end of July 2020, the RECOVERY trial had successfully established over 175 sites across the UK and recruited over 11,000 participants treated in hospital for COVID-19. The data linkage already established to received SUS+ and other data is providing important information to the Data Monitoring Committee on a weekly basis about patients' recovery (i.e. discharge from hospital), in-hospital death and procedures required. Provision of complete and reliable data to the DMC through May and early June 2020 is critical to allow robust assessment of the effects of the the trial treatments with a major contribution to these data expected from analysis of the routine health care data requested under this agreement.

On Thursday 4 June, in response to a request from the UK Medicines and Healthcare Products Regulatory Agency (MHRA), the independent Data Monitoring Committee conducted a further review of the data. The DMC recommended the chief investigators review the unblinded data on the hydroxychloroquine arm of the trial. The results included a total of 1542 patients randomised to hydroxychloroquine compared with 3132 patients randomised to usual care alone. There was no significant difference in the primary endpoint of 28-day mortality (25.7% hydroxychloroquine vs. 23.5% usual care; hazard ratio 1.11 [95% confidence interval 0.98-1.26]; p=0.10). There was also no evidence of beneficial effects on hospital stay duration or other outcomes. These results were released to the public and on the 15 July 2020, the U.S. Food and Drug Administration (FDA) revoked the emergency use authorization (EUA) that allowed for chloroquine phosphate and hydroxychloroquine sulfate donated to the Strategic National Stockpile to be used to treat certain hospitalized patients with COVID-19 when a clinical trial was unavailable, or participation in a clinical trial was not feasible. This result has major implications for countries around the world who were planning to scale manufacturing of these drugs in order to treat COVID patients.

On 8 June 2020, recruitment to the dexamethasone arm was halted since, in the view of the trial Steering Committee, sufficient patients had been enrolled to establish whether or not the drug had a meaningful benefit. On 16 June 2020, preliminary results were released. A total of 2104 patients were randomised to receive dexamethasone 6 mg once per day (either by mouth or by intravenous injection) for ten days and were compared with 4321 patients randomised to usual care alone. Among the patients who received usual care alone, 28-day mortality was highest in those who required ventilation (41%), intermediate in those patients who required oxygen only (25%), and lowest among those who did not require any respiratory intervention (13%). Dexamethasone reduced deaths by one-third in ventilated patients (rate ratio 0.65 [95% confidence interval 0.48 to 0.88]; p=0.0003) and by one fifth in other patients receiving oxygen only (0.80 [0.67 to 0.96]; p=0.0021). There was no benefit among those patients who did not require respiratory support (1.22 [0.86 to 1.75]; p=0.14). Based on these results, 1 death would be prevented by treatment of around 8 ventilated patients or around 25 patients requiring oxygen alone.

On the 16 June 2020, on the basis of these results, the MHRA issued an alert to health care providers in the UK recommending the use of dexamethasone in hospitalised patients with COVID who require oxygen or ventilation. In addition dexamethasone has also been added to the government’s parallel export list, which bans companies from buying medicines meant for UK patients and selling them on for a higher price in another country.

A recent analysis estimates that treatment with Dexamethasone may save between 4,000 and 27,000 lives in the UK by January 20201 (depending on the number of COVID-19 cases during Q3-4 2020) and potentially over half a million lives worldwide (https://www.medrxiv.org/content/10.1101/2020.07.29.20164269v1). Without the results from RECOVERY the effectiveness of this treatment would not be known.

A paper was recently published in the New England Journal of Medicine entitled "Dexamethasone in Hospitalized Patients with Covid-19 — Preliminary Report" (17/07/2020) detailing the results.

University of Oxford are requesting to use the NHS Digital Clinical Trials Service for access to data for a study entitled Randomised Evaluation of COVid-19 thERapY (RECOVERY).

In 2019 a novel coronavirus-induced disease (COVID-19) emerged in Wuhan, China. A month later the Chinese Center for Disease Control and Prevention identified a new betacoronavirus (SARS [Severe Acute Respiratory Syndrome] coronavirus 2, or SARS-CoV-2) as the aetiological (causing or contributing to the development of a disease or condition) agent. The clinical manifestations of COVID-19 range from asymptomatic infection or mild, transient symptoms to severe viral pneumonia with respiratory failure. As many patients do not progress to severe disease the overall case fatality rate per infected individual is low, but hospitals in areas with significant community transmission experienced a major increase in the number of hospitalized pneumonia patients, and the frequency of severe disease in hospitalised patients was recorded as high as 30%. The progression from prodrome (an early symptom indicating the onset of a disease or illness - in this case usually fever, fatigue and cough) to severe pneumonia requiring oxygen support or mechanical ventilation often takes one to two weeks after the onset of symptoms. The kinetics of viral replication in the respiratory tract are not well characterized, but this relatively slow progression provides a potential time window in which antiviral therapies could influence the course of disease.

This study aims to compare several different treatments that may be useful for patients with COVID-19. These treatments have been recommended by the expert panel that advises the Chief Medical Officer (CMO) in England.

From version 6.0 of the protocol, a factorial design will be used such that eligible and consenting participants may be randomised to one of the treatment arms in Randomisation A and, simultaneously, to one of the treatment arms in Randomisation B. Treatments and randiomisation are as follows:

Randomisation may be between the following treatment arms in randomisation A (although not all arms may be available at any one time):

1. Usual care plus hydroxychloroquine vs usual care alone (no longer recruiting)

Hydroxychloroquine, a derivative of chloroquine, has been used for many decades to treat malaria and rheumatological diseases. It has antiviral activity against SARS-CoV-2 in cell culture. This comparison has now reported preliminary results (see benefits section).

2. Usual care plus Low dose corticosteroids (dexamethasone) vs usual care alone (no longer recruiting)

Favourable immune response modulation by low-dose corticosteroids might help treat severe acute respiratory coronavirus infections, including COVID-19, SARS and MERS [Middle East Respiratory Syndrome (MERS]. This comparison has now reported preliminary results (see benefits section).

3. Usual care plus lopinavir-ritnovavir vs usual care along (arm now closed - no longer recruiting)

Lopinavir-Ritonavir: (commonly used to treat human immunodeficiency viruses - HIV). Lopinavir is a human immunodeficiency virus 1 (HIV-1) protease inhibitor, which is combined with ritonavir to increase lopinavir’s plasma half-life. Lopinavir-Ritonavir has shown activity against SARS and MERS CoVs.

4. • Azithromycin: Usual care plus azithromyin vs usual care along (still recruiting)

Azithromycin is a macrolide antibiotic. The macrolides inhibit the growth of bacteria and are often prescribed to treat rather common bacterial infections. Azithromycin has immunomodulatory properties that has shown benefit in inflammatory lung disease.

Randomisation part B: Simultaneously, eligible patients will be randomly allocated between the following treatment arms (provided there are no contraindications and the appropriate consent has been given):

• Usual care plus convalescent plasma vs usual care: Plasma from patients who have recovered from SARS-CoV-2 infection may contain antibodies that can bind to and neutralise the virus. Infusion of convalescent plasma containing high concentrations of neutralising antibody may accelerate clearance of the virus and clinical improvement. Convalescent plasma therapy had been given to at least 245 COVID-19 patients by the end of February 2020, and, according to a Chinese health official, 91 cases had shown improvement in clinical indicators and symptoms (http://www.xinhuanet.com/english/2020-02/28/c_138828177.htm). Five small case series (26 patients in total) have been published that report the use of convalescent plasma in people with COVID-19 infection.59-63 These studies have reported clinical and radiological improvements after treatment with convalescent plasma. However, these small uncontrolled studies have significant flaws and the reported effects are unreliable. Convalescent plasma is currently being tested in the REMAP-CAP trial (a seperate trial) among patients on intensive care units.

A subset of participants with progressive COVID-19 (as evidenced by hypoxia and an inflammatory state) may undergo an optional second randomisation between the following treatment arms:

Usual care plus Tocilizumab vs usual care

Tocilizumab: (an immunosuppressive drug, mainly for the treatment of rheumatoid arthritis) or control (both in addition to the treatment assigned at the first randomisation).Tocilizumab is an interleukin-6 (IL-6) receptor antibody, which blocks a component of the immune response that may drive progression to Acute Respiratory Distress Syndrome.

