Standard versus Accelerated initiation of Renal Replacement Therapy in Acute Kidney Injury(STARRT-AKI): UK arm of a multi-centre randomized controlled trial

The STARRT-AKI trial (Standard versus Accelerated initiation of Renal Replacement Therapy in Acute Kidney Injury) is an international multi-centre randomised control trial (RCT) designed to address the clinical question of the optimal timing of initiation of renal replacement therapy (RRT) in patients in critical care with acute kidney injury. This international trial commenced in 2016 and recruitment closed in 2019 following recruitment of 3,000 patients. The international arm of the trial is coordinated by investigators from the University of Alberta in Canada, and statistical analysis is being undertaken in Canada. Alongside the international trial, the National Institute for Health Research (NIHR) in the UK agreed to fund an expanded UK arm for the purpose of supporting an economic evaluation of accelerated initiation of renal replacement therapy vs standard implementation of renal replacement therapy in a UK setting.

This data sharing agreement is intended to provide data on secondary care services received by patients enrolled in the UK arm of the STARRT-AKI trial and in the observational cohort study in order to allow quantification of the overall cost of each patient’s care. Inclusion of the cost of secondary care services is essential to determine the impact of accelerated vs standard initiation of renal replacement therapy on total costs of care as well as outcomes of care which will facilitate an economic evaluation. The study will determine the incremental cost and the incremental health benefits, in terms of quality adjusted life-years (QALYs) [a] of accelerated initiation of renal replacement therapy compared to standard initiation.

[a] One quality-adjusted life year (QALY) is equal to 1 year of life in perfect health. QALYs are calculated by estimating the years of life remaining for a patient following a particular treatment or intervention and weighting each year with a quality-of-life score (on a 0 to 1 scale)

The analysis will be undertaken in two parts. The first part will consider the difference in outcomes and costs over the period of observations of patients in the trial. The second part will utilise data from the trial to parameterise a simulation model which will estimate the incremental impact of accelerated or standard initiation of renal replacement therapy on lifetime costs and quality adjusted life expectancy.

The objective of the trial is to test the effectiveness of accelerated vs standard initiation of renal replacement following Kidney injury. The requested data will inform the economic analysis of the UK arm of the international trial STARRT-AKI.

The expanded UK arm aimed to recruit 580 patients, however due to the pandemic, only 191 patients were recruited in total. The international trial closed recruitment before the UK arm of the trial was able to meet its recruitment target. At that time, 189 UK patients had been enrolled. With NIHR funding, UK investigators agreed to include data from a parallel observational study to support the accurate estimation of the cost impact of the timing of initiation of renal replacement therapy in a UK setting. The continuation of the study and the inclusion of the observational data alongside the trial data ensured the study team had sufficient patient data to address the original study question – the cost-effectiveness of accelerated initiation of RRT for acute kidney injury in the UK. The observational data will NOT form part of the international trial. The purpose of this analysis is to provide UK specific data to inform decision making and clinical guidelines in the UK.

COHORT

The UK arm of the study consists of two parts.

1. A cohort of 191 participants with acute kidney injury randomised to accelerated or standard renal replacement therapy will provide data to both the UK study and the international trial. The first group are enrolled in the UK arm of STARRT-AKI who have been randomised to receive either accelerated or standard initiation of renal replacement therapy.

2. In addition, the UK data will be supplemented with a separate cohort of 568 patients meeting the inclusion criteria for the trial but not randomly assigned to treatment. The second group consists of patients eligible, but not enrolled in the trial due to a clinician decision that either accelerated or standard initiation of renal replacement therapy was in the patient’s best interest.

Under the Common-law Duty of Confidentiality, from the start of recruitment in July 2018 participants, dependent upon whether they were undergoing the observational study or randomised trial, were asked to sign a consent form. If the participant did not have the capacity to consent for either the observational study or randomised trial, a friend/relative or nominated consultee declaration was sought under the Mental Capacity Act. All patients who participated in the trial/study following declaration by a personal or professional consultee and survived were consulted by a member of the research team and invited to give “consent to continue” in the trial. If a participant declined, they were withdrawn. If a participant died before they gained capacity, they could not be consulted. In these cases, the study team received approval from the Research Ethics Committee to keep the participant in the trial and include the data in the analysis.

Both groups of patients are aged 18 or over and admitted to an intensive care unit due to acute kidney injury. They have evidence of kidney disfunction (serum creatinine ≥100 µmol/L in women and ≥ 130 µmol/L in men) and evidence of acute kidney injury defined by at least one of the following three criteria:

i) 2-fold increase in serum creatinine from a known pre-morbid baseline or result obtained during the current hospitalization;

ii) Achievement of a serum creatinine > 354 µmol/L with evidence of either a minimum increase of 27 µmol/L or an increase of 50% from pre-morbid baseline or result obtained during the current hospitalization;

iii) Urine output <6.0 mL/kg over the preceding 12 hours

Consent/consultee declaration has been sought to collect data for these patients in the same manner as patients partaking in the trial. The decision to include the observational data was taken to combat difficulties in recruitment to the trial and after careful consideration of how to mitigate the impact of selection bias.

The study team obtained patient consent/consultee declaration to provide longer term follow-up data at 1 year and beyond as necessary for the analysis. As Consultee declaration cannot override the rights of an individual’s National data opt-out request, National Data Opt Outs will be applied to study participants who have taken part in the study via Consultee Direction.

MINIMISATION

The study team have requested only the data they need to quantify the costs of secondary care received by patients in the trial and observational study. The study team have not requested data on any fields surplus to these requirements.

The study team have sought to minimise the data requested in the following ways:

1. requested data only on patients enrolled in the trial

2. restricted request to fields which will enable the study team to accurately quantify the resource use associated with critical care each patient in the trial has received following enrolment, and hence accurately estimate the cost to the NHS.

