Long-term Follow Up of Patients in the Birmingham and Lambeth Liver Evaluation Strategies (BALLETS) Study

The original Birmingham and Lambeth Liver Evaluation Strategies (BALLETS) study was a prospective study of people in General Practice who had abnormal Liver Function Tests (LFTs). These tests consist of a panel of six to eight individual tests. A cohort of 1,290 patients within the BALLETS study, recruited between 2005 and 2008, were found to have abnormal results, but with no clear clinical reason (or diagnosis) for this abnormality. This cohort was made up of patients with and without ‘fatty liver disease’. The other patients (not of interest in this study) had clear diagnoses of severe liver disease.

Non-alcoholic fatty liver disease (NAFLD) is the most common liver condition in high-income countries, affecting an estimated 25% to 30% of adults (although in the BALLETS cohort this was 40%, this cohort had presented with symptoms that the GP thought warranted an LFT).

Although if caught early NAFLD is thought to be reversible, NAFLD can progress to more severe liver disease. While the possibility of such progression is not in doubt, the risk of progression is unclear and disputed. There is good evidence that the incidence of cirrhosis has risen over the last few decades, that it is linked to diet, and that fatty liver has overtaken alcohol as the most common cause of cirrhosis [2]. The gradual accumulation of fat in the liver leading to cirrhosis of the liver is the main driver of the increasing incidence in serious liver disease seen over the last decades.

It is important to provide scientific evidence regarding the actual scale of the risk of progression as it has been recommended that all people with NAFLD should be investigated with regularly repeated blood tests and scans of the liver[3]. The implementation of such a policy would have not just a large financial cost, but also may have a cost to the general well-being of patients under regular testing.

To date, follow-up studies have been carried out on highly selected populations, who have already tested positive on supplementary tests [4]. A systematic review of cohort studies involving people with paired liver biopsies at least one year apart, carried out by Singh, et al. (2015), included only 131 patients [5]. The rate of progression seen in this cohort is likely to be an overestimate on the grounds that, patients being given repeated liver biopsies are likely to have had more worrying features than those of the typical NAFLD patient seen in primary care.

To summarise, the existing literature on NAFLD cohorts is likely to be subjected to selection effects and hence to over-estimate the risk of progression for a more ‘routine’ group of people. The BALLETS study provides a population of patients that might typically be identified in the course of routine practice in primary care. The population has been comprehensively screened to exclude serious liver diseases and to identify factors that may be associated with the more or less rapid progress of fatty liver cases. These risk factors include the degree of fatty infiltration, Body Mass Index (BMI), and alcohol intake. The study has an ‘internal control’ in the form of BALLETS patients who do not have fatty livers (accepting that their prognosis may be worse than that of people of the same age and sex who have not been tested). Lastly, the study team believes that their proposed study would be much the largest cohort study of people with NAFLD.

University of Birmingham (UoB) holds the BALLETS study data set and will supply University Hospital Birmingham NHS Foundation Trust (UHB) with the required identifiers (NHS number and Date of Birth). UoB will not take part in the analysis of the record level pseudonymised data which will be received by UHB from NHS Digital. UHB will provide the analysis and will provide UoB with an anonymous summary of results and analysis to UoB, please see processing details for further clarification.

The Study Team at the University of Birmingham (UoB) aims to describe the risk of progression to severe liver disease in patients with NAFLD over a medium to long-term follow-up period and understand whether this cohort is at an increased risk of death.

To achieve this aim the UoB Study Team wants to understand any interactions with secondary care relating to liver disease and the survival of patients in these two groups and is requesting follow-up data on:

1. Death (with or without liver disease as the cause of death)

2. Inpatient admission: for liver or liver-related disease.

3. Outpatient attendance: for liver or liver-related disease.

The data sets being requested are:

Civil Registration (Deaths) extract

Hospital Episode Statistics (HES) Admitted Patient Care data set (APC)

Hospital Episode Statistics Admitted Outpatient data set (HES OP)

The severe liver disease will be at least identified if not actively managed in secondary care settings so the above data elements will allow the UoB Study Team to identify any progressions to severe liver disease, any deaths or other liver-related conditions which require investigation or management in a secondary healthcare setting.