Other arms can be added if evidence emerges that there are suitable candidate therapeutics. Conversely, in some patient populations, not all trial arms are appropriate (e.g. due to contraindications based on co-morbid conditions or concomitant medication); in some hospitals, not all treatment arms will be available (e.g. due to manufacturing and supply shortages); and at some times, not all treatment arms will be active (e.g. due to lack of relevant approvals and contractual agreements). The Trial Steering Committee may elect to pause one or more of the arms in order to increase trial efficiency during a fluctuating epidemic. Therefore,other treatments may be added to the protocol as time goes on and more information is gathered. This will not impact or change the level of data that is required from NHS Digital - as the cohort in the trial remains the same, regardless of what drug is being trialed.

Data from the trial will be regularly reviewed so that any effective treatment can be identified quickly and made available to all patients. The RECOVERY Trial team will constantly review information on new drugs and include promising ones in the trial.

Patients are eligible for the study if all of the following are true:

(i) Hospitalised

(ii) SARS-CoV-2 infection (clinically suspected or laboratory confirmed)

(iii) No medical history that might, in the opinion of the attending clinician, put the patient at significant risk if he/she were to participate in the trial

In addition, if the attending clinician believes that there is a specific contra-indication to one of the active drug treatment arms or that the patient should definitely be receiving one of the active drug treatment arms then that arm will not be available for randomisation for that patient. For patients who lack capacity, an advanced directive or behaviour that clearly indicates that they would not wish to participate in the trial would be considered sufficient reason to exclude them from the trial.

Informed consent should be obtained from each patient 16 years and over before enrolment into the study. However, if the patient lacks capacity to give consent due to the severity of their medical condition (e.g. acute respiratory failure or need for immediate ventilation) or prior disease, then consent may be obtained from a relative acting as the patient’s legally designated representative or independent doctor. Further consent will then be sought with the patient if they recover sufficiently. For children aged <16 years old consent will be sought from their parents or legal guardian. Where possible, children aged between 10-15 years old will also be asked for assent. Children aged ≥16 years old will be asked for consent as for adults. Witnessed consent may be obtained over the telephone or web video link if hospital visiting rules or parental infection mean a parent/guardian cannot be physically present.

Where Local Clinical Centre's (LCC's) plan to recruit children the principal investigator will co-opt support from a local paediatrician and/or neonatologists to oversee the management of children and infants in the trial.

By July 2020, the RECOVERY trial had successfully recruited over 12,000 participants, making it one of the largest trials assessing treatments for COVID-19 worldwide. Already this rapid recruitment and resulting large sample size has enabled the study to produce important results (see yielded benefits section). This remarkable participant recruitment is the result of a number of factors. Eligibility criteria are simple and trial processes (including paperwork) are minimised. Critically, the trial (and subsequent data collection) is designed to minimise the burden on front-line hospital staff working within an overstretched care system during a major pandemic. Given the high mortality rate among patients hospitalised with acute COVID-19, even treatments with only a moderate impact on survival or on hospital resources could be worthwhile.

The large sample size allows the focus of the COVID-19 Core Protocol to be the impact of candidate treatments on mortality, the key outcome which smaller studies are not able to assess.

For each pairwise comparison with the ‘no additional treatment’ arm, the primary objective is to provide reliable estimates of the effect of study treatments on all-cause mortality at 28 days after first randomisation (with subsidiary analyses of cause of death and of death at various timepoints following discharge).

The secondary objectives are to assess the effects of study treatments on duration of hospital stay; the need for (and duration of) ventilation; and, among patients not on ventilation at baseline, the composite endpoint of death or need for mechanical ventilation or ECMO (extracorporeal membrane oxygenation)

Other objectives include the assessment of the effects of study treatments on the need for renal replacement therapy and new major cardiac arrhythmias.

Study outcomes will be assessed based on data recorded up to 28 days and up to 6 months after the main randomization.

Data from routine healthcare records (including linkage to medical databases held by organisations such as NHS Digital) and from relevant research studies (such as UK Biobank and Genomics England (possibly in the future) will allow subsidiary analyses of the effect of the study treatments on particular non-fatal events (e.g. ascertained through linkage to Hospital Episode Statistics), the influence of pre-existing major co-morbidity (e.g. diabetes, heart disease, lung disease, hepatic insufficiency, severe depression, severe kidney impairment, immunosuppression), and longer-term outcomes (e.g. 6 month survival) as well as in particular sub-categories of patient (e.g. by genotype). For clarification - this agreement does not permit the linkage of NHS Digital data to data held by UK Biobank or Genomics England, and a subsequent amendment to this agreement will be put in place should this occur.

Follow-up information is to be collected on all study participants, irrespective of whether or not they complete the scheduled course of allocated study treatment. Study staff will seek Follow-up information through various means including medical staff, reviewing information from medical notes, routine healthcare systems, and registries.

The study team require the following information from NHS Digital:

- Prompt feed from SUS (Secondary Use Services) on hospital discharge

- Quarterly feed on HES (Hospital Episode Statistics) data

- Quarterly feeds on civil registration data (death certificate information)

- CHESS (COVID-19 Hospitalization in England Surveillance System) Data

- SGSS (COVID-19 Second Generation Surveillance System) Data

- GDPPR Data (General Practice Extraction Service [GPES] Data for Pandemic Planning and Research)

- Primary Care Medicines Data Set (disseminated when available)

The GDPPR and Primary Care Medicines data sets are requested for two key reasons. First they will be used to assess whether the effects of the treatments being tested in Recovery are different in people with certain medical conditions (e.g. lung disease or diabetes) or who were receiving particular medications (e.g. immunosupression) prior to joining the trial. To avoid burdening front-line staff, only limited data about other medical conditions is collected from the site staff at the point of enrollment and information about usual medications is not collected. Therefore these data sets will be invaluable in identifying relevant subgroups and will supplement information derived from the HES data regarding prior hospitalisations.

Second, information in these data sets after randomization will be used to determine the effects of the treatments on outcomes after discharge from hospital. For example, effects on the development on the development of long term lung problems, diabetes (particularly relevant to the steroid comparison), liver disease (particularly relevant to the lopinavir/ritonavir comparison) or on infections (particularly relevant to the tocolizumab comparison). It is understood that the GDPPR data is available for a limit time period only but will allow analysis of these and other outcomes following discharge from hospital.

Identifiable data is necessary for the purpose of data linkage and quality assurance of that linkage – the Data Sharing Agreement (DSA) requires measures are taken to ensure subsequent uses of the data will involve only data that has been appropriately pseudonymised.

Data requested includes a range of detailed information that is necessary for analysis and sub-group analysis of the effectiveness of the treatments – in particular to enable accounting for a range of biases and confounding factors relevant to the primary and secondary outcomes. Specifically:

• The trial has taken a pragmatic approach to recruitment and as a result may have included some patients who do not have confirmed Covid 19. For sub-group analysis and interpretation of the trial results, they need to know which patients have a confirmed positive Covid 19 test.

• Disease onset, dates for admission into hospital, ICU [Intensive Care Unit] and HDU [High Dependency Units] are important to track progression of the disease to understand the efficacy of treatment esp for those which inhibit viral replication.

• These data will allow the trial to assess the impact of trial drugs on the precise care and treatment delivered e.g the need for mechanical ventilation and any complications such as secondary infections.

• Risk factors and underlying conditions enable sub-group analysis of other factors that may have a bearing on the treatments being tested.

This study is supported by a grant to the University of Oxford from UK Research and Innovation/National Institute for Health Research (NIHR) and by core funding provided by NIHR Oxford Biomedical Research Centre, the Wellcome Trust, the Bill and Melinda Gates Foundation, Health Data Research UK (HDRUK), and the Medical Research Council Population Health Research Unit, and NIHR Clinical Trials Unit Support Funding.

The new trial has been classed as an Urgent Public Health Research Study. It is one of a round of projects to receive £10.5 million as part of the £20 million rapid research response funded by UK Research and Innovation, and by the Department of Health and Social Care through the National Institute for Health Research.

Independent External Validation of Results:

The RECOVERY trial is generating high profile results which change practice and there will be considerable external scrutiny of the trial processes. In order to provide an independent assessment of the validity of the research findings from the RECOVERY trial, the researchers at the University of Oxford have asked an independent statistician and associated team at the University of Bristol to reproduce the key findings using a copy of the trial analysis data-set.