The proposed economic evaluation applies methodology routinely applied to new and existing medical therapies and technologies in the UK to assess value for money. Appropriate data security will be put in place to ensure that no identifiable record-level data is disseminated beyond the research team.

The study team are requesting data on all secondary care use for patients in the UK arm of the STARRT-AKI trial and the observational study. The data will allow quantification of the NHS secondary care resource use by each patient. The data will be supplemented by data on primary health care and social care provided by patients completing a bespoke questionnaire. These data will allow estimation of the costs of care for each patient in the trial and in the observational study. The data will be used to quantify the impact on costs of accelerated vs standard initiation of renal replacement therapy in critically ill patients including determination of uncertainty in the estimates of the cost difference. These data will inform the economic evaluation.

Data from both the trial and observational cohorts in the UK will be used to estimate the impact of the timing of initiation of renal replacement therapy on costs, and hence to underpin the economic evaluation. Patients enrolled in the trial in the UK are part of the main STARRT-AKI trial, and data on their clinical outcomes, collected as part of the trial, will inform the trial analysis undertaken at the University of Alberta in Canada. NHS Digital Record-level health data, requested as part of this application, will NOT be shared with the University of Alberta, Canada. These data will be analysed in the UK under the direction of Kings College London for the expanded UK arm and the lead health economist, who is a substantive employee of KCL.

Estimation of the cost of care and determination of the impact of the timing of the initiation of renal replacement therapy on the cost of care is essential to inform the economic evaluation. This will consist of two related sub-components: a ‘trial based’ analysis which will examine costs and outcomes over the period of direct observation for patients in the trial and observational cohorts; and a ‘model based’ analysis in which a simulation model will be used to extrapolate costs and health outcomes over the simulated lifetime of the cohort receiving treatment. The two components represent the standard approach to economic evaluation alongside a clinical trial in which the costs and/or benefits of intervention are likely to extend beyond the observation period in the clinical trial. These analyses together will constitute the whole of the trial-based analysis funded by the NIHR. Clinical data from patients in the trial will contribute the international trial.

NHS Digital pseudonymised record level data from patients in the observational study will not contribute to the international trial. NHS Digital pseudonymised record level data requested in this agreement will NOT be analysed as part of the international trial, or disseminated beyond the UK trial team, unless in an aggregated and suppressed form in line with the HES analysis guide.

The study team have requested NHS Digital pseudonymised record level data to enable the estimation of the cost of secondary care provided for patients in the trial and observational cohorts. The study team wish to cost all secondary care, so have requested data from the Hospital Episode Statistics (HES) Admitted Patient Care data set (APC or Inpatients), HES Outpatient (OP) data set, HES Accident & Emergency (A&E) (and Emergency Care Data Set or ECDS) data set and the HES Critical Care (CC) data set. The data requested is that which is necessary to determine the amount of reimbursement of the provider for each episode of care. The reimbursement will be considered to be the cost to the NHS of providing that episode of care. The study team require data to link finished consultant episodes into spells to allow appropriate assignment of a spell level reimbursement.

The data requested also consists primarily of dates of care, diagnosis codes and procedure codes to allow assignment of the appropriate Healthcare Resource Group (HRG) code. This code will be used to assign a cost to inpatient spells, and to outpatient and accident & emergency visits. The study group have requested some additional data fields related to the care patients received such as the main speciality and the treating speciality. This data will be helpful to estimate costs where other data, such as procedure codes, are missing for the episode of care.

The study team have also requested a small amount of data to help them link episodes and identify episodes which may be erroneous, such as method of admission and discharge destination.

In addition to the above data, the study team have requested detailed data on organ support during critical care from Intensive Care National Audit & Research Centre (ICNARC). Patients in both the trial and the observational cohort will be in critical care at their enrolment into the study. The cost of this care is likely to be substantial and it is important to capture any impact that accelerated initiation of RRT may have on these costs. The study team are planning to link NHS Digital pseudonymised record level to the data requested from from ICNARC to accurately quantify hospital reimbursement for each day spent in critical care as a function of the number of organs supported.

Data on hospital episodes will be used to determine the cost of the index admission to critical care as a function of age, gender, and timing of RRT. The study team will use the data on hospital episodes to determine the subsequent cost of care as a function of kidney function, age, and gender. These data, along with other data from the trial and from the literature will be used to create the parameters for a decision analytic model. The model will determine the cost-effectiveness of accelerated initiation of RRT over a time horizon of the lifetime of the patient.

NHS Digital pseudonymised record level data will allow the study team to quantify the impact of the timing of initiation of RRT on secondary care costs, which will underpin the economic evaluation of accelerated versus standard RRT. The study team need to assess costs for each patient in the trial and the observational study, so they are able to quantify the impact of the timing of RRT with as much precision as possible, and to allow them to quantify uncertainty in costs. The study team have requested data for all patients in the trial and in the observational cohort. The individual patient data is necessary to assess the cost of care for each patient. The study team will need to supply identifying data (NHS Number and Date of Birth) to NHS Digital to allow identification of the relevant patient records. These identifiers will be removed prior to NHS Digital supplying the relevant data files as requested in this agreement. Study team will also provide a pseudonymised study ID which will remain in the files when they are returned to the study team.

The study team have requested record level pseudonymised data for the period April 2017 to May 2022. These data will provide the study team with a minimum of one year’s data prior to the index admission at enrolment into the trial or observational study for each patient. This will enable the study team to assess the costs of care accrued by patients in the year prior to admission. The study team will use these data to adjust their estimates of the difference in costs between patients receiving accelerated versus standard timing of RRT for any differences between patients in their costs for the year prior to the index admission. This is particularly important for patients in the observational data set for whom assignment to accelerated or standard initiation of RRT is not randomised. This will maximise the length of follow-up for which the study team are able to access cost data on patients. The economic analysis will apply a lifetime time horizon which will require the extrapolation of data on costs and outcomes. The study team will use the available data on each patient to better inform the estimates of the cost impact of accelerated vs standard initiation of RRT in the years following the index admission to critical care and reduce the need to extrapolate cost data in the analysis.

