HES and NICOR data linkage for cardiac failure population analysis

There are an estimated 81,000 people living with heart failure, with admissions accounting for an estimated 1 million inpatient bed-days (around 2% of the entire NHS total). Up to 70% of these costs are due to hospitalisation costs, so attempts to reduce costs must be focused on reduced admissions.

King’s College London (KCL)* requires 5 years of HES data (linked with NICOR data) for use in a research study quantifying heart failure patients who have to deal with repeat readmissions and evaluating the risk factors for repeat readmissions.

*Please note: KiTEC (King’s Technology Evaluation Centre) is a part of King's College London (KCL), based in the School of Biomedical Engineering & Imaging Sciences. The King’s Technology Evaluation Centre (KiTEC) is a collaboration between several King’s College London departments and the Medical Physics Department of Guy’s and St Thomas’ NHS Foundation Trust (GSTT) Kings Collage London are the Data Controller for KiTEC. All analysis will be done within KiTEC and no other part of KCL will be involved in this project. All KiTEC staff are employed by KCL and KiTEC is covered by KCL’s employers, public and professional liability insurance. For clarity, the term KCL has been used in all instances.

KiTEC is King’s College London’s Health Technology Assessment Centre (HTA) with experience in carrying out Medical Technology (MedTech) evaluations. It is a collaboration between the King’s College London School of Biomedical Engineering & Imaging Sciences, the School of Population Health and Environmental Sciences, and King’s Health Economics.

The lawful processing under GDPR is Article 6(1)(e) ‘processing is necessary for the performance of a task in the public interest and Article 9(2)(j) ‘processing is necessary for archiving purposes in the public interest, scientific or historical research purposes or statistical purposes…’ as the data are required for research purposes in the public interest and judged subject to the appropriate safeguards.

The National Institute for Cardiovascular Outcomes Research (NICOR) based within Barts Health NHS Trust has been added as a Data Processor as they will be receiving from NHS Digital a bridging file of pseudonymised HES-IDs derived from the NICOR cohort Patient Identifiable data in order for NICOR to create a data set for KiTEC to link to the 5 years of pseudonymised HES Data being provided directly to KiTEC from NHS Digital.

Heartfelt Technologies (a commercial company based out of the University of Cambridge) expressed mutual interest to KCL (as part of a workshop on innovation organised by the Health Foundry) regarding this project with the aim of establishing an accurate picture of repeat readmissions in heart failure patients, and the costs incurred by avoidable readmissions. Heartfelt Technologies has developed an in-home monitoring device for heart failure patients. Heartfelt Technologies wish to understand various aspects of the likely target patient population, in order to ascertain which patient groups are most likely to benefit from the technology, and in which patient groups the greatest avoidable costs arise that their technology may assist in reducing. KCL agree that this is a useful line of enquiry but also believe that many other research questions can be answered using the requested data. The primary goal of this research is to provide health benefit to heart failure patients and KCL believe that several journal articles can be developed using the requested data.

KCL briefly searched existing literature discovering that more comprehensive answers could be found by using registry data, specifically by linking the HES and NICOR datasets. Due to the significant cost burden heart failure admissions place on the NHS, it is expected that KCL's research - if able to identify a sub-population responsible for repeat hospital admissions - will identify an unmet need with potential savings leading to a benefit to health and social care. The outcome of the publication will go as far as identifying this unmet need and not a solution or a product for it. It is therefore, not related to any commercial gain for Heartfelt Technologies.

Heartfelt Technologies have stated that they will not be the data controller and will have no involvement in the analysis or processing of the data. Heartfelt Technologies will act as research collaborators in an advisory role only (having expressed mutual interest to KCL to start this project), providing input on statistical analysis methodologies, but will not have access to the data, do not have an active decision-making role and KCL will remain in complete control of the data analysis plan. Moreover, although Heartfelt have provided their thoughts on a potential statistical analysis, KCL have developed the analysis plan independently to answer multiple research questions. KCL’s statistical analysis plan will both inform about current rehospitalizations in subgroups of populations and generate answers to other questions of interest to KCL. For example, KCL will also look to build a predictive model to enable the accurate prediction of short term re-hospitalisation (and death) during the first year from the start of the follow-up process.