This data-set does not include the full files provided to the RECOVERY team under this agreement, but does include individual participant data derived from those files, such as the date on which a particular medical procedure took place for a participant. In line with data minimization principles, only those fields required to replicate the analyses will be provided and direct identifiers (such as name, date of birth and NHS number) will be removed. The data will be transferred securely using password protection and encryption. This amendment adds the University of Bristol as a data processor so they can receive the data for this purpose. An appropriate controller processor agreement will be in place between the University of Oxford and the University of Bristol prior to any data transfer.

The University of Oxford are unable to provide external researchers with secure remote access to the data within the University of Oxford servers for technical reasons. This is something that will be considered in the future but has to be weighed against other demands on the department. The analysis team have shared individual participant data with external researchers before (outside of this agreement in terms of other research projects) and have a process for securely transferring encrypted files. A controller - processor agreement with Bristol University will be in place prior to data transfer. The duration of processing will be 12 months to enable the Professor to validate the results of future comparisons as needed.

The University of Bristol will access the data via remote connection to the University of Bristol Server - using University issued encrypted lap tops from home addresses. Data will be downloaded to these encrypted lap tops also.

The University of Oxford invited this Leading Independent Statistician to undertake this external validation due to previous academic and clinical experience. The statistician is leading the World Health Organization’s meta-analysis of dexamethasone trials and as such is an expert in the clinical context and methodological issues in relation to a trial such as RECOVERY. The Lead Statistician (from the team at Bristol) is a Professor of Medical Statistics and Epidemiology at University of Bristol and also the Director of Health Data Research UK South-West.

The task is sufficiently well defined to make the independent validation possible whist Oxford remain the data controller, and Bristol the data processor. The University of Oxford is asking the University of Bristol to validate the published results and the type of analyses to be done are already pre-specified (as they always would be in this type of trial) in the Statistical Analysis Plan (published by the RECOVERY team on the website prior to analyzing the results https://www.recoverytrial.net/results).

COVID-19 is an emerging pathogen which presents a significant threat to the population in terms of increased morbidity and mortality, particularly among vulnerable groups such as those with pre-existing disease.

The primary objective is to provide reliable estimates of the effect of study treatments on in-hospital death (with subsidiary analyses of cause of death and death at various timepoints following discharge).

The secondary objectives are to assess the effects of study treatments on duration of hospital stay; the need for (and duration of) ventilation; and the need for renal replacement therapy.

The interim trial results will be monitored by an independent Data Monitoring Committee (DMC). The most important task for the DMC will be to assess whether the randomised comparisons in the study have provided evidence on mortality that is strong enough (with a range of uncertainty around the results that is narrow enough) to affect national and global treatment strategies. In such a circumstance, the DMC will inform the Trial Steering Committee who will make the results available to the public and amend the trial arms accordingly.

The data requested will be used to evaluate the efficacy and safety of the study treatments and will help shape the public health response. The rapid feed will be used to ensure that the trial Data Monitoring Committee have complete, up-to-date information on the major outcomes in the trial on which to base their decisions. If they find compelling evidence of efficacy or safety then that arm may be stopped and – if effective – added to standard care across the NHS.

All outputs produced will be in the form of aggregated reports with small number suppression applied.

A paper was recently published in the New England Journal of Medicine entitled "Dexamethasone in Hospitalized Patients with Covid-19 — Preliminary Report" (17/07/2020) detailing the results.

By the end of July 2020, the RECOVERY trial had successfully established over 175 sites across the UK and recruited over 11,000 participants treated in hospital for COVID-19. The data linkage already established to received SUS+ and other data is providing important information to the Data Monitoring Committee on a weekly basis about patients' recovery (i.e. discharge from hospital), in-hospital death and procedures required. Provision of complete and reliable data to the DMC through May and early June 2020 is critical to allow robust assessment of the effects of the the trial treatments with a major contribution to these data expected from analysis of the routine health care data requested under this agreement.

On Thursday 4 June, in response to a request from the UK Medicines and Healthcare Products Regulatory Agency (MHRA), the independent Data Monitoring Committee conducted a further review of the data. The DMC recommended the chief investigators review the unblinded data on the hydroxychloroquine arm of the trial. The results included a total of 1542 patients randomised to hydroxychloroquine compared with 3132 patients randomised to usual care alone. There was no significant difference in the primary endpoint of 28-day mortality (25.7% hydroxychloroquine vs. 23.5% usual care; hazard ratio 1.11 [95% confidence interval 0.98-1.26]; p=0.10). There was also no evidence of beneficial effects on hospital stay duration or other outcomes. These results were released to the public and on the 15 July 2020, the U.S. Food and Drug Administration (FDA) revoked the emergency use authorization (EUA) that allowed for chloroquine phosphate and hydroxychloroquine sulfate donated to the Strategic National Stockpile to be used to treat certain hospitalized patients with COVID-19 when a clinical trial was unavailable, or participation in a clinical trial was not feasible. This result has major implications for countries around the world who were planning to scale manufacturing of these drugs in order to treat COVID patients.

On 8 June 2020, recruitment to the dexamethasone arm was halted since, in the view of the trial Steering Committee, sufficient patients had been enrolled to establish whether or not the drug had a meaningful benefit. On 16 June 2020, preliminary results were released. A total of 2104 patients were randomised to receive dexamethasone 6 mg once per day (either by mouth or by intravenous injection) for ten days and were compared with 4321 patients randomised to usual care alone. Among the patients who received usual care alone, 28-day mortality was highest in those who required ventilation (41%), intermediate in those patients who required oxygen only (25%), and lowest among those who did not require any respiratory intervention (13%). Dexamethasone reduced deaths by one-third in ventilated patients (rate ratio 0.65 [95% confidence interval 0.48 to 0.88]; p=0.0003) and by one fifth in other patients receiving oxygen only (0.80 [0.67 to 0.96]; p=0.0021). There was no benefit among those patients who did not require respiratory support (1.22 [0.86 to 1.75]; p=0.14). Based on these results, 1 death would be prevented by treatment of around 8 ventilated patients or around 25 patients requiring oxygen alone.

On the 16 June 2020, on the basis of these results, the MHRA issued an alert to health care providers in the UK recommending the use of dexamethasone in hospitalised patients with COVID who require oxygen or ventilation. In addition dexamethasone has also been added to the government’s parallel export list, which bans companies from buying medicines meant for UK patients and selling them on for a higher price in another country.

A recent analysis estimates that treatment with Dexamethasone may save between 4,000 and 27,000 lives in the UK by January 20201 (depending on the number of COVID-19 cases during Q3-4 2020) and potentially over half a million lives worldwide (https://www.medrxiv.org/content/10.1101/2020.07.29.20164269v1). Without the results from RECOVERY the effectiveness of this treatment would not be known.

A paper was recently published in the New England Journal of Medicine entitled "Dexamethasone in Hospitalized Patients with Covid-19 — Preliminary Report" (17/07/2020) detailing the results.

University of Oxford are requesting to use the NHS Digital Clinical Trials Service for access to data for a study entitled Randomised Evaluation of COVid-19 thERapY (RECOVERY).

In 2019 a novel coronavirus-induced disease (COVID-19) emerged in Wuhan, China. A month later the Chinese Center for Disease Control and Prevention identified a new betacoronavirus (SARS coronavirus 2, or SARS-CoV-2) as the aetiological (causing or contributing to the development of a disease or condition) agent. The clinical manifestations of COVID-19 range from asymptomatic infection or mild, transient symptoms to severe viral pneumonia with respiratory failure. As many patients do not progress to severe disease the overall case fatality rate per infected individual is low, but hospitals in areas with significant community transmission have experienced a major increase in the number of hospitalized pneumonia patients, and the frequency of severe disease in hospitalised patients can be as high as 30%. The progression from prodrome (an early symptom indicating the onset of a disease or illness - in this case usually fever, fatigue and cough) to severe pneumonia requiring oxygen support or mechanical ventilation often takes one to two weeks after the onset of symptoms. The kinetics of viral replication in the respiratory tract are not well characterized, but this relatively slow progression provides a potential time window in which antiviral therapies could influence the course of disease.

This study aims to compare several different treatments that may be useful for patients with COVID-19. These treatments have been recommended by the expert panel that advises the Chief Medical Officer in England.