The economic analysis is predicated on the use of individual patient data on resource use, quality of life and mortality to assess the cost-effectiveness of accelerated initiation of RRT. The study team sought to reduce the burden on patients completing questionnaires by utilising administrative data, from Hospital Episode Statistics (HES), Renal Registry, ICNARC and Scottish HES, to collect data on resource use. Patients complete a brief questionnaire on health-related quality of life and a bespoke questionnaire collecting data on primary health and social care received. Data on secondary health care will be taken from administrative sources as detailed above. This approach was implemented to minimize the demands on patient time and help ensure the study team receive data as accurate as possible on resource use to reduce bias in the estimation of the cost impact of accelerated initiation of RRT.

The data from the NHS Digital HES Critical Care data set is essential to quantify the impact of the timing of renal replacement therapy on the costs of critical care during the index hospitalization. Specifically, the analysis will exploit detailed data on organ support to estimate costs reflecting the level of support during the critical care episode. Consequently, the study team have requested data on organ support. Data enabling them to link the critical episode with the admitted inpatient episode will enable the study team to cost the entire spell covering the index admission.

DATA CONTROLLERS, DATA PROCESSORS AND FUNDING

The Data Controllers for this data sharing agreement are Guy's and St Thomas’ NHS Foundation Trust and King’s College London. The economic evaluation will be led by researchers at King’s College London. The Principal Investigator (PI) is a consultant at Guy's & St Thomas’ Hospital within Guy's & St Thomas’ NHS Foundation Trust. The PI is supported by a trial manager at Guy's & St Thomas’ Hospital within Guy's & St Thomas’NHS Foundation Trust and the Clinical Trials Unit at King’s College London. The PI will not have access to NHS Digital record-level data.

Participants in the trial are also part of the international trial led by University of Toronto and University of Alberta, both in Canada. Patients in the trial are contributing clinical data to the international trial. Patients in the observational study are NOT contributing clinical data to the international trial. Both groups will contribute clinical data to the economic evaluation in the UK setting to be undertaken by the applicants. The NHS Digital record level data which are the subject of this application will NOT be part of the main trial. NHS Digital record level data will only be used to undertake an economic evaluation of accelerated initiation of RRT in a UK setting.

The funding for the study is provided solely by the National Institute for Health Research (NIHR). The NIHR approved funding for the continuation of the study to include an observational cohort subject to reassurance from the applicants that the data can be used to address the original study question whilst limiting any potential bias arising from the inclusion of observational data. The NIHR do not make any decisions determining the purposes and means of the processing of the data and are therefore not considered a data controller for this agreement. The UK study team (Guys and St Thomas’ NHS Foundation Trust and King’s College London) have complete authority and control on the analysis of this data to inform the cost-effectiveness analysis. The sole Data Processor for this agreement is King’s College London and Guys and St Thomas’ NHS Foundation Trust will have no access to record-level NHS Digital data.

LEGAL BASIS

NHS Digital record level data is required for the study to assess the resource implications of the timing of initiation of renal replacement therapy. This is an essential component of the analysis of cost-effectiveness, which is the purpose and justification for the expanded UK recruitment to the STARRT-AKI trial. This analysis is in the public interest since it seeks to provide important data to inform clinical guidelines on the treatment of patients with acute kidney injury in the NHS. Therefore, both Data Controllers use the GDPR Article 6(1)(e) for the Legal Basis for Processing.

Additionally, the data is health data which is a Special Category of Personal data. The data request is necessary to allow the study team to determine with the best available precision the resources consumed by patients receiving accelerated or standard initiation of renal replacement therapy. Renal replacement therapy is resource intensive and modest differences will have important consequences for the NHS. An accurate determination of the impact of the timing of initiation of renal replacement therapy will increase the robustness and value of information from this study. Therefore, both Data Controllers additionally use the GDPR Article 9(2)(j) for Legal Basis for Processing.

- The data is required for research purposes 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. The Data Protection Act 2018 Schedule 1 Part 1 states that when processing special category data under the condition relating to research, the condition is met when:

(a) 'necessary for archiving purposes, scientific or historical research purposes or statistical purposes'. As described above the data is considered necessary for the performance of the task

(b) 'carried out in accordance with Article 89(1) of the GDPR'. In accordance with this article, processing is subject to appropriate safeguards. These include:

i. The data will be pseudonymised prior to dissemination by NHS Digital to the data recipient;

ii. 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);

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

iv. 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);

(c) 'is in the public interest'. NHS Digital is content that the information set out in the Benefits section of the application evidences that the data processing will be in the public interest - There is a strong public interest in ensuring that clinical guidelines in the UK recommend treatments which demonstrate value for money in order to ensure that public money is spent as effectively as possible in providing NHS services. The data is necessary to accurately quantify the impact of accelerated RRT on the cost of patients of supporting patients in critical care with acute kidney injury. The subsequent analysis hope to inform clinical guidelines in the UK and help commissioners to anticipate resource demands for patients with acute kidney injury.

The data requested aims to allow the accurate estimation of the impact of the timing of initiation of renal replacement therapy on resource use and costs of caring for patients with acute kidney injury. This data aims to be combined with individual patient data collected on outcomes (primarily mortality) for patients in the trial and observational data. The study team plan to analyse these data with the hope to address the cost-effectiveness of accelerated initiation of renal replacement therapy in an NHS setting over the period of the trial. Data is planned to also be used to estimate cost parameters for a decision model which aims to be used to extrapolate costs and outcomes over patients’ lifetimes and allow a lifetime analysis. The international trial ended in 2020 and the trial results have been published. This analysis aims to help contextualise the findings within a NHS setting and support decision making that is relevant to the UK context.

As a result of the data processing the study team intend to create various reports, submissions to peer reviewed journals, presentations, and representation at conferences.