The UoB study team will not diagnose any conditions the patients are unaware of. Neither UoB nor University Hospitals Birmingham NHS Foundation Trust (UHB) will contact or re-identify any patients. Any outputs will be published only in an aggregated format with small number suppression applied as per the HES Analysis guidance.

The UoB study team believes that this research is of importance to the general public as NAFLD is a condition thought to affect at least one in four adults in the UK. They believe it is important that these patients, when first identified in primary care are given realistic and reliable information about their future risk of severe liver disease and survival. When fully informed about the potential risks patients may be more proactive in their treatment or monitoring plans. The results of this research will provide the evidence to determine whether routine assessments of the patient's liver well-being are justified. As a secondary benefit of this work, UoB will be able to provide reliable information about the use of hospital liver services that patients with NAFLD will have, allowing hospitals to more accurately estimate the availability of these services if the prevalence of NAFLD continues to rise. The study team is requesting the minimum data necessary to be able to achieve their aims. Obtaining any inpatient and outpatient visits to the hospital in this cohort will allow the team to identify whether a patient has progressed to severe liver disease. UoB will also be able to understand how often patients are accessing hospital liver services both before and after a diagnosis of severe liver disease. The death data will allow UoB to understand the mortality in this cohort, and provide estimates of survival at regular time points following a diagnosis of NAFLD. The UoB Study Team has also requested the cause of death to understand if any of the patients are dying as a result of any liver disease.

These datasets are required in order to identify whether patients involved in the original BALLETS study have had any of the outcomes identified above. UHB require extracts of the dataset outlined above specifically for the cohort they are studying. Record-level, identifiable data are required in order to link with the dataset collected for the original BALLETS cohort. UoB is not asking for any direct identifiers that they do not already hold. The outcomes of interest may occur after several years and UoB has a long follow-up time since the original study. Therefore, UoB is requesting data from 2007/08 to the present to allow them to follow up patients from the time of the original study to the present date.

The original BALLETS study was funded by the National Institute for Health Research (NIHR) Health Technology Assessment programme at a cost of over £1m. Eight primary care practices in Birmingham and three in Lambeth contributed to the study. To be eligible, patients had to have no known liver disease and at least one abnormal test on their LFT panel. Recruitment took place between 2005 and 2008, and patients were followed up for two years following recruitment.

The current study is planned to be a standalone follow-up to the original BALLETS study and will follow up on 1,237 patients who received an abnormal result in their LFT and specific, severe causes (hepato-cellular disease, hepatobiliary disease, and tumour) of liver disease were ruled out. Ninety-eight percent of patients in this category went on to have ultrasound scans and NAFLD persisted at two years in the majority (67%) of patients when follow-up scans were carried out. Patients who were found to have evidence of NAFLD will be the treatment group in the study and those without will be the control group.

The patients recruited for the original study were registered at GP practices in Birmingham and Lambeth. However, given that so many years have passed, many of them are likely to have moved or to have attended other hospitals around the country for various reasons. Therefore, it is necessary to request data for the whole of England to minimise the number of hospital interactions and civil registration deaths that are missed. UoB is also only asking for a small subset of the datasets in order to answer achieve its aims.

UoB only needs records of hospital admissions and attendances relating to liver disease for the cohort of patients identified by the BALLETS study. Therefore, they have minimised the data request by only asking for admissions and attendances with a diagnosis (ICD-10*) code K70-K77 or B15-B17 or the treatment specialty was liver-related.

*The ICD-10 (International Classification of Diseases, Tenth Revision ) is a system used by physicians and other healthcare providers to classify and code all diagnoses, symptoms and procedures recorded in conjunction with hospital care.