From version 6.0 of the protocol, a factorial design will be used such that eligible and consenting participants may be randomised to one of the treatment arms in Randomisation A and, simultaneously, to one of the treatment arms in Randomisation B. Treatments and randiomisation are as follows:

Randomisation may be between the following treatment arms in randomisation A (although not all arms may be available at any one time):

1. Usual care plus hydroxychloroquine vs usual care alone (no longer recruiting)

Hydroxychloroquine, a derivative of chloroquine, has been used for many decades to treat malaria and rheumatological diseases. It has antiviral activity against SARS-CoV-2 in cell culture. This comparison has now reported preliminary results (see benefits section).

2. Usual care plus Low dose corticosteroids (dexamethasone) vs usual care alone (no longer recruiting)

Favourable immune response modulation by low-dose corticosteroids might help treat severe acute respiratory coronavirus infections, including COVID-19, SARS and MERS. This comparison has now reported preliminary results (see benefits section).

3. Usual care plus lopinavir-ritnovavir vs usual care along (still recruiting)

Lopinavir-Ritonavir: (commonly used to treat HIV). Lopinavir is a human immunodeficiency virus 1 (HIV-1) protease inhibitor, which is combined with ritonavir to increase lopinavir’s plasma half-life. Lopinavir-Ritonavir has shown activity against SARS and MERS CoVs.

4. • Azithromycin: Usual care plus azithromyin vs usual care along (still recruiting)

Azithromycin is a macrolide antibiotic. The macrolides inhibit the growth of bacteria and are often prescribed to treat rather common bacterial infections. Azithromycin has immunomodulatory properties that has shown benefit in inflammatory lung disease.

Randomisation part B: Simultaneously, eligible patients will be randomly allocated between the following treatment arms (provided there are no contraindications and the appropriate consent has been given):

• Usual care plus convalescent plasma vs usual care: Plasma from patients who have recovered from SARS-CoV-2 infection may contain antibodies that can bind to and neutralise the virus. Infusion of convalescent plasma containing high concentrations of neutralising antibody may accelerate clearance of the virus and clinical improvement. Convalescent plasma therapy had been given to at least 245 COVID-19 patients by the end of February 2020, and, according to a Chinese health official, 91 cases had shown improvement in clinical indicators and symptoms (http://www.xinhuanet.com/english/2020-02/28/c_138828177.htm). Five small case series (26 patients in total) have been published that report the use of convalescent plasma in people with COVID-19 infection.59-63 These studies have reported clinical and radiological improvements after treatment with convalescent plasma. However, these small uncontrolled studies have significant flaws and the reported effects are unreliable. Convalescent plasma is currently being tested in the REMAP-CAP trial (a seperate trial) among patients on intensive care units.

A subset of participants with progressive COVID-19 (as evidenced by hypoxia and an inflammatory state) may undergo an optional second randomisation between the following treatment arms:

Usual care plus Tocilizumab vs usual care

Tocilizumab: (an immunosuppressive drug, mainly for the treatment of rheumatoid arthritis) or control (both in addition to the treatment assigned at the first randomisation).Tocilizumab is an interleukin-6 (IL-6) receptor antibody, which blocks a component of the immune response that may drive progression to Acute Respiratory Distress Syndrome.

Other arms can be added if evidence emerges that there are suitable candidate therapeutics. Conversely, in some patient populations, not all trial arms are appropriate (e.g. due to contraindications based on co-morbid conditions or concomitant medication); in some hospitals, not all treatment arms will be available (e.g. due to manufacturing and supply shortages); and at some times, not all treatment arms will be active (e.g. due to lack of relevant approvals and contractual agreements). The Trial Steering Committee may elect to pause one or more of the arms in order to increase trial efficiency during a fluctuating epidemic. Therefore,other treatments may be added to the protocol as time goes on and more information is gathered. This will not impact or change the level of data that is required from NHS Digital - as the cohort in the trial remains the same, regardless of what drug is being trialed.

Data from the trial will be regularly reviewed so that any effective treatment can be identified quickly and made available to all patients. The RECOVERY Trial team will constantly review information on new drugs and include promising ones in the trial.

Patients are eligible for the study if all of the following are true:

(i) Hospitalised

(ii) SARS-CoV-2 infection (clinically suspected or laboratory confirmed)

(iii) No medical history that might, in the opinion of the attending clinician, put the patient at significant risk if he/she were to participate in the trial

In addition, if the attending clinician believes that there is a specific contra-indication to one of the active drug treatment arms or that the patient should definitely be receiving one of the active drug treatment arms then that arm will not be available for randomisation for that patient. For patients who lack capacity, an advanced directive or behaviour that clearly indicates that they would not wish to participate in the trial would be considered sufficient reason to exclude them from the trial.

Informed consent should be obtained from each patient 16 years and over before enrolment into the study. However, if the patient lacks capacity to give consent due to the severity of their medical condition (e.g. acute respiratory failure or need for immediate ventilation) or prior disease, then consent may be obtained from a relative acting as the patient’s legally designated representative or independent doctor. Further consent will then be sought with the patient if they recover sufficiently. For children aged <16 years old consent will be sought from their parents or legal guardian. Where possible, children aged between 10-15 years old will also be asked for assent. Children aged ≥16 years old will be asked for consent as for adults. Witnessed consent may be obtained over the telephone or web video link if hospital visiting rules or parental infection mean a parent/guardian cannot be physically present.

Where Local Clinical Centre's (LCC's) plan to recruit children the principal investigator will co-opt support from a local paediatrician and/or neonatologists to oversee the management of children and infants in the trial.

The anticipated scale of the epidemic is such that hospitals, and particularly intensive care facilities, may be massively overstretched. Under some models of pandemic spread, up to 50% of the adult population may fall sick over a period of 8-12 weeks, of whom around 10% may require hospitalisation. In this situation, even treatments with only a moderate impact on survival or on hospital resources could be worthwhile. Therefore, the focus of the COVID-19 Core Protocol is the impact of candidate treatments on mortality and on the need for hospitalisation or ventilation. Critically, the trial (and subsequent data collection) is designed to minimise the burden on front-line hospital staff working within an overstretched care system during a major epidemic. Eligibility criteria are therefore simple and trial processes (including paperwork) are minimised.

For each pairwise comparison with the ‘no additional treatment’ arm, the primary objective is to provide reliable estimates of the effect of study treatments on all-cause mortality at 28 days after first randomisation (with subsidiary analyses of cause of death and of death at various timepoints following discharge).

The secondary objectives are to assess the effects of study treatments on duration of hospital stay; the need for (and duration of) ventilation; and, among patients not on ventilation at baseline, the composite endpoint of death or need for mechanical ventilation or ECMO (extracorporeal membrane oxygenation)

Other objectives include the assessment of the effects of study treatments on the need for renal replacement therapy and new major cardiac arrhythmias.

Study outcomes will be assessed based on data recorded up to 28 days and up to 6 months after the main randomization.

Data from routine healthcare records (including linkage to medical databases held by organisations such as NHS Digital) and from relevant research studies (such as UK Biobank and Genomics England) will allow subsidiary analyses of the effect of the study treatments on particular non-fatal events (e.g. ascertained through linkage to Hospital Episode Statistics), the influence of pre-existing major co-morbidity (e.g. diabetes, heart disease, lung disease, hepatic insufficiency, severe depression, severe kidney impairment, immunosuppression), and longer-term outcomes (e.g. 6 month survival) as well as in particular sub-categories of patient (e.g. by genotype).

Follow-up information is to be collected on all study participants, irrespective of whether or not they complete the scheduled course of allocated study treatment. Study staff will seek Follow-up information through various means including medical staff, reviewing information from medical notes, routine healthcare systems, and registries.

The study team require the following information from NHS Digital:

- Prompt feed from SUS on hospital discharge

- Quarterly feed on HES data

- Quarterly feeds on civil registration data (death certificate information)

- CHESS (COVID-19 Hospitalization in England Surveillance System) Data

- SGSS (COVID-19 Second Generation Surveillance System) Data

- GDPPR Data (GPES Data for Pandemic Planning and Research

- Primary Care Medicines Data Set (disseminated when available)

The GDPPR and Primary Care Medicines data sets are requested for two key reasons. First they will be used to assess whether the effects of the treatments being tested in Recovery are different in people with certain medical conditions (e.g. lung disease or diabetes) or who were receiving particular medications (e.g. immunosupression) prior to joining the trial. To avoid burdening front-line staff, only limited data about other medical conditions is collected from the site staff at the point of enrollment and information about usual medications is not collected. Therefore these data sets will be invaluable in identifying relevant subgroups and will supplement information derived from the HES data regarding prior hospitalisations.