The data processing aims to allow estimation of the mean costs of care for patients in the accelerated and standard initiation of RRT arms of the trial. In addition, it aims to allow estimation of the mean costs of specific episodes of care, such as the index admission to critical care, as a function of patient characteristics including age and Sequential Organ Failure Assessment (SOFA) score. These mean costs aim to be reported, alongside quantification of uncertainty in the cost estimates. They hope to also inform the economic analysis of accelerated initiation of RRT. That analysis is planned to consist of a trial-based analysis utilising the observed data and a model-based evaluation in which costs and outcomes are extrapolated over patient lifetimes.

The study team intends to publish the mean costs described above and the evaluation results supported by those costs in a report to the Human Tissue Authority (HTA). These also hope to form the basis of journal articles and conference presentations. The study team intend to target appropriate peer reviewed journals and conferences with relevant clinical audiences. Finally, the simulation model aims to be made available to the main trial team in Canada to allow customization and application to different jurisdictions by researchers considered appropriate. The UK study team and the Canadian study team will attempt to ensure that the value of the model in informing decision-making is maximised by ensuring that those who are given access are able and intend to apply and adapt the model appropriately.

No record level individual patient data will be published in any outputs. All data will be aggregate data with small number suppression applied in accordance with the HES Analysis guide. Primarily, the data derived from HES data will be the costs associated with accelerated and standard initiation of RRT. The study team intend to report the mean costs of key episodes of care such as the index admission, and the relationship with patient characteristics. The study team also intend to report data relating to outcomes such as mortality.

The study team intend to publish regular progress reports on the analysis of the trial on the STARRT-AKI website. Once the study team have completed the analysis, their primary dissemination strategy hopes to consist of publication in appropriate high impact journals alongside submission of a detailed report of the full findings of the study to the NIHR. In addition, the study team aims to seek to disseminate findings at appropriate international conferences, for instance the annual congress of the Intensive Care Society (ICS) UK, the annual International Symposium on Intensive Care & Emergency Medicine (ISICEM), the annual congress of the European Society of Intensive Care (ESICM) and UK Kidney Week of the Renal Association.

The study team expect to keep trial participants informed of publications. Where possible they will ensure publications are open access to maximise accessibility. The modelling work they aim to undertake, which will utilise data from HES for estimation of some parameters, aims to be made available to the wider international trial investigators to support adaptation and application to different jurisdictions.

The study team intend to communicate trial progress in the UK through NIHR networks, Intensive Care Unit (ICU ) and Renal networks and via the newsletters of the Kidney Patient Association, Kidney Care UK, the Biomedical Research Centre at Guys and St Thomas’ NHS Foundation Trust and "Think Kidneys" (a national programme supported by NHS England).

After completion of the trial, the results aim to be reported in accordance with the Consolidated Standards of Reporting Trials (CONSORT) guidelines. The findings are expected to be presented to collaborators / investigators, and subsequently at national and international meetings, for instance the annual congress of the Intensive Care Society (ICS) UK, the annual International Symposium on Intensive Care & Emergency Medicine (ISICEM), the annual congress of the European Society of Intensive Care (ESICM), UK Kidney Week of the Renal Association and the annual meeting of the American Society of Nephrology. The study team anticipate publishing the main trial results in a major high impact journal. The Health economic evaluation is expected to be published with this report, or in a separate detailed paper. These publications aim to supplement the final report published in the NIHR Health Technology Assessment (HTA ) journal.

With the help of Personal and Public Involvement (PPIs), a lay person’s summary is intended to be sent to relevant local and national patient support and liaison groups and charities. One of the co-applicants of the HTA grant is acting as PPI for this study. In addition, the study team is working closely with members of Guy’s Hospital Kidney Patient Association and ICUsteps, a charity supporting ICU survivors, who have previously helped with the conduct of the STARRT AKI trial in the UK. The study team aim to seek their input regarding dissemination of the study results. A report is also expected to be sent to the INVOLVE registry (an open-access database which registers research health care projects involving members of the public as partners in the research process). This ensures full transparency of the results and is recommended by research authorities.

Following peer reviewed publication, the key findings are intended to be posted on institutional websites available to the general public and communicated through press releases to ensure dissemination to research participants, patients and the broader public (e.g., through scientific reports, conference presentations and renal charity newsletters). In addition, the study team aim to work with professional societies and ensure that the conclusions are included in their official guidelines, including the AKI guideline by the Renal Association UK, recommendations by Kidney Disease Improving Global Outcomes (KDIGO) and the AKI guideline by the European Society of Intensive Care Medicine. Furthermore, the original HTA application was supported by National Institute for Health and Care Excellence (NICE ). It is very likely that the results of this analysis will also be incorporated in future NICE guidance.

As part of the analysis the study team aim to construct a simulation model to capture the longer-term impacts of the renal replacement therapy, and in particular, the impact of the timing of renal replacement therapy. The model is intended to allow evaluation of the cost-effectiveness of accelerated initiation of renal replacement therapy to inform decision making. The model will be primarily intended to inform decision making in the UK. However, the study team aim to make the model available to other investigators within the international trial who may wish to adapt the model to local settings. Along with the PI of the international trial the study team will consider the suitability of any requests to use the model. The study team aim to supply the model to any group they think will use the model appropriately to help inform decision making in their jurisdiction. The study team aim to annotate the model and provide documentation to help researchers adapt the model to their settings and they hope to provide advice to them to support adaptation and application of the model. The economic evaluation is funded by the NIHR with an expectation that the findings will be made freely available to inform decision making in the UK and beyond. The funders have no specific requirements regarding access or ownership of data.

The study team plan to have their analysis completed by January 2023. The study team expect to commence drafting of publications shortly after. The study team aim to submit the first draft of the final report to the NIHR in 2023. The study team aim to submit a draft manuscript of the findings of the economic evaluation to a high impact, peer reviewed open access journal in 2023, including The Journal of the American Medical Association (JAMA), Clinical Journal of the American Society of Nephrology (CJASN) and Intensive Care Medicine.

