Prioritising patients for Emergency Surgery Or Not: the impact of COVID-19 (ESORT-C19)

BACKGROUND

Patients with acute conditions who present as emergency hospital admissions may receive emergency surgery (operative) or non-operative care. Within the emergency general surgery specialty, some patients with acute conditions have improved health following emergency surgery and others from non-operative care. However, for many patients the relative benefits, risks and costs of emergency surgery versus non-operative care are unknown.

The Getting it Right First Time (GIRFT) report for emergency general surgery (https://gettingitrightfirsttime.co.uk/wp-content/uploads/2018/08/GIRFT-GeneralSurgery-Aug17-O1.pdf), found wide variation across NHS trusts in care quality and outcomes after emergency surgery, which reflect local logistical and resource constraints, but also clinical uncertainty. For common acute conditions, such as diverticular disease, there are well-developed non-operative strategies and little evidence that emergency surgery leads to better outcomes.

THIS STUDY

The Prioritising Patients for Emergency Surgery Or Not: the impact of COVID-19 (ESORT-C19) study aims to estimate the effectiveness and cost-effectiveness of emergency surgery (ES) versus non-emergency surgery (NES) to inform which patients should be prioritised for ES during the COVID-19 recovery period. The study will consider emergency hospital admissions for five acute conditions where there is great clinical uncertainty about which patients should have ES. The five conditions are: acute appendicitis, cholelithiasis, diverticulitis, abdominal wall hernia, and intestinal bowel obstruction. ESORT-C19 will consider access to ES versus NES according to patient subgroups (e.g. age >=70 or not; ethnic origin, deprivation, co-morbidities). Access to ES versus NES will be contrasted across different time periods to reflect the different ‘stages’ of the COVID-19 pandemic. The final decision as to the number and timing of the different stages will be made following initial descriptive analysis of the data, but it is planned that there will be five time periods defined as: ‘pre’ COVID-19 pandemic (February 1 2015 - 31 January 2020); ‘first peak’ (February 1 2020 to June 30 2020), ‘first recovery phase’ (July 1, 2020- September 30, 2020), ‘second peak’ (October 1, 2020 to February 28, 2021), and ‘second recovery phase’ (March 1, 2021 to August 31, 2021).

ESORT-C19 will also aim to build on the ongoing Emergency Surgery OR noT (ESORT) (https://www.lshtm.ac.uk/research/centres-projects-groups/esort) study funded by the National Institute for Health Research (NIHR) with data provided under an existing data sharing agreement (DARS-NIC-185179-V0B0T). Data provided under NIC-185179 have been used to supplement the analysis used in the final report for the ESORT study (currently under review) and three papers (one published, one accepted, one provisionally accepted). Key findings from ESORT are differences in the effectiveness and cost-effectiveness of ES for specific subgroups – in particular, level of frailty. Two patient and public involvement workshops were held in September 2021 and resulted in the dissemination of study findings aimed at patients and the public. Further dissemination and subsequent yielded benefits will follow publication of the main effectiveness and cost-effectiveness results.

Health Research Authority Research Ethics Committee approval for the ESORT-C19 study is not required as this data application does not include the flow of confidential data. Local institutional ethics approval is however adequate and in place and evidence provided to NHS Digital.

The ESORT-C19 study (https://www.lshtm.ac.uk/research/centres-projects-groups/esort#esort-c19), funded by the Health Foundation (COVID-19 Research Programme) hopes to examine the impact of COVID-19 on the effectiveness and cost-effectiveness of ES. The Health Foundation have no role in decisions made over the processing or analysis of NHS Digital data. The Health Foundation are funders only.

The data requested for ESORT-C19 are identical in scope with the data provided for ESORT with the exception of the time-period covered and will be provided under this new data sharing agreement to ensure full separation between the two studies. Cause of death is required so that the number and proportion of deaths attributed to COVID-19 can be reported and sensitivity analyses conducted to examine the impact of both COVID-19 diagnoses and deaths on study results. Duplication of data flowing under NIC-583534-X7S2N and NIC-185179-V0B0T is required as the core objective of the respective studies are the same (with the exception of the COVID-19 purpose to NIC-583534-X7S2N) - to estimate the relative effectiveness and cost-effectiveness of ES versus non-operative care for common acute conditions. The same five conditions will be assessed across both studies, with identical ICD-10 codes supplied to NHS Digital. The same variables will also be assessed across the two studies – eg. patient characteristics, surgical volume for each acute condition.