Second, information in these data sets after randomization will be used to determine the effects of the treatments on outcomes after discharge from hospital. For example, effects on the development on the development of long term lung problems, diabetes (particularly relevant to the steroid comparison), liver disease (particularly relevant to the lopinavir/ritonavir comparison) or on infections (particularly relevant to the tocolizumab comparison). It is understood that the GDPPR data is available for a limit time period only but will allow analysis of these and other outcomes following discharge from hospital.

Identifiable data is necessary for the purpose of data linkage and quality assurance of that linkage – the DSA requires measures are taken to ensure subsequent uses of the data will involve only data that has been appropriately pseudonymised.

Data requested includes a range of detailed information that is necessary for analysis and sub-group analysis of the effectiveness of the treatments – in particular to enable accounting for a range of biases and confounding factors relevant to the primary and secondary outcomes. Specifically:

• The trial has taken a pragmatic approach to recruitment and as a result may have included some patients who do not have confirmed Covid 19. For sub-group analysis they need to know which patients have a confirmed positive Covid 19 test.

• Disease onset, dates for admission into hospital, ICU and HDU are important to track progression of the disease to understand the efficacy of treatment esp for those which inhibit viral replication.

• These data will allow the trial to assess the impact of trial drugs on the precise care and treatment delivered eg the need for mechanical ventilation and any complications such as secondary infections.

• Risk factors and underlying conditions enable sub-group analysis of other factors that may have a bearing on the treatments being tested.

This study is supported by a grant to the University of Oxford from UK Research and Innovation/National Institute for Health Research (NIHR) and by core funding provided by NIHR Oxford Biomedical Research Centre, the Wellcome Trust, the Bill and Melinda Gates Foundation, Health Data Research UK, and the Medical Research Council Population Health Research Unit, and NIHR Clinical Trials Unit Support Funding.

The new trial has been classed as an Urgent Public Health Research Study. It is one of a round of projects to receive £10.5 million as part of the £20 million rapid research response funded by UK Research and Innovation, and by the Department of Health and Social Care through the National Institute for Health Research.

COVID-19 is an emerging pathogen which presents a significant threat to the population in terms of increased morbidity and mortality, particularly among vulnerable groups such as those with pre-existing disease.

The primary objective is to provide reliable estimates of the effect of study treatments on in-hospital death (with subsidiary analyses of cause of death and death at various timepoints following discharge).

The secondary objectives are to assess the effects of study treatments on duration of hospital stay; the need for (and duration of) ventilation; and the need for renal replacement therapy.

The interim trial results will be monitored by an independent Data Monitoring Committee (DMC). The most important task for the DMC will be to assess whether the randomised comparisons in the study have provided evidence on mortality that is strong enough (with a range of uncertainty around the results that is narrow enough) to affect national and global treatment strategies. In such a circumstance, the DMC will inform the Trial Steering Committee who will make the results available to the public and amend the trial arms accordingly.

The data requested will be used to evaluate the efficacy and safety of the study treatments and will help shape the public health response. The rapid feed will be used to ensure that the trial Data Monitoring Committee have complete, up-to-date information on the major outcomes in the trial on which to base their decisions. If they find compelling evidence of efficacy or safety then that arm may be stopped and – if effective – added to standard care across the NHS.

All outputs produced will be in the form of aggregated reports with small number suppression applied.

By the end of May 2020, the RECOVERY trial had successfully established over 175 sites across the UK and recruited over 11,000 participants treated in hospital for COVID-19. The data linkage already established to received SUS+ and other data is providing important information to the Data Monitoring Committee on a weekly basis about patients' recovery (i.e. discharge from hospital), in-hospital death and procedures required. Provision of complete and reliable data to the DMC through May and early June 2020 is critical to allow robust assessment of the effects of the the trial treatments with a major contribution to these data expected from analysis of the routine health care data requested under this agreement.

On Thursday 4 June, in response to a request from the UK Medicines and Healthcare Products Regulatory Agency (MHRA), the independent Data Monitoring Committee conducted a further review of the data. The DMC recommended the chief investigators review the unblinded data on the hydroxychloroquine arm of the trial. The results included a total of 1542 patients randomised to hydroxychloroquine compared with 3132 patients randomised to usual care alone. There was no significant difference in the primary endpoint of 28-day mortality (25.7% hydroxychloroquine vs. 23.5% usual care; hazard ratio 1.11 [95% confidence interval 0.98-1.26]; p=0.10). There was also no evidence of beneficial effects on hospital stay duration or other outcomes. These results were released to the public and on the 15 July 2020, the U.S. Food and Drug Administration (FDA) revoked the emergency use authorization (EUA) that allowed for chloroquine phosphate and hydroxychloroquine sulfate donated to the Strategic National Stockpile to be used to treat certain hospitalized patients with COVID-19 when a clinical trial was unavailable, or participation in a clinical trial was not feasible. This result has major implications for countries around the world who were planning to scale manufacturing of these drugs in order to treat COVID patients.

On 8 June 2020, recruitment to the dexamethasone arm was halted since, in the view of the trial Steering Committee, sufficient patients had been enrolled to establish whether or not the drug had a meaningful benefit. On 16 June 2020, preliminary results were released. A total of 2104 patients were randomised to receive dexamethasone 6 mg once per day (either by mouth or by intravenous injection) for ten days and were compared with 4321 patients randomised to usual care alone. Among the patients who received usual care alone, 28-day mortality was highest in those who required ventilation (41%), intermediate in those patients who required oxygen only (25%), and lowest among those who did not require any respiratory intervention (13%). Dexamethasone reduced deaths by one-third in ventilated patients (rate ratio 0.65 [95% confidence interval 0.48 to 0.88]; p=0.0003) and by one fifth in other patients receiving oxygen only (0.80 [0.67 to 0.96]; p=0.0021). There was no benefit among those patients who did not require respiratory support (1.22 [0.86 to 1.75]; p=0.14). Based on these results, 1 death would be prevented by treatment of around 8 ventilated patients or around 25 patients requiring oxygen alone.

On the 16 June 2020, on the basis of these results, the MHRA issued an alert to health care providers in the UK recommending the use of dexamethasone in hospitalised patients with COVID who require oxygen or ventilation. In addition dexamethasone has also been added to the government’s parallel export list, which bans companies from buying medicines meant for UK patients and selling them on for a higher price in another country.

This is a new application from the University of Oxford. University of Oxford request to use the NHS Digital Clinical Trials Service for access to data for a study entitled Randomised Evaluation of COVid-19 thERapY (RECOVERY).

In 2019 a novel coronavirus-induced disease (COVID-19) emerged in Wuhan, China. A month later the Chinese Center for Disease Control and Prevention identified a new betacoronavirus (SARS coronavirus 2, or SARS-CoV-2) as the aetiological (causing or contributing to the development of a disease or condition) agent. The clinical manifestations of COVID-19 range from asymptomatic infection or mild, transient symptoms to severe viral pneumonia with respiratory failure. As many patients do not progress to severe disease the overall case fatality rate per infected individual is low, but hospitals in areas with significant community transmission have experienced a major increase in the number of hospitalized pneumonia patients, and the frequency of severe disease in hospitalised patients can be as high as 30%. The progression from prodrome (an early symptom indicating the onset of a disease or illness - in this case usually fever, fatigue and cough) to severe pneumonia requiring oxygen support or mechanical ventilation often takes one to two weeks after the onset of symptoms. The kinetics of viral replication in the respiratory tract are not well characterized, but this relatively slow progression provides a potential time window in which antiviral therapies could influence the course of disease.

In early 2020, as the protocol for this trial was being developed, there were no approved treatments for COVID-19, a disease induced by the novel coronavirus SARS-CoV-2 that emerged in China in late 2019. The UK New and Emerging Respiratory Virus Threats Advisory Group (NERVTAG) advised that several possible treatments should be evaluated, including (but not limited to) Lopinavir-Ritonavir, Interferon β, and low-dose corticosteroids. These groups also advised that other treatments will soon emerge that require evaluation. A World Health Organization (WHO) expert group issued broadly similar advice. This trial allows reliable assessment of the effects of multiple different treatments (including re-purposed and novel drugs) on major outcomes in COVID-19.