The STARRT-AKI trial (Standard versus Accelerated initiation of Renal Replacement Therapy in Acute Kidney Injury) is an international multi-centre randomised control trial (RCT) designed to address the clinical question of the optimal timing of initiation of renal replacement therapy (RRT) in patients in critical care with acute kidney injury. This international trial commenced in 2016 and recruitment closed in 2019 following recruitment of 3,000 patients. The international arm of the trial is coordinated by investigators from the University of Alberta in Canada, and statistical analysis is being undertaken in Canada. Alongside the international trial, the National Institute for Health Research (NIHR) in the UK agreed to fund an expanded UK arm for the purpose of supporting an economic evaluation of accelerated initiation of renal replacement therapy vs standard implementation of renal replacement therapy in a UK setting.

This data sharing agreement is intended to provide data on secondary care services received by patients enrolled in the UK arm of the STARRT-AKI trial and in the observational cohort study in order to allow quantification of the overall cost of each patient’s care. Inclusion of the cost of secondary care services is essential to determine the impact of accelerated vs standard initiation of renal replacement therapy on total costs of care as well as outcomes of care which will facilitate an economic evaluation. The study will determine the incremental cost and the incremental health benefits, in terms of quality adjusted life-years (QALYs) [a] of accelerated initiation of renal replacement therapy compared to standard initiation.

[a] One quality-adjusted life year (QALY) is equal to 1 year of life in perfect health. QALYs are calculated by estimating the years of life remaining for a patient following a particular treatment or intervention and weighting each year with a quality-of-life score (on a 0 to 1 scale)

The analysis will be undertaken in two parts. The first part will consider the difference in outcomes and costs over the period of observations of patients in the trial. The second part will utilise data from the trial to parameterise a simulation model which will estimate the incremental impact of accelerated or standard initiation of renal replacement therapy on lifetime costs and quality adjusted life expectancy.

The objective of the trial is to test the effectiveness of accelerated vs standard initiation of renal replacement following Kidney injury. The requested data will inform the economic analysis of the UK arm of the international trial STARRT-AKI.

The expanded UK arm aimed to recruit 580 patients, however due to the pandemic, only 191 patients were recruited in total. The international trial closed recruitment before the UK arm of the trial was able to meet its recruitment target. At that time, 189 UK patients had been enrolled. With NIHR funding, UK investigators agreed to include data from a parallel observational study to support the accurate estimation of the cost impact of the timing of initiation of renal replacement therapy in a UK setting. The continuation of the study and the inclusion of the observational data alongside the trial data ensured the study team had sufficient patient data to address the original study question – the cost-effectiveness of accelerated initiation of RRT for acute kidney injury in the UK. The observational data will NOT form part of the international trial. The purpose of this analysis is to provide UK specific data to inform decision making and clinical guidelines in the UK.

COHORT

The UK arm of the study consists of two parts.

1. A cohort of 191 participants with acute kidney injury randomised to accelerated or standard renal replacement therapy will provide data to both the UK study and the international trial. The first group are enrolled in the UK arm of STARRT-AKI who have been randomised to receive either accelerated or standard initiation of renal replacement therapy.

2. In addition, the UK data will be supplemented with a separate cohort of 568 patients meeting the inclusion criteria for the trial but not randomly assigned to treatment. The second group consists of patients eligible, but not enrolled in the trial due to a clinician decision that either accelerated or standard initiation of renal replacement therapy was in the patient’s best interest.

Under the Common-law Duty of Confidentiality, from the start of recruitment in July 2018 participants, dependent upon whether they were undergoing the observational study or randomised trial, were asked to sign a consent form. If the participant did not have the capacity to consent for either the observational study or randomised trial, a friend/relative or nominated consultee declaration was sought under the Mental Capacity Act. All patients who participated in the trial/study following declaration by a personal or professional consultee and survived were consulted by a member of the research team and invited to give “consent to continue” in the trial. If a participant declined, they were withdrawn. If a participant died before they gained capacity, they could not be consulted. In these cases, the study team received approval from the Research Ethics Committee to keep the participant in the trial and include the data in the analysis.

Both groups of patients are aged 18 or over and admitted to an intensive care unit due to acute kidney injury. They have evidence of kidney disfunction (serum creatinine ≥100 µmol/L in women and ≥ 130 µmol/L in men) and evidence of acute kidney injury defined by at least one of the following three criteria:

i) 2-fold increase in serum creatinine from a known pre-morbid baseline or result obtained during the current hospitalization;

ii) Achievement of a serum creatinine > 354 µmol/L with evidence of either a minimum increase of 27 µmol/L or an increase of 50% from pre-morbid baseline or result obtained during the current hospitalization;

iii) Urine output <6.0 mL/kg over the preceding 12 hours

Consent/consultee declaration has been sought to collect data for these patients in the same manner as patients partaking in the trial. The decision to include the observational data was taken to combat difficulties in recruitment to the trial and after careful consideration of how to mitigate the impact of selection bias.

The study team obtained patient consent/consultee declaration to provide longer term follow-up data at 1 year and beyond as necessary for the analysis. As consultee declaration cannot override the rights of an individual’s National data opt-out request, the study team has agreed to National Data Opt Outs being applied to the whole cohort for this agreement.

MINIMISATION

The study team have requested only the data they need to quantify the costs of secondary care received by patients in the trial and observational study. The study team have not requested data on any fields surplus to these requirements.

The study team have sought to minimise the data requested in the following ways:

1. requested data only on patients enrolled in the trial

2. restricted request to fields which will enable the study team to accurately quantify the resource use associated with critical care each patient in the trial has received following enrolment, and hence accurately estimate the cost to the NHS.

The proposed economic evaluation applies methodology routinely applied to new and existing medical therapies and technologies in the UK to assess value for money. Appropriate data security will be put in place to ensure that no identifiable record-level data is disseminated beyond the research team.