The ESORT-19 study requires a separate Data Sharing Agreement and cannot be covered by the data supplied under NIC-185179 for a number of reasons. This includes having separate ethical approval, a different funder (Health Foundation for ESORT-19 rather than National Institute for Health Research for ESORT), a different Data Sharing Agreement end date, and a different purpose for processing NHS Digital data (the effectiveness and cost-effectiveness of ES versus NES is related specifically to the COVID-19 pandemic).

London School of Hygiene & Tropical Medicine (LSHTM) is the sole data controller who will process data. Only LSHTM will determine the purpose for and the manner in which the data in ESORT-C19 are processed. Only LSHTM will have access to the pseudonymised data requested and supplied by NHS Digital. Data will not flow to other partners (listed on the ESORT website homepage – https://www.lshtm.ac.uk/research/centres-projects-groups/esort) who will only act in an advisory capacity. Clinical co-applicants comprise representatives of University of Nottingham, Royal Alexander Hospital, University of Bristol, University College London and Royal Devon and Exeter NHS Foundation Trust. These clinical co-applicants provide specialist guidance to the study only and do not carry any responsibility for the purpose and manner in which data provided under this Agreement are processed. This responsibility lies solely with LSHTM. Clinical co-applicant and research partner responsibilities include attending investigators' meetings, providing input on relevant clinical and policy matters, responding to clinical queries from LSHTM staff processing the data, and providing a clinical interpretation of the outputs of data analyses.

A patient and public involvement (PPI) group has informed the design of the study through the involvement of the co-investigator. The group are involved in this study in an advisory capacity only. Two virtual PPI panels were held on the 6th and 9th July 2020. These panels sought the views of members of the public and individuals who had experience of the conditions being studied in ESORT and ESORT-19. Panelists discussed outcomes of treatment, as well as quality of life issues. Topics of discussion included pain-management and the psychological impact of treatment, as well as the potential impact of COVID-19 on decisions about care. Further panels were held in September 2021 where a Plain English Summary of the study was shared with participants.

OBJECTIVES

The aim of the ESORT-C19 study is to estimate the effectiveness and cost-effectiveness of ES versus NES strategies to inform which patients should be prioritised for ES during the COVID-19 recovery period. Its specific objectives are:

1 - To estimate the effects of the ‘peak’ and ‘recovery’ periods versus ‘pre’ COVID-19 period, on access to ES according to patient- level characteristics, for common acute conditions presenting as emergency admissions.

2 - To estimate the effectiveness and cost-effectiveness of ES versus NES for common acute conditions presenting as emergency admissions during the ‘peak’ and ‘recovery’ versus ‘pre’ periods, according to patient-level characteristics.

3 - To recommend which patients to prioritise for ES for common acute conditions presenting as emergency admissions during the COVID-19 recovery period.

The study objectives were written in 2020 before the second COVID-19 wave in the UK and assumed an on-going recovery period following the first wave of COVID-19 that peaked in March-May 2020. It was anticipated that the recovery period would see a return to normal services for emergency surgery and other parts of the health system. Subsequent COVID-19 waves require the study to consider multiple ‘peak’ periods coinciding with high COVID hospital admissions that are interspersed with ‘non-peak’ periods rather than the anticipated single ‘recovery’ period.

DATA MINIMISATION:

1) All patients with an emergency admission which includes a diagnosis (via International Classification of Diseases version 10 (ICD-10) code) of one of 5 acute conditions (appendicitis, gallstones, diverticulitis, hernia, intestinal obstruction) in any diagnosis field of any episode of the admission, with an admission date between 1 April 2013 and latest available.

2) All episodes of all admissions (emergency and elective) which include a diagnosis (via International Classification of Diseases version 10 (ICD-10) code) of one of the above 5 acute conditions, with an admission date between 1 April 2013 and latest available.