UoB have requested patient-level data with some identifiable elements (such as date of death). Record level information is needed to understand how many contacts a patient has with hospital(s) for liver related disease, and the course of their treatments. Date of death is needed to calculate survival time from a number of time points collected in the original BALLETS study.

Due to the length of time since the start of the study, it would not be practical to follow-up patients manually. There are multiples reasons for this including:

1. This would be very costly.

2. Many patients will have moved and would, therefore, not have their contact details

3. Some patients may have died. Sending information may be distressing for family members.

4. Issues with patient recall – patients may not remember exactly when they were diagnosed with severe liver disease, nor all of their contacts with hospitals, thus determining risks of severe liver disease at certain time points would not be feasible.

The sole Data Controller is University of Birmingham and the Data Processor is University Hospitals Birmingham NHS Foundation Trust. There are no other organisations involved in the project. The study will be funded by the National Institute for Health Research (NIHR) Applied Research Collaboration (ARC) West Midlands.

LEGAL BASIS

GDPR Article 6(1)(e) – “processing is necessary for the performance of a task in the public interest or in the exercise of official authority vested in the controller”

The University of Birmingham is classed under the Data Protection Act 2018 as a Public Authority (see s7(1)(a) defines ‘public bodies’ for the purpose of the GDPR as “a public authority as defined by the Freedom of Information Act 2000”). The FOI Act Schedule 1 Section 53(1)(b) states that 'a university receiving financial support under section 65 of the Further and Higher Education Act 1992' is considered a public authority, and the Higher Education Funding Council (now the Office for Students) names the University of Birmingham as one of the institutions to which they provide funding. The Data Protection Act 2018 Part 2, Chapter 2 Section 8(e) states that ‘the exercise of a function of the Crown, a Minister of the Crown or a government department’ has a basis in law for Lawfulness of processing. The University of Birmingham falls into this category as they have a Royal Charter which includes "The University shall be both a teaching and an examining University and shall further the prosecution of original research." Furthermore, NHS Digital DARS application assessment team are satisfied that this request is appropriate, necessary and proportionate for the performance of the task described in the Purpose statement and a task in the public interest.

Additionally, as Health data is a special category of Personal data, University of Birmingham will be using GDPR Article 9(2)(j) - “processing is necessary for archiving purposes in the public interest, scientific or historical research purposes or statistical purposes in accordance with Article 89(1) based on Union or Member State law which shall be proportionate to the aim pursued, respect the essence of the right to data protection and provide for suitable and specific measures to safeguard the fundamental rights and the interests of the data subject” as the data are required for research purposes in the public interest (meeting the conditions in the DPA 2018 Schedule 1 Part 1 (4) - which GDPR Recital 52(2) determines is an appropriate derogation from the prohibition on processing special categories of personal data) and data processing is subject to appropriate safeguards (- In accordance with GDPR Article 89(1)) .

During the study Patient and Public Involvement (PPI) discussion, their public contributors agreed that long-term follow-up of patients on the BALLETS Study would help the NHS make better decisions about how to manage care for patients with abnormal liver function tests. Several contributors agreed that if they had consented to be part of the BALLETS study, as explained in the Patient Information Sheet, they would ‘expect’ to be followed-up. The public contributors stated that the methods of follow-up were acceptable. In particular, public contributors highlighted the fact that the methods were ‘non-intrusive’ and would ‘not cause any harm’ that consented to take part.

REFERENCES

[2] Wree A, Broderick L, Canbay A, Hoffman HM, Feldstein AE. From NAFLD to NASH to cirrhosis-new insights into disease mechanisms. Nat Rev Gastroenterol Hepatol. 2013;10(11):G627-36.

[3] Younossi ZM, Koenig AB, Abdelatif D, Fazel Y, Henry L, Wymer M. Global epidemiology of