The study team are requesting data on all secondary care use for patients in the UK arm of the STARRT-AKI trial and the observational study. The data will allow quantification of the NHS secondary care resource use by each patient. The data will be supplemented by data on primary health care and social care provided by patients completing a bespoke questionnaire. These data will allow estimation of the costs of care for each patient in the trial and in the observational study. The data will be used to quantify the impact on costs of accelerated vs standard initiation of renal replacement therapy in critically ill patients including determination of uncertainty in the estimates of the cost difference. These data will inform the economic evaluation.

Data from both the trial and observational cohorts in the UK will be used to estimate the impact of the timing of initiation of renal replacement therapy on costs, and hence to underpin the economic evaluation. Patients enrolled in the trial in the UK are part of the main STARRT-AKI trial, and data on their clinical outcomes, collected as part of the trial, will inform the trial analysis undertaken at the University of Alberta in Canada. NHS Digital Record-level health data, requested as part of this application, will NOT be shared with the University of Alberta, Canada. These data will be analysed in the UK under the direction of Kings College London for the expanded UK arm and the lead health economist, who is a substantive employee of KCL.

Estimation of the cost of care and determination of the impact of the timing of the initiation of renal replacement therapy on the cost of care is essential to inform the economic evaluation. This will consist of two related sub-components: a ‘trial based’ analysis which will examine costs and outcomes over the period of direct observation for patients in the trial and observational cohorts; and a ‘model based’ analysis in which a simulation model will be used to extrapolate costs and health outcomes over the simulated lifetime of the cohort receiving treatment. The two components represent the standard approach to economic evaluation alongside a clinical trial in which the costs and/or benefits of intervention are likely to extend beyond the observation period in the clinical trial. These analyses together will constitute the whole of the trial-based analysis funded by the NIHR. Clinical data from patients in the trial will contribute the international trial.

NHS Digital pseudonymised record level data from patients in the observational study will not contribute to the international trial. NHS Digital pseudonymised record level data requested in this application will NOT be analysed as part of the international trial, or disseminated beyond the UK trial team, unless in an aggregated and suppressed form in line with the HES analysis guide.

The study team have requested NHS Digital pseudonymised record level data to enable the estimation of the cost of secondary care provided for patients in the trial and observational cohorts. The study team wish to cost all secondary care, so have requested data from the Hospital Episode Statistics (HES) Admitted Patient Care data set (APC or Inpatients), HES Outpatient (OP) data set, HES Accident & Emergency (A&E) (and Emergency Care Data Set or ECDS) data set and the HES Critical Care (CC) data set. The data requested is that which is necessary to determine the amount of reimbursement of the provider for each episode of care. The reimbursement will be considered to be the cost to the NHS of providing that episode of care. The study team require data to link finished consultant episodes into spells to allow appropriate assignment of a spell level reimbursement.

The data requested also consists primarily of dates of care, diagnosis codes and procedure codes to allow assignment of the appropriate Healthcare Resource Group (HRG) code. This code will be used to assign a cost to inpatient spells, and to outpatient and accident & emergency visits. The study group have requested some additional data fields related to the care patients received such as the main speciality and the treating speciality. This data will be helpful to estimate costs where other data, such as procedure codes, are missing for the episode of care.

The study team have also requested a small amount of data to help them link episodes and identify episodes which may be erroneous, such as method of admission and discharge destination.

In addition to the above data, the study team have requested detailed data on organ support during critical care from Intensive Care National Audit & Research Centre (ICNARC). Patients in both the trial and the observational cohort will be in critical care at their enrolment into the study. The cost of this care is likely to be substantial and it is important to capture any impact that accelerated initiation of RRT may have on these costs. The study team are planning to link NHS Digital pseudonymised record level to the data requested from from ICNARC to accurately quantify hospital reimbursement for each day spent in critical care as a function of the number of organs supported.

Data on hospital episodes will be used to determine the cost of the index admission to critical care as a function of age, gender, and timing of RRT. The study team will use the data on hospital episodes to determine the subsequent cost of care as a function of kidney function, age, and gender. These data, along with other data from the trial and from the literature will be used to create the parameters for a decision analytic model. The model will determine the cost-effectiveness of accelerated initiation of RRT over a time horizon of the lifetime of the patient.

NHS Digital pseudonymised record level data will allow the study team to quantify the impact of the timing of initiation of RRT on secondary care costs, which will underpin the economic evaluation of accelerated versus standard RRT. The study team need to assess costs for each patient in the trial and the observational study, so they are able to quantify the impact of the timing of RRT with as much precision as possible, and to allow them to quantify uncertainty in costs. The study team have requested data for all patients in the trial and in the observational cohort. The individual patient data is necessary to assess the cost of care for each patient. The study team will need to supply identifying data (NHS Number and Date of Birth) to NHS Digital to allow identification of the relevant patient records. These identifiers will be removed prior to NHS Digital supplying the relevant data files as requested in this agreement. Study team will also provide a pseudonymised study ID which will remain in the files when they are returned to the study team.

The study team have requested record level pseudonymised data for the period April 2017 to May 2022. These data will provide the study team with a minimum of one year’s data prior to the index admission at enrolment into the trial or observational study for each patient. This will enable the study team to assess the costs of care accrued by patients in the year prior to admission. The study team will use these data to adjust their estimates of the difference in costs between patients receiving accelerated versus standard timing of RRT for any differences between patients in their costs for the year prior to the index admission. This is particularly important for patients in the observational data set for whom assignment to accelerated or standard initiation of RRT is not randomised. This will maximise the length of follow-up for which the study team are able to access cost data on patients. The economic analysis will apply a lifetime time horizon which will require the extrapolation of data on costs and outcomes. The study team will use the available data on each patient to better inform the estimates of the cost impact of accelerated vs standard initiation of RRT in the years following the index admission to critical care and reduce the need to extrapolate cost data in the analysis.