Data are required from 2013/2014 – latest available to provide an analysis cohort split into pre-pandemic (5 years from 2014/15) and peak/recovery periods during the pandemic for analysis. The additional year of the cohort (from 2013/14) is required to calculate the instrumental variable, the tendency to operate, for all patients in the analysis cohort. These data are required to describe and control for longer-term trends in the receipt of ES and outcomes both prior to and during the COVID-19 era, and to investigate changes in the instrumental variable, the tendency to operate, during this period. An instrumental variable is a variable that allows causal relationships to be estimated without the need for a randomised controlled trial.

Overall justification for the data requested:

Hospital Episode Statistics, Admitted Patient Care (HES APC) data are required to determine the exposure (emergency surgery or non-operative care), to identify patient characteristics (age, sex, ethnicity, index of multiple deprivation) and comorbidities, to identify subsequent admissions (emergency or elective), to derive the surgical volume (a measure of quality) for each acute condition, and to derive and the proportion of emergency admissions for each acute condition where ES is undertaken (the tendency to operate, subsequently used as an instrumental variable). HES APC and HES Critical Care (CC) data are required for estimating resource use and costs associated with ES and non-operative care. National data are required to ensure sufficient events for analysis and provide nationally representative findings.

The HES APC data being requested relate to:

i) a specific cohort of adult patients (18 years+) relating to the above 5 acute conditions only, and

ii) a broader extract of adult patients (18 years+) relating to the above 5 acute conditions only which will allow the LSHTM to assess the quality of care experienced by the patients in the cohort.

i) Cohort

The cohort of patients is defined as all adult patients with an emergency admission which includes a diagnosis (via ICD-10 code) of one of 5 acute conditions (appendicitis, gallstones, diverticulitis, hernia, intestinal obstruction) in any diagnosis field of any episode of the admission, with an admission date between 1 April 2013 and latest available. This is the index admission. LSHTM are also requesting data relating to this cohort of patients for all episodes of all admissions (emergency and elective) from 1 year prior to their index admission to 1 year after their index admission (i.e. covering the period 1 April 2012 to latest available). Data at the index admission will be used to determine the exposure (ES or non-operative care); data from 1 year prior to the index admission up to and including the index admission will be used to identify patient characteristics (age, sex, ethnicity, index of multiple deprivation) and comorbidities, and for applying exclusion criteria such as prior surgery; data from the 1 year following the index admission will be used to identify subsequent admissions (emergency or elective).

The size of this data request has been minimised by requesting only data relating to the cohort patients (i.e. for the above acute conditions only); the number of years of data requested has been minimised by requesting only data relating to the period within 1 year of the index admission for each cohort patient; the number of variables requested has been minimised by only requesting those which are necessary for the proposed analysis.

Exclusion criteria for defining the index admission will include: age less than 18, a previous emergency admission for the condition within the year prior to the index admission, and referrals from tertiary referral centres.

The rationale for each of these exclusions is as follows.

a) For patients aged under 18 years old, the string consensus amongst the clinical co-applicants is that there is no clinical equipoise for the decision whether to provide ES or alternative strategies. It was therefore judged inappropriate to include this population in the study and hence in the data request;

b) Patients who had had a previous emergency admission within the preceding 12 months are more likely to have a chronic condition rather than acute presentation, and were also judged to represent a different target population;

c) Patients referred from a tertiary referral centre were also judged to be quite different in their prognosis and more likely to be triaged for ES. Hence, the inclusion of this subgroup of patients would likely lead to biased estimates of the relative effectiveness and cost-effectiveness of ES, and hence provide less useful results for NHS decision-making.

The study will include clear justification for these exclusions in the outputs from the research. LSHTM will offer clear guidance as to how the groups excluded could be considered in future studies that had different objectives, in particular with respect to the comparator strategies or service decision-makers of interest. For example, a future research programme targeted at patients under 18, may focus on those acute hospitals that provide specialist children’s services.