This study aims to compare several different treatments that may be useful for patients with COVID-19. These treatments have been recommended by the expert panel that advises the Chief Medical Officer in England. Some are tablets and some are injections. Although these treatments show promise, nobody knows if any of them will turn out to be more effective in helping patients recover than the usual standard of care at hospitals (which all patients will receive).

The treatments, given in addition to the usual care at hospital, are: Lopinavir-Ritonavir (commonly used to treat HIV); dexathasone (a type of steroid, which is used in a range of conditions typically to reduce inflammation) and Hydroxychloroquine. Hydroxychloroquine, a derivative of chloroquine, has been used for many decades to treat malaria and rheumatological diseases. It has antiviral activity against SARS-CoV-2 in cell culture. As of 07/04/2020 - Azithromycin has been added as an arm to the trial (replacing a previous drug type called interferon beta). Azithromycin is a macrolide antibiotic. The macrolides inhibit the growth of bacteria and are often prescribed to treat rather common bacterial infections. Azithromycin has immunomodulatory properties that has shown benefit in inflammatory lung disease.

As of 16/04/2020 - a second randomisation arm has been added to the protocol.

This is for participants who meet certain simple physiological and inflammatory criteria - for example those in the cohort receiving oxygen therapy or oxygen saturations <92%) who can be randomised for a second time to receive either tocilizumab (an immunosuppressive drug, mainly for the treatment of rheumatoid arthritis) or control (both in addition to the treatment assigned at the first randomisation)

Other arms can be added if evidence emerges that there are suitable candidate therapeutics. Conversely, in some patient

populations, not all trial arms are appropriate (e.g. due to contraindications based on co-morbid conditions or concomitant medication); in some hospitals, not all treatment arms will be available (e.g. due to manufacturing and supply shortages); and at some times, not all treatment arms will be active (e.g. due to lack of relevant approvals and contractual agreements). Therefore,other treatments may be added to the protocol as time goes on and more information is gathered. This will not impact or change the level of data that is required from NHS Digital - as the cohort in the trial remains the same, regardless of what drug is being trialed.

Data from the trial will be regularly reviewed so that any effective treatment can be identified quickly and made available to all patients. The RECOVERY Trial team will constantly review information on new drugs and include promising ones in the trial.

Patients are eligible for the study if all of the following are true:

(i) Hospitalised

(ii) SARS-CoV-2 infection (clinically suspected or laboratory confirmed)

(iii) No medical history that might, in the opinion of the attending clinician, put the patient at significant risk if he/she were to participate in the trial

In addition, if the attending clinician believes that there is a specific contra-indication to one of the active drug treatment arms or that the patient should definitely be receiving one of the active drug treatment arms then that arm will not be available for randomisation for that patient. For patients who lack capacity, an advanced directive or behaviour that clearly indicates that they would not wish to participate in the trial would be considered sufficient reason to exclude them from the trial.

Informed consent should be obtained from each patient 16 years and over before enrolment into the study. However, if the patient lacks capacity to give consent due to the severity of their medical condition (e.g. acute respiratory failure or need for immediate ventilation) or prior disease, then consent may be obtained from a relative acting as the patient’s legally designated representative or independent doctor. Further consent will then be sought with the patient if they recover sufficiently. For children aged <16 years old consent will be sought from their parents or legal guardian. Where possible, children aged between 10-15 years old will also be asked for assent. Children aged ≥16 years old will be asked for consent as for adults. Witnessed consent may be obtained over the telephone or web video link if hospital visiting rules or parental infection mean a parent/guardian cannot be physically present.

Where Local Clinical Centre's (LCC's) plan to recruit children the principal investigator will co-opt support from a local paediatrician and/or neonatologists to oversee the management of children and infants in the trial.

The anticipated scale of the epidemic is such that hospitals, and particularly intensive care facilities, may be massively overstretched. Under some models of pandemic spread, up to 50% of the adult population may fall sick over a period of 8-12 weeks, of whom around 10% may require hospitalisation. In this situation, even treatments with only a moderate impact on survival or on hospital resources could be worthwhile. Therefore, the focus of the COVID-19 Core Protocol is the impact of candidate treatments on mortality and on the need for hospitalisation or ventilation. Critically, the trial (and subsequent data collection) is designed to minimise the burden on front-line hospital staff working within an overstretched care system during a major epidemic. Eligibility criteria are therefore simple and trial processes (including paperwork) are minimised.

The primary objective is to provide reliable estimates of the effect of study treatments on in-hospital death (with subsidiary analyses of cause of death and death at various timepoints following discharge).

The secondary objectives are to assess the effects of study treatments on duration of hospital stay; the need for (and duration of) ventilation; and the need for renal replacement therapy.

Data from routine healthcare records (including linkage to medical databases held by organisations such as NHS Digital) and from relevant research studies (such as UK Biobank and Genomics England) will allow subsidiary analyses of the effect of the study treatments on particular non-fatal events (e.g. ascertained through linkage to Hospital Episode Statistics), the influence of pre-existing major co-morbidity (e.g. diabetes, heart disease, lung disease, hepatic insufficiency, severe depression, severe kidney impairment, immunosuppression), and longer-term outcomes (e.g. 6 month survival) as well as in particular sub-categories of patient (e.g. by genotype).

Follow-up information is to be collected on all study participants, irrespective of whether or not they complete the scheduled course of allocated study treatment. Study staff will seek Follow-up information through various means including medical staff, reviewing information from medical notes, routine healthcare systems, and registries.

The study team require the following information from NHS Digital:

- Prompt feed from SUS on hospital discharge

- Quarterly feed on HES data

- Quarterly feeds on civil registration data (death certificate information)

- CHESS (COVID-19 Hospitalization in England Surveillance System) Data

- SGSS (COVID-19 Second Generation Surveillance System) Data

Identifiable data is necessary for the purpose of data linkage and quality assurance of that linkage – the DSA requires measures are taken to ensure subsequent uses of the data will involve only data that has been appropriately pseudonymised.

Data requested includes a range of detailed information that is necessary for analysis and sub-group analysis of the effectiveness of the treatments – in particular to enable accounting for a range of biases and confounding factors relevant to the primary and secondary outcomes. Specifically:

• The trial has taken a pragmatic approach to recruitment and as a result may have included some patients who do not have confirmed Covid 19. For sub-group analysis they need to know which patients have a confirmed positive Covid 19 test.

• Disease onset, dates for admission into hospital, ICU and HDU are important to track progression of the disease to understand the efficacy of treatment esp for those which inhibit viral replication.

• These data will allow the trial to assess the impact of trial drugs on the precise care and treatment delivered eg the need for mechanical ventilation and any complications such as secondary infections.

• Risk factors and underlying conditions enable sub-group analysis of other factors that may have a bearing on the treatments being tested.

This study is supported by a grant to the University of Oxford from UK Research and Innovation/National Institute for Health Research (NIHR) and by core funding provided by NIHR Oxford Biomedical Research Centre, the Wellcome Trust, the Bill and Melinda Gates Foundation, Health Data Research UK, and the Medical Research Council Population Health Research Unit, and NIHR Clinical Trials Unit Support Funding.

The new trial has been classed as an Urgent Public Health Research Study. It is one of a round of projects to receive £10.5 million as part of the £20 million rapid research response funded by UK Research and Innovation, and by the Department of Health and Social Care through the National Institute for Health Research.

COVID-19 is an emerging pathogen which presents a significant threat to the population in terms of increased morbidity and mortality, particularly among vulnerable groups such as those with pre-existing disease.

The primary objective is to provide reliable estimates of the effect of study treatments on in-hospital death (with subsidiary analyses of cause of death and death at various timepoints following discharge).

The secondary objectives are to assess the effects of study treatments on duration of hospital stay; the need for (and duration of) ventilation; and the need for renal replacement therapy.

The interim trial results will be monitored by an independent Data Monitoring Committee (DMC). The most important task for the DMC will be to assess whether the randomised comparisons in the study have provided evidence on mortality that is strong enough (with a range of uncertainty around the results that is narrow enough) to affect national and global treatment strategies. In such a circumstance, the DMC will inform the Trial Steering Committee who will make the results available to the public and amend the trial arms accordingly.

The data requested will be used to evaluate the efficacy and safety of the study treatments and will help shape the public health response. The rapid feed will be used to ensure that the trial Data Monitoring Committee have complete, up-to-date information on the major outcomes in the trial on which to base their decisions. If they find compelling evidence of efficacy or safety then that arm may be stopped and – if effective – added to standard care across the NHS.