KCL will not deliver an internal report to Heartfelt Technologies; the only output will be a publication (or publications) in an open access journal. These publications will focus entirely on the unmet need of identifying the distribution of and risk factors for repeat readmissions in heart failure patients and prediction of hospitalisations. The publication will not focus on any technology and will be of no more relevance to Heartfelt Technologies than any other company or research group working on the field of heart failure. Heartfelt Technologies hope to find out it if there is a clinical need to be addressed by the technology which will help these patients and the NHS.

Heartfelt Technologies have agreed to fund the costs of obtaining and linking the data only and KCL will not receive any other funding for this project. Heartfelt Technologies will benefit from this research as it will allow them to focus their device to the patient groups that will find it of most benefit. This will also allow for more accurate cost modelling. This is the first project of this kind KCL have undertaken and the benefit to KCL is in demonstrating their ability to carry out this kind of research, plus generating high quality articles for publication. KCL are named as Data Controller and Data Processor. No KCL staff are receiving any reimbursement from Heartfelt Technologies for analysing the data.

The study will test the hypothesis that there are specific risk factors associated with repeat readmissions. It is KCL's hypothesis that a minority of patients account for the majority of re-hospitalisations, and therefore are at higher risk for worst clinical outcomes and the majority of the cost implications. Real world data are becoming an important source of information for patient outcomes and combining two complementary registries is a novel way to conduct research. There is a notable lack of evidence on both the frequency and distribution of readmissions, particularly in the UK, and very little information on risk factors for readmissions.

By identifying risk factors, it will become apparent which, if any, are avoidable. It will then be possible to calculate the potential cost savings associated with preventing such avoidable admissions. Heartfelt Technologies research questions include:

• What is the distribution of hospital admissions on a per-patient basis over the average life expectancy for this patient cohort?

• What are the objective patient selection criteria that capture the maximum number of repeat-admissions and minimum number of single-admission patients?

• What is the distribution of further admissions following that selection point?

• What are the symptoms, or clusters of symptoms, that correlate between hospitalisation events on a per-patient basis?

• What is the clinical utility of peripheral oedema monitoring in predicting heart failure readmission?

This study is the first of its kind – there has been no preliminary or background work for this research. The study is not part of a bigger study.

HES data is required, across a 5-year period, in order to identify patients who, have multiple admissions for heart failure, as well as the distribution and dates of admissions. NICOR offers classification of peripheral oedema which can be used, along with other clinical parameters, to identify patients that are more likely to be readmitted for heart failure. Linking this data will allow us to determine whether oedema monitoring can reduce emergency readmissions for heart failure. KCL will also quantify which population characteristics influence the chance of re-admission for heart failure. Quantifying the impact of age, sex and comorbidities is vital to ensure that support for patients with heart failure is appropriately targeted. Controlling for comorbidities and case-mix also allows KCL to quantify the impact of differences in socioeconomic status on access to care and outcomes. They will then determine the costs associated with patient subgroups. The research will identify groups that would benefit from existing NHS pathways to improve health outcomes in the most vulnerable and the cost implications of changes to care pathways.

As the research involves health data, which is included in the definition of special categories of personal data, it requires an additional condition for processing. Based on guidance, for health research this is article 9(2)(j), which details that processing is necessary for scientific and research purposes, subject to appropriate safeguards.

The legal basis for the flow of identifiable data from NICOR to NHS Digital are section 251 support.

HES DISCLOSURE CONTROL / SMALL NUMBER SUPPRESSION

In order to protect patient confidentiality, when presenting results calculated from HES record level data, outputs will contain only aggregate level data with small numbers suppressed in line with HES Analysis Guide. When publishing HES data, you must make sure that:

· cell values from 1 to 7 are suppressed at a local level to prevent possible identification of individuals from small counts within the table.

· Zeros (0) do not need to be suppressed.

· All other counts will be rounded to the nearest 5.

Data will not be made available to any third parties other than those specified except in the form of aggregated outputs with small numbers suppressed in line with the HES Analysis Guide.

KCL submits applications to NHS Digital and NICOR in parallel. The reference for the NICOR application is HQIP295 (18-HF-01).