Health Inequalities Impact Assessment

The ESORT-19 project will provide direct evidence on the effect that the COVID-19 period has had on inequalities in the provision of ES. Specifically, in addressing objective 1, LSHTM will provide the NHS with direct evidence on the proportion of patients, and the delay to receiving ES, for these common acute conditions according to socioeconomic characteristics, such as the patients’ age, ethnicity, and index of multiple deprivation. In addressing objective 2, LSHTM will present effectiveness and cost-effectiveness results according to these same socioeconomic characteristics, in order to identify those subgroups for whom ES may be under- (or over-) provided following the outbreak of COVID-19. Under objective 3, LSHTM will provide clear recommendations on how the future provision of acute surgical services should be targeted to particular subgroups to improve efficiency, but also equity, in service provision. In making these recommendations, LSHTM will draw from their Public and Patient Involvement (PPI) findings to help ensure that recommendations, especially with respect to health inequalities, reflect the concerns of a representative group of ex-patients and public contributors.

ii) Broader extract

LSHTM are also requesting data on a very limited number of variables relating to all episodes of all admissions (emergency and elective) which include a diagnosis (via ICD-10 code) of one of 5 acute conditions (appendicitis, gallstones, diverticulitis, hernia, intestinal obstruction) in any diagnosis field of any episode of the admission, with an admission date between 1 April 2013 and latest available (i.e. not just the cohort patients). These data will be used to derive the surgical volume (a measure of quality) for each acute condition for inclusion in the instrumental variable analysis.

The size of this data request has been minimised by requesting only data relating to the diagnoses of interest; the number of years of data requested has been minimised by requesting only data corresponding to the cohort (2013/14 to latest available); the number of variables requested has been minimised by only requesting those which are necessary for the proposed analysis.

Civil Registration (Deaths)

Date of death is required for two reasons:

(i) deriving a primary outcome, and

(ii) allocation to the emergency surgery or conservative management arm.

The ESORT study's patient and public involvement (PPI) meetings discussed potential outcomes and highlighted 'days alive and out of hospital' as an important outcome. Accurate measurement of this outcome requires date of death. The second reason for requiring date of death is to enable an allowance to be made in the analysis for any bias that might arise if death occurs before surgery can occur. This relates to the use of a time window within which surgery must occur (e.g. 7 days) to be defined as 'emergency surgery' and the possibility that a patient may die without having surgery during this window. Date of death will allow the study to identify these patients for each day in the window.

The General Data Protection Regulation (GDPR) Article 6 (1) (e) and Article 9 (2) (j) are the legal basis for the processing of the data. Article 6(1)(e) - task in the public interest: LSHTM is a public authority as described under Schedule 1 of the Freedom of Information Act 2000. LSHTM has a basis in law for the lawfulness of processing as LSHTM holds a Royal Charter. The Royal Charter states “there shall be one Body Corporate and Politic under the name of the London School of Hygiene and Tropical Medicine (“London School”) for the purpose of and with the objects of promoting original research, consultancy and the study of and education in public health and tropical medicine”. It is, based upon evidence this research will generate, in the public interest to provide and prioritise effective and cost-effective health care.

Article 9(2)(j) Archiving in the public interest/scientific or historical research/statistical purposes: The data are required for the purposes of a study analysing the impact that COVID-19 has had on the effectiveness and cost-effectiveness of ES. The outcomes of this study are in the public interest as it will allow health care services to better plan and prioritise care for patients. To this end, the study meets the conditions detailed in Data Protection Act 2018 Schedule 1 Part 1 (4). The GDPR Recital 52(2) determines that this is an appropriate derogation from the prohibition on processing special categories of personal data, as demonstrated in the Purpose section of this application.

The process will involve:

1. NHS Digital provide the LSHTM with bespoke data extracts of HES APC linked to mortality data, and HES CC episodes including Master Person Service (MPS) ID.

2. Extract will be received by the LSHTM by file transfer and stored on the LSHTM secure server.

3. Eligible index episodes/admissions and derived variables such as comorbidities, surgical volume and tendency to operate will be identified from the HES APC cohort data and 1-year prior data.

4. Subsequent HES episodes will be used to identify patient outcomes in addition to those provided via linkage to Civil Registration (Deaths) data.

5. A single patient-level dataset for each condition will be created including the above derived variables. The main analysis will be on these patient-level datasets.

LSHTM will store the data on a secure server in London (at LSHTM's Keppel Street building with back-ups at LSHTM's 15-17 Tavistock Place building) which can only be accessed at LSHTM.