All outputs produced will be in the form of aggregated reports with small number suppression applied.

By the end of April 2020, the Recovery trial had successfully established over 160 sites across the UK and recruited over 8000 participants treated in hospital for Covid-19. The data linkage already established to received SUS+ and other data is providing important information to the Data Monitoring Committee on a weekly basis about patients' recovery (i.e. discharge from hospital), in-hospital death and procedures required. Provision of complete and reliable data to the DMC through May and early June 2020 is critical to allow robust assessment of the effects of the the trial treatments with a major contribution to these data expected from analysis of the routine health care data requested under this agreement.

This is a new application from the University of Oxford. University of Oxford request to use the NHS Digital Clinical Trials Service for access to data for a study entitled Randomised Evaluation of COVid-19 thERapY (RECOVERY).

In 2019 a novel coronavirus-induced disease (COVID-19) emerged in Wuhan, China. A month later the Chinese Center for Disease Control and Prevention identified a new betacoronavirus (SARS coronavirus 2, or SARS-CoV-2) as the aetiological (causing or contributing to the development of a disease or condition) agent. The clinical manifestations of COVID-19 range from asymptomatic infection or mild, transient symptoms to severe viral pneumonia with respiratory failure. As many patients do not progress to severe disease the overall case fatality rate per infected individual is low, but hospitals in areas with significant community transmission have experienced a major increase in the number of hospitalized pneumonia patients, and the frequency of severe disease in hospitalised patients can be as high as 30%. The progression from prodrome (an early symptom indicating the onset of a disease or illness - in this case usually fever, fatigue and cough) to severe pneumonia requiring oxygen support or mechanical ventilation often takes one to two weeks after the onset of symptoms. The kinetics of viral replication in the respiratory tract are not well characterized, but this relatively slow progression provides a potential time window in which antiviral therapies could influence the course of disease.

In early 2020, as the protocol for this trial was being developed, there were no approved treatments for COVID-19, a disease induced by the novel coronavirus SARS-CoV-2 that emerged in China in late 2019. The UK New and Emerging Respiratory Virus Threats Advisory Group (NERVTAG) advised that several possible treatments should be evaluated, including (but not limited to) Lopinavir-Ritonavir, Interferon β, and low-dose corticosteroids. These groups also advised that other treatments will soon emerge that require evaluation. A World Health Organization (WHO) expert group issued broadly similar advice. This trial allows reliable assessment of the effects of multiple different treatments (including re-purposed and novel drugs) on major outcomes in COVID-19.

This study aims to compare several different treatments that may be useful for patients with COVID-19. These treatments have been recommended by the expert panel that advises the Chief Medical Officer in England. Some are tablets and some are injections. Although these treatments show promise, nobody knows if any of them will turn out to be more effective in helping patients recover than the usual standard of care at hospitals (which all patients will receive).

The treatments, given in addition to the usual care at hospital, are: Lopinavir-Ritonavir (commonly used to treat HIV); dexathasone (a type of steroid, which is used in a range of conditions typically to reduce inflammation) and Hydroxychloroquine. Hydroxychloroquine, a derivative of chloroquine, has been used for many decades to treat malaria and rheumatological diseases. It has antiviral activity against SARS-CoV-2 in cell culture. As of 07/04/2020 - Azithromycin has been added as an arm to the trial (replacing a previous drug type called interferon beta). Azithromycin is a macrolide antibiotic. The macrolides inhibit the growth of bacteria and are often prescribed to treat rather common bacterial infections. Azithromycin has immunomodulatory properties that has shown benefit in inflammatory lung disease.

As of 16/04/2020 - a second randomisation arm has been added to the protocol.

This is for participants who meet certain simple physiological and inflammatory criteria - for example those in the cohort receiving oxygen therapy or oxygen saturations <92%) who can be randomised for a second time to receive either tocilizumab (an immunosuppressive drug, mainly for the treatment of rheumatoid arthritis) or control (both in addition to the treatment assigned at the first randomisation)

Other arms can be added if evidence emerges that there are suitable candidate therapeutics. Conversely, in some patient

populations, not all trial arms are appropriate (e.g. due to contraindications based on co-morbid conditions or concomitant medication); in some hospitals, not all treatment arms will be available (e.g. due to manufacturing and supply shortages); and at some times, not all treatment arms will be active (e.g. due to lack of relevant approvals and contractual agreements). Therefore,other treatments may be added to the protocol as time goes on and more information is gathered. This will not impact or change the level of data that is required from NHS Digital - as the cohort in the trial remains the same, regardless of what drug is being trialed.

Data from the trial will be regularly reviewed so that any effective treatment can be identified quickly and made available to all patients. The RECOVERY Trial team will constantly review information on new drugs and include promising ones in the trial.

Patients are eligible for the study if all of the following are true:

(i) Aged at least 18 years

(ii) Hospitalised

(iii) Proven SARS-CoV-2 infection.

(iv) Suspected SARS-Cov-2 infection. In general, SARS-CoV-2 infection should be suspected when a patient presents with

(i) typical symptoms (e.g. influenza-like illness with fever and muscle pain, or respiratory illness with cough and shortness of breath); and (ii) compatible chest X-ray findings (consolidation or ground-glass shadowing); and (iii) alternative causes have been considered unlikely or excluded (e.g. heart failure, influenza). However, the diagnosis remains a clinical one based on the opinion of the managing doctor.

(v) No medical history that might, in the opinion of the attending clinician, put the patient at significant risk if he/she were to participate in the trial

The anticipated scale of the epidemic is such that hospitals, and particularly intensive care facilities, may be massively overstretched. Under some models of pandemic spread, up to 50% of the adult population may fall sick over a period of 8-12 weeks, of whom around 10% may require hospitalisation. In this situation, even treatments with only a moderate impact on survival or on hospital resources could be worthwhile. Therefore, the focus of the COVID-19 Core Protocol is the impact of candidate treatments on mortality and on the need for hospitalisation or ventilation. Critically, the trial (and subsequent data collection) is designed to minimise the burden on front-line hospital staff working within an overstretched care system during a major epidemic. Eligibility criteria are therefore simple and trial processes (including paperwork) are minimised.

The primary objective is to provide reliable estimates of the effect of study treatments on in-hospital death (with subsidiary analyses of cause of death and death at various timepoints following discharge).

The secondary objectives are to assess the effects of study treatments on duration of hospital stay; the need for (and duration of) ventilation; and the need for renal replacement therapy.

Data from routine healthcare records (including linkage to medical databases held by organisations such as NHS Digital) and from relevant research studies (such as UK Biobank and Genomics England) will allow subsidiary analyses of the effect of the study treatments on particular non-fatal events (e.g. ascertained through linkage to Hospital Episode Statistics), the influence of pre-existing major co-morbidity (e.g. diabetes, heart disease, lung disease, hepatic insufficiency, severe depression, severe kidney impairment, immunosuppression), and longer-term outcomes (e.g. 6 month survival) as well as in particular sub-categories of patient (e.g. by genotype).

Follow-up information is to be collected on all study participants, irrespective of whether or not they complete the scheduled course of allocated study treatment. Study staff will seek Follow-up information through various means including medical staff, reviewing information from medical notes, routine healthcare systems, and registries.

The study team require the following information from NHS Digital:

- Prompt feed from SUS on hospital discharge

- Quarterly feed on HES data

- Quarterly feeds on civil registration data (death certificate information)

This study is supported by a grant to the University of Oxford from UK Research and Innovation/National Institute for Health Research (NIHR) and by core funding provided by NIHR Oxford Biomedical Research Centre, the Wellcome Trust, the Bill and Melinda Gates Foundation, Health Data Research UK, and the Medical Research Council Population Health Research Unit, and NIHR Clinical Trials Unit Support Funding.

The new trial has been classed as an Urgent Public Health Research Study. It is one of a round of projects to receive £10.5 million as part of the £20 million rapid research response funded by UK Research and Innovation, and by the Department of Health and Social Care through the National Institute for Health Research.

COVID-19 is an emerging pathogen which presents a significant threat to the population in terms of increased morbidity and mortality, particularly among vulnerable groups such as those with pre-existing disease.

The primary objective is to provide reliable estimates of the effect of study treatments on in-hospital death (with subsidiary analyses of cause of death and death at various timepoints following discharge).

The secondary objectives are to assess the effects of study treatments on duration of hospital stay; the need for (and duration of) ventilation; and the need for renal replacement therapy.