STEP 1: NICOR send the Patient Identifiable Data (NHS Number, Date of Birth, Postcode) plus a study specific ID for the specific study cohort to NHS Digital. The NICOR study cohort contains only patients who are aged over 18 years old with heart failure in England between 2013/14 and 2017/18. The NICOR study cohort will contain approx. 81,000 people per year for 5 years.

STEP 2: NHS Digital

2a: Creates a bridging file of the pseudonymised STUDY_ID for the patients included in the NICOR cohort provided above and send this to NICOR.

2b: Creates a full set of 5 years of pseudonymised HES data from the NICOR Patient Identifiable Data and sends this to KiTEC.

NB: (separately to this NHS Digital agreement) NICOR will use the pseudonymised STUDY_IDs to create a full set of 5 years of pseudonymised NICOR data and send this separately to KiTEC.

KiTEC will therefore receive two separate data sets which will be linkable only via the pseudonymised STUDY_ID and therefore individual patients cannot be identified.

KiTEC does not have the ability to re-identify any individuals within the dataset and will make no attempt to re-identify individuals.

The data will not be made available to any third parties except in the form of aggregated outputs, in the form of academic paper(s) published in an open access journal, with small numbers suppressed in line with the HES Analysis Guide.

KCL is requesting linked HES Admitted Patient Care and NICOR data, linked using sensitive fields (NHS number, date of birth, post code) which have section 251 approval as documented in the separate NICOR application - HQIP295 (18-HF-01). No sensitive or identifiable fields will flow to KCL: the linked HES-NICOR data set is pseudonymised.

KCL requires linked HES-NICOR data from the years 2013-14 to 2017-18 (5 years), filtered by condition: heart failure patients (where 4-character codes i50.0-i50-9 appear in the diagnosis fields (DIAG_4_CONCAT)); additionally, linking HES data to the NICOR dataset will automatically filter the cohort by condition because all patients in the NICOR dataset have the codes i50.0-i50-9. KCL require this length of time in order to adequately cover the period in which numerous repeat readmissions might occur. Note that the data requested has not been minimised to exclude one-off admissions (i.e. patients who are not readmitted) because KCL requires data on these patients in order to perform comparisons between them and patients who are readmitted, either once or multiple times.

Existing literature on the topic has employed a similar length of follow-up, i.e. 5 years (Leva, F., et al. (2017). "Multi-state modelling of repeated hospitalisation and death in patients with heart failure: The use of large administrative databases in clinical epidemiology." Statistical Methods in Medical Research 26(3): 1350-1372). KCL require national data because the study is a national one and is not limited to specific geographical areas; in addition, the study will consider whether geographical region is a potential risk factor for repeat readmissions.

All organisations party to this agreement must comply with the Data Sharing Framework Contract requirements, including those regarding the use (and purposes of that use) by “Personnel” (as defined within the Data Sharing Framework Contract ie: employees, agents and contractors of the Data Recipient who may have access to that data).

Redcentric PLC [also known as Redcentric Managed Solutions, and Redcentric Solutions Limited as well as Redcentric PLC (Harrogate)] are NOT able to access the data stored on NICOR (Barts Health NHS Trust) servers. The data servers are managed by NICOR staff, including all server maintenance, backups etc. Redcentric only provide the secure facility (bricks and mortar) with power and internet connectivity for us to house our servers. Servers are in a secure locked environment which Redcentric do not access. Therefore, any access to the data held under this agreement by Redcentric Ltd would be considered a breach of the agreement. This includes granting of access to the database[s] containing the data. Barts Health NHS Trust staff have access to the data via remote access.

The back-ups are managed entirely by NICOR (either remotely or on site) and are not conducted by the data centre. The data centre functions only in providing an IT infrastructure.

There will be no data linkage undertaken with NHS Digital data provided under this agreement that is not already noted in the agreement.

Data will only be accessed and processed by substantive employees of Kings College London and will not be accessed or processed by any other third parties not mentioned in this agreement.

Academic papers will be published open-access in a high impact cardiology journal (such as The Journal of the American College of Cardiology (JACC): Heart Failure) on KCL's methodology, analyses and results, on the impact of repeat readmissions on costs and the attendant risk factors for repeat readmissions. The intention is to publish early in 2024, approximately eighteen months after the receipt of the data, and there is no plan to publish interim results.