No organisations other than LSHTM are involved in the planned data analysis. All those involved in the processing of the data are substantive employees of LSHTM or students/honorary staff of LSHTM with contracts meeting NHS Digital's requirements. LSHTM staff and students are only granted access to the secure server if they have undergone Information Security and Awareness training. This is a requirement of LSHTM's Data Security and Protection Toolkit.

There will be no requirement nor attempt to re-identify individuals from the data.

No elements of the work will take place outside the England and Wales.

The aim is to complete the study by September 2022 with all the below outputs scheduled between approximately April 2022 and December 2023. The request for the data sharing agreement to run for three years to 2025 is to allow for potential delays in the peer review process and post-publication queries that may require further analysis of the data.

The clinical co-applicants, who are listed as collaborating partners on the ESORT study website, hope to ensure the study recommendations can quickly feed into NHS guidelines, and shape practice. The research questions tackle priorities designated by policy-makers at NHS digital, the Getting it Right First Time (GIRFT) initiative, clinical opinion leaders, service commissioners, leaders of surgical networks and Patient and Public representatives. These views have shaped the study objectives and strategy for impact. The dissemination plan has the following key components:

a) Workshops and discussions at NHS England (NHSE)

One co-applicant, a National Clinical Director at NHSE hope to lead communication of results to national policy-makers and inform imminent policy decisions, for example about Best Practice Tariffs (e.g. for emergency laparotomy) and Commissioning for Quality and Innovation. An advisory group member, a Medical Director at NHSE, fully supports the proposed study, and has agreed to chair webinars at NHSE to raise awareness of the study’s early results.

b) Future updates to GIRFT initiative

Another advisory group member, former lead of the highly influential GIRFT initiative, hopes to ensure that future updates of the GIRFT report for emergency general surgery, draw on this study’s recommendations about emergency general surgery provision in the COVID-19 recovery phases.

c) Presentation to networks of clinical decision-makers

The incoming President of the Association of Surgeons of Great Britain and Ireland (ASGBI) requested that the main results of ESORT-C19 are presented during a COVID-19 symposium at the ASGBI annual conference. The incoming president of the Association of Coloproctology of Great Britain and Ireland (ACPGBI) would like results of the ESORT-19 study to be presented at the societies’ annual meeting in 2022. The study team hope to give informal webinars to these groups to ensure the study’s early results are widely available.

d) Presentation to networks of health service managers and researchers

The study team hope to present findings to national (Health Services Research Network) and international (e.g. Academy Health) health services research conferences in 2022.

e) Wider clinical, policy and public audiences

The study's clinical and PPI panels hope to help ensure the widespread presentation and interpretation of the study findings, and that the implications for clinical guidelines are clear for clinicians, patients and carers. The study team will write blogs discussing the issues raised by emergency general surgery provision following COVID-19 for the Department of Health and Social Care support unit, Policy-Innovation Research Unit (PIRU), https://piru.ac.uk/ and for the RCS COVID-19 website, https://www.rcseng.ac.uk/coronavirus/blog-views-from-the-nhs-frontline/.

Outputs will contain only aggregate level data with small numbers suppressed in line with HES analysis guide.

The main results from the ESORT-19 study are expected to be delivered within 12 months of receipt of data with initial results within three months.

Benefits to patients:

The five emergency general surgery conditions included in ESORT-19 are typically responsible for a total of around 2,000 hospital admissions per week in England. The provision of data may allow the effectiveness and cost-effectiveness of ES to be compared with alternatives such as medical management or later planned surgery during different stages of the COVID pandemic. The study design, in particular the study populations included, has been carefully specified, drawing on clinical expertise, to generate reliable evidence for those populations where the choice of ES versus alternatives is most important during the COVID-19 periods.

This information hopes to allow NHSE and its hospitals and staff to better plan the provision and prioritisation of care given the uncertainties with the future direct and indirect impact of COVID-19. For example, alternatives to ES may be more effective in some circumstances but ES more effective in others depending on the condition, the diagnosis or other characteristics such as patient frailty, leading to greater effectiveness and better outcomes overall.

Key findings from the study hope to be promptly disseminated to NHSE, e.g. via webinar.

Inclusion of key senior NHSE staff as co-investigators or advisory group members hopes to ensure findings are presented appropriately and reach relevant decision-makers.

Yielded Benefits is not a requirement for new applications.