The interim trial results will be monitored by an independent Data Monitoring Committee (DMC). The most important task for the DMC will be to assess whether the randomised comparisons in the study have provided evidence on mortality that is strong enough (with a range of uncertainty around the results that is narrow enough) to affect national and global treatment strategies. In such a circumstance, the DMC will inform the Trial Steering Committee who will make the results available to the public and amend the trial arms accordingly.

The data requested will be used to evaluate the efficacy and safety of the study treatments and will help shape the public health response. The rapid feed will be used to ensure that the trial Data Monitoring Committee have complete, up-to-date information on the major outcomes in the trial on which to base their decisions. If they find compelling evidence of efficacy or safety then that arm may be stopped and – if effective – added to standard care across the NHS.

All outputs produced will be in the form of aggregated reports with small number suppression applied.

This is a new application from the University of Oxford. University of Oxford request to use the NHS Digital Clinical Trials Service for access to data for a study entitled Randomised Evaluation of COVid-19 thERapY (RECOVERY).

In 2019 a novel coronavirus-induced disease (COVID-19) emerged in Wuhan, China. A month later the Chinese Center for Disease Control and Prevention identified a new betacoronavirus (SARS coronavirus 2, or SARS-CoV-2) as the aetiological (causing or contributing to the development of a disease or condition) agent. The clinical manifestations of COVID-19 range from asymptomatic infection or mild, transient symptoms to severe viral pneumonia with respiratory failure. As many patients do not progress to severe disease the overall case fatality rate per infected individual is low, but hospitals in areas with significant community transmission have experienced a major increase in the number of hospitalized pneumonia patients, and the frequency of severe disease in hospitalised patients can be as high as 30%. The progression from prodrome (an early symptom indicating the onset of a disease or illness - in this case usually fever, fatigue and cough) to severe pneumonia requiring oxygen support or mechanical ventilation often takes one to two weeks after the onset of symptoms. The kinetics of viral replication in the respiratory tract are not well characterized, but this relatively slow progression provides a potential time window in which antiviral therapies could influence the course of disease.

In early 2020, as the protocol for this trial was being developed, there were no approved treatments for COVID-19, a disease induced by the novel coronavirus SARS-CoV-2 that emerged in China in late 2019. The UK New and Emerging Respiratory Virus Threats Advisory Group (NERVTAG) advised that several possible treatments should be evaluated, including (but not limited to) Lopinavir-Ritonavir, Interferon β, and low-dose corticosteroids. These groups also advised that other treatments will soon emerge that require evaluation. A World Health Organization (WHO) expert group issued broadly similar advice. This trial allows reliable assessment of the effects of multiple different treatments (including re-purposed and novel drugs) on major outcomes in COVID-19.

This study aims to compare several different treatments that may be useful for patients with COVID-19. These treatments have been recommended by the expert panel that advises the Chief Medical Officer in England. Some are tablets and some are injections. Although these treatments show promise, nobody knows if any of them will turn out to be more effective in helping patients recover than the usual standard of care at hospitals (which all patients will receive). The treatments, given in addition to the usual care at hospital, are: Lopinavir-Ritonavir (commonly used to treat HIV); inhaled Interferon (usually given by injection to treat multiple sclerosis, hepatitis C, and some blood disorders); dexathasone (a type of steroid, which is used in a range of conditions typically to reduce inflammation) and Hydroxychloroquine: Hydroxychloroquine, a derivative of chloroquine, has been used for many decades to treat malaria and rheumatological diseases. It has antiviral activity against SARS-CoV-2 in cell culture.

Other arms can be added if evidence emerges that there are suitable candidate therapeutics. Conversely, in some patient

populations, not all trial arms are appropriate (e.g. due to contraindications based on co-morbid conditions or concomitant medication); in some hospitals, not all treatment arms will be available (e.g. due to manufacturing and supply shortages); and at some times, not all treatment arms will be active (e.g. due to lack of relevant approvals and contractual agreements). Therefore,other treatments may be added to the protocol as time goes on and more information is gathered. This will not impact or change the level of data that is required from NHS Digital - as the cohort in the trial remains the same, regardless of what drug is being trialed.

Data from the trial will be regularly reviewed so that any effective treatment can be identified quickly and made available to all patients. The RECOVERY Trial team will constantly review information on new drugs and include promising ones in the trial.

Patients are eligible for the study if all of the following are true:

(i) Aged at least 18 years

(ii) Hospitalised

(iii) SARS-CoV-2 infection

(iv) No medical history that might, in the opinion of the attending clinician, put the patient at significant risk if he/she were to participate in the trial

The anticipated scale of the epidemic is such that hospitals, and particularly intensive care facilities, may be massively overstretched. Under some models of pandemic spread, up to 50% of the adult population may fall sick over a period of 8-12 weeks, of whom around 10% may require hospitalisation. In this situation, even treatments with only a moderate impact on survival or on hospital resources could be worthwhile. Therefore, the focus of the COVID-19 Core Protocol is the impact of candidate treatments on mortality and on the need for hospitalisation or ventilation. Critically, the trial (and subsequent data collection) is designed to minimise the burden on front-line hospital staff working within an overstretched care system during a major epidemic. Eligibility criteria are therefore simple and trial processes (including paperwork) are minimised.

The primary objective is to provide reliable estimates of the effect of study treatments on in-hospital death (with subsidiary analyses of cause of death and death at various timepoints following discharge).

The secondary objectives are to assess the effects of study treatments on duration of hospital stay; the need for (and duration of) ventilation; and the need for renal replacement therapy.

Data from routine healthcare records (including linkage to medical databases held by organisations such as NHS Digital) and from relevant research studies (such as UK Biobank and Genomics England) will allow subsidiary analyses of the effect of the study treatments on particular non-fatal events (e.g. ascertained through linkage to Hospital Episode Statistics), the influence of pre-existing major co-morbidity (e.g. diabetes, heart disease, lung disease, hepatic insufficiency, severe depression, severe kidney impairment, immunosuppression), and longer-term outcomes (e.g. 6 month survival) as well as in particular sub-categories of patient (e.g. by genotype).

Follow-up information is to be collected on all study participants, irrespective of whether or not they complete the scheduled course of allocated study treatment. Study staff will seek Follow-up information through various means including medical staff, reviewing information from medical notes, routine healthcare systems, and registries.

The study team require the following information from NHS Digital:

- Prompt feed from SUS on hospital discharge

- Quarterly feed on HES data

- Quarterly feeds on civil registration data (death certificate information)

This study is supported by a grant to the University of Oxford from UK Research and Innovation/National Institute for Health Research (NIHR) and by core funding provided by NIHR Oxford Biomedical Research Centre, the Wellcome Trust, the Bill and Melinda Gates Foundation, Health Data Research UK, and the Medical Research Council Population Health Research Unit, and NIHR Clinical Trials Unit Support Funding.

The new trial has been classed as an Urgent Public Health Research Study. It is one of a round of projects to receive £10.5 million as part of the £20 million rapid research response funded by UK Research and Innovation, and by the Department of Health and Social Care through the National Institute for Health Research.

COVID-19 is an emerging pathogen which presents a significant threat to the population in terms of increased morbidity and mortality, particularly among vulnerable groups such as those with pre-existing disease.

The primary objective is to provide reliable estimates of the effect of study treatments on in-hospital death (with subsidiary analyses of cause of death and death at various timepoints following discharge).

The secondary objectives are to assess the effects of study treatments on duration of hospital stay; the need for (and duration of) ventilation; and the need for renal replacement therapy.

The interim trial results will be monitored by an independent Data Monitoring Committee (DMC). The most important task for the DMC will be to assess whether the randomised comparisons in the study have provided evidence on mortality that is strong enough (with a range of uncertainty around the results that is narrow enough) to affect national and global treatment strategies. In such a circumstance, the DMC will inform the Trial Steering Committee who will make the results available to the public and amend the trial arms accordingly.

The data requested will be used to evaluate the efficacy and safety of the study treatments and will help shape the public health response. The rapid feed will be used to ensure that the trial Data Monitoring Committee have complete, up-to-date information on the major outcomes in the trial on which to base their decisions. If they find compelling evidence of efficacy or safety then that arm may be stopped and – if effective – added to standard care across the NHS.

All outputs produced will be in the form of aggregated reports with small number suppression applied.

Yielded Benefits is not a requirement for new applications.