The economic analysis is predicated on the use of individual patient data on resource use, quality of life and mortality to assess the cost-effectiveness of accelerated initiation of RRT. The study team sought to reduce the burden on patients completing questionnaires by utilising administrative data, from Hospital Episode Statistics (HES), Renal Registry, ICNARC and Scottish HES, to collect data on resource use. Patients complete a brief questionnaire on health-related quality of life and a bespoke questionnaire collecting data on primary health and social care received. Data on secondary health care will be taken from administrative sources as detailed above. This approach was implemented to minimize the demands on patient time and help ensure the study team receive data as accurate as possible on resource use to reduce bias in the estimation of the cost impact of accelerated initiation of RRT.

The data from the NHS Digital HES Critical Care data set is essential to quantify the impact of the timing of renal replacement therapy on the costs of critical care during the index hospitalization. Specifically, the analysis will exploit detailed data on organ support to estimate costs reflecting the level of support during the critical care episode. Consequently, the study team have requested data on organ support. Data enabling them to link the critical episode with the admitted inpatient episode will enable the study team to cost the entire spell covering the index admission.

DATA CONTROLLERS, DATA PROCESSORS AND FUNDING

The Data Controllers for this data sharing agreement are Guy's and St Thomas’ NHS Foundation Trust and King’s College London. The economic evaluation will be led by researchers at King’s College London. The Principal Investigator (PI) is a consultant at Guy's & St Thomas’ Hospital within Guy's & St Thomas’ NHS Foundation Trust. The PI is supported by a trial manager at Guy's & St Thomas’ Hospital within Guy's & St Thomas’NHS Foundation Trust and the Clinical Trials Unit at King’s College London. The PI will not have access to NHS Digital record-level data.

Participants in the trial are also part of the international trial led by University of Toronto and University of Alberta, both in Canada. Patients in the trial are contributing clinical data to the international trial. Patients in the observational study are NOT contributing clinical data to the international trial. Both groups will contribute clinical data to the economic evaluation in the UK setting to be undertaken by the applicants. The NHS Digital record level data which are the subject of this application will NOT be part of the main trial. NHS Digital record level data will only be used to undertake an economic evaluation of accelerated initiation of RRT in a UK setting.

The funding for the study is provided solely by the National Institute for Health Research (NIHR). The NIHR approved funding for the continuation of the study to include an observational cohort subject to reassurance from the applicants that the data can be used to address the original study question whilst limiting any potential bias arising from the inclusion of observational data. The NIHR do not make any decisions determining the purposes and means of the processing of the data and are therefore not considered a data controller for this agreement. The UK study team (Guys and St Thomas’ NHS Foundation Trust and King’s College London) have complete authority and control on the analysis of this data to inform the cost-effectiveness analysis. The sole Data Processor for this agreement is King’s College London and Guys and St Thomas’ NHS Foundation Trust will have no access to record-level NHS Digital data.

LEGAL BASIS

NHS Digital record level data is required for the study to assess the resource implications of the timing of initiation of renal replacement therapy. This is an essential component of the analysis of cost-effectiveness, which is the purpose and justification for the expanded UK recruitment to the STARRT-AKI trial. This analysis is in the public interest since it seeks to provide important data to inform clinical guidelines on the treatment of patients with acute kidney injury in the NHS. Therefore, both Data Controllers use the GDPR Article 6(1)(e) for the Legal Basis for Processing.

Additionally, the data is health data which is a Special Category of Personal data. The data request is necessary to allow the study team to determine with the best available precision the resources consumed by patients receiving accelerated or standard initiation of renal replacement therapy. Renal replacement therapy is resource intensive and modest differences will have important consequences for the NHS. An accurate determination of the impact of the timing of initiation of renal replacement therapy will increase the robustness and value of information from this study. Therefore, both Data Controllers additionally use the GDPR Article 9(2)(j) for Legal Basis for Processing.

- The data is required for research purposes 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. The Data Protection Act 2018 Schedule 1 Part 1 states that when processing special category data under the condition relating to research, the condition is met when:

(a) 'necessary for archiving purposes, scientific or historical research purposes or statistical purposes'. As described above the data is considered necessary for the performance of the task

(b) 'carried out in accordance with Article 89(1) of the GDPR'. In accordance with this article, processing is subject to appropriate safeguards. These include:

i. The data will be pseudonymised prior to dissemination by NHS Digital to the data recipient;

ii. 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);

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

iv. 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);

(c) 'is in the public interest'. NHS Digital is content that the information set out in the Benefits section of the application evidences that the data processing will be in the public interest - There is a strong public interest in ensuring that clinical guidelines in the UK recommend treatments which demonstrate value for money in order to ensure that public money is spent as effectively as possible in providing NHS services. The data is necessary to accurately quantify the impact of accelerated RRT on the cost of patients of supporting patients in critical care with acute kidney injury. The subsequent analysis hope to inform clinical guidelines in the UK and help commissioners to anticipate resource demands for patients with acute kidney injury.

The data requested aims to allow the accurate estimation of the impact of the timing of initiation of renal replacement therapy on resource use and costs of caring for patients with acute kidney injury. This data aims to be combined with individual patient data collected on outcomes (primarily mortality) for patients in the trial and observational data. The study team plan to analyse these data with the hope to address the cost-effectiveness of accelerated initiation of renal replacement therapy in an NHS setting over the period of the trial. Data is planned to also be used to estimate cost parameters for a decision model which aims to be used to extrapolate costs and outcomes over patients’ lifetimes and allow a lifetime analysis. The international trial ended in 2020 and the trial results have been published. This analysis aims to help contextualise the findings within a NHS setting and support decision making that is relevant to the UK context.

As a result of the data processing the study team intend to create various reports, submissions to peer reviewed journals, presentations, and representation at conferences.