For each paper published, a short presentation is developed to summarise the findings for a range of stakeholders, including health technology assessment practitioners and the funder, Heartfelt Technologies. Findings will be presented at the Health Technology Assessment International (HTAi) Annual Meeting in Beijing.

The KiTEC website provides links to the open access papers and summaries of findings. Publications are also listed on individuals’ KCL Pure pages. KCL Pure is KCL's public-facing Research Portal [https://kclpure.kcl.ac.uk/portal/en/], where the public can find out information about KCL's research such as funding, researcher biographies and outputs such as published books and peer-reviewed journal articles.

All publications and conference presentations are promoted on LinkedIn, via the KiTEC account.

All outputs will contain only data that is aggregated with small numbers suppressed in line with the HES Analysis Guide.

Study findings will be instructive to clinicians treating heart failure patients in terms of highlighting risk factors for avoidable repeat readmissions. There is currently a lack of evidence on risk factors for readmissions. These findings will also highlight possible opportunities for remote care and telehealth initiatives that could prevent readmissions. The logical sequence of events would include the effective dissemination of KCL's findings, an interpretation of them leading to changes in clinical practice, and finally the adoption of new clinical practice in treating patients with heart failure.

It is reasonable to expect that clinicians will act on these findings; KCL will disseminate the key findings from the publication in relevant departments of the university (KiTEC, the team carrying out this project within KCL is part of the School of Biomedical Engineering & Imaging Sciences and works closely with cardiologists from Guy’s and St Thomas’ NHS Trust). KiTEC also has a track record of presenting research at conferences (e.g. the Health Technology Assessment International annual meeting) and is planning to do the same with the results of this study. This is in addition to the publication that will be in a relevant high-impact open access journal. KCL's patient engagement work has shown that the project is of interest to patients and the results will be welcomed. The key findings will be disseminated among relevant patient organisations.

Preventing avoidable readmissions would lead to substantial reductions in resource use and expenditure, as well as better patient experience and quality of life.

There are an estimated 81,000 people living with heart failure, with admissions accounting for an estimated 1 million inpatient bed-days (around 2% of the entire NHS total). Up to 70% of these costs are due to hospitalisation costs, so attempts to reduce costs must be focused on reduced admissions.

It is reasonable to expect that KCL's publications will be taken into consideration by NICE as evidence for the next update of the NICE clinical guideline on chronic heart failure in adults: diagnosis and management (NG106). NICE guidance is updated every 3 years by means of a systematic review (which would include our publication) and through means of open consultation (contributions to which would be supported by the publication). KiTEC is an external assessment centre for NICE and therefore has very close links to these activities, and is well placed to contribute to guidance updates.

KCL (KiTEC) have advertised on People in Research (a website run by the NIHR), looking to involve members of the public at all stages of the research. To promote the results and findings, a Plain English document will be distributed to interested members of the public. KCL have made use of social media (twitter, LinkedIn) in the past for such dissemination and will use these avenues again. KCL currently have two responses to the advert from patients willing to be involved in the research. This involvement will include reviewing of the study protocol and any outputs from the research, particularly the Plain English summary of results.

KiTEC staff have training in Patient and Public Involvement (PPI), run by the Wellcome Engineering and Physical Sciences Research Council (EPSRC) Centre for Medical Engineering (CME) and have access to public engagement resources through this centre and also within the school of Biomedical Engineering and Imaging Sciences at King’s College London.

Two patients have so far agreed in principle to be involved with this research. Both have reviewed the study protocol and provided feedback to the research team. In addition to these 2 patients, it is hoped that at least 3 more can be recruited to form a PPI consultation group. The members of this group will primarily participate in this research by reviewing the outputs of the research, to ensure their readability and understandability. Outputs will include at a minimum: a Plain English Summary of results, social media posts and information on KiTEC’s website (www.kitec.co.uk). If possible, the group will meet (or discuss over MS Teams), to discuss the findings of the research and request feedback from the group on other possible analysis or other possible ways to disseminate the research.

KCL are yet to receive the data and therefore no benefits have been realised as a result of the data requested in this Agreement.