The data processing aims to allow estimation of the mean costs of care for patients in the accelerated and standard initiation of RRT arms of the trial. In addition, it aims to allow estimation of the mean costs of specific episodes of care, such as the index admission to critical care, as a function of patient characteristics including age and Sequential Organ Failure Assessment (SOFA) score. These mean costs aim to be reported, alongside quantification of uncertainty in the cost estimates. They hope to also inform the economic analysis of accelerated initiation of RRT. That analysis is planned to consist of a trial-based analysis utilising the observed data and a model-based evaluation in which costs and outcomes are extrapolated over patient lifetimes.

The study team intends to publish the mean costs described above and the evaluation results supported by those costs in a report to the Human Tissue Authority (HTA). These also hope to form the basis of journal articles and conference presentations. The study team intend to target appropriate peer reviewed journals and conferences with relevant clinical audiences. Finally, the simulation model aims to be made available to the main trial team in Canada to allow customization and application to different jurisdictions by researchers considered appropriate. The UK study team and the Canadian study team will attempt to ensure that the value of the model in informing decision-making is maximised by ensuring that those who are given access are able and intend to apply and adapt the model appropriately.

No record level individual patient data will be published in any outputs. All data will be aggregate data with small number suppression applied in accordance with the HES Analysis guide. Primarily, the data derived from HES data will be the costs associated with accelerated and standard initiation of RRT. The study team intend to report the mean costs of key episodes of care such as the index admission, and the relationship with patient characteristics. The study team also intend to report data relating to outcomes such as mortality.

The study team intend to publish regular progress reports on the analysis of the trial on the STARRT-AKI website. Once the study team have completed the analysis, their primary dissemination strategy hopes to consist of publication in appropriate high impact journals alongside submission of a detailed report of the full findings of the study to the NIHR. In addition, the study team aims to seek to disseminate findings at appropriate international conferences, for instance the annual congress of the Intensive Care Society (ICS) UK, the annual International Symposium on Intensive Care & Emergency Medicine (ISICEM), the annual congress of the European Society of Intensive Care (ESICM) and UK Kidney Week of the Renal Association.

The study team expect to keep trial participants informed of publications. Where possible they will ensure publications are open access to maximise accessibility. The modelling work they aim to undertake, which will utilise data from HES for estimation of some parameters, aims to be made available to the wider international trial investigators to support adaptation and application to different jurisdictions.

The study team intend to communicate trial progress in the UK through NIHR networks, Intensive Care Unit (ICU ) and Renal networks and via the newsletters of the Kidney Patient Association, Kidney Care UK, the Biomedical Research Centre at Guys and St Thomas’ NHS Foundation Trust and "Think Kidneys" (a national programme supported by NHS England).

After completion of the trial, the results aim to be reported in accordance with the Consolidated Standards of Reporting Trials (CONSORT) guidelines. The findings are expected to be presented to collaborators / investigators, and subsequently at national and international meetings, for instance the annual congress of the Intensive Care Society (ICS) UK, the annual International Symposium on Intensive Care & Emergency Medicine (ISICEM), the annual congress of the European Society of Intensive Care (ESICM), UK Kidney Week of the Renal Association and the annual meeting of the American Society of Nephrology. The study team anticipate publishing the main trial results in a major high impact journal. The Health economic evaluation is expected to be published with this report, or in a separate detailed paper. These publications aim to supplement the final report published in the NIHR Health Technology Assessment (HTA ) journal.

With the help of Personal and Public Involvement (PPIs), a lay person’s summary is intended to be sent to relevant local and national patient support and liaison groups and charities. One of the co-applicants of the HTA grant is acting as PPI for this study. In addition, the study team is working closely with members of Guy’s Hospital Kidney Patient Association and ICUsteps, a charity supporting ICU survivors, who have previously helped with the conduct of the STARRT AKI trial in the UK. The study team aim to seek their input regarding dissemination of the study results. A report is also expected to be sent to the INVOLVE registry (an open-access database which registers research health care projects involving members of the public as partners in the research process). This ensures full transparency of the results and is recommended by research authorities.

Following peer reviewed publication, the key findings are intended to be posted on institutional websites available to the general public and communicated through press releases to ensure dissemination to research participants, patients and the broader public (e.g., through scientific reports, conference presentations and renal charity newsletters). In addition, the study team aim to work with professional societies and ensure that the conclusions are included in their official guidelines, including the AKI guideline by the Renal Association UK, recommendations by Kidney Disease Improving Global Outcomes (KDIGO) and the AKI guideline by the European Society of Intensive Care Medicine. Furthermore, the original HTA application was supported by National Institute for Health and Care Excellence (NICE ). It is very likely that the results of this analysis will also be incorporated in future NICE guidance.

As part of the analysis the study team aim to construct a simulation model to capture the longer-term impacts of the renal replacement therapy, and in particular, the impact of the timing of renal replacement therapy. The model is intended to allow evaluation of the cost-effectiveness of accelerated initiation of renal replacement therapy to inform decision making. The model will be primarily intended to inform decision making in the UK. However, the study team aim to make the model available to other investigators within the international trial who may wish to adapt the model to local settings. Along with the PI of the international trial the study team will consider the suitability of any requests to use the model. The study team aim to supply the model to any group they think will use the model appropriately to help inform decision making in their jurisdiction. The study team aim to annotate the model and provide documentation to help researchers adapt the model to their settings and they hope to provide advice to them to support adaptation and application of the model. The economic evaluation is funded by the NIHR with an expectation that the findings will be made freely available to inform decision making in the UK and beyond. The funders have no specific requirements regarding access or ownership of data.

The study team plan to have their analysis completed by January 2023. The study team expect to commence drafting of publications shortly after. The study team aim to submit the first draft of the final report to the NIHR in 2023. The study team aim to submit a draft manuscript of the findings of the economic evaluation to a high impact, peer reviewed open access journal in 2023, including The Journal of the American Medical Association (JAMA), Clinical Journal of the American Society of Nephrology (CJASN) and Intensive Care Medicine.

Yielded Benefits is not a requirement for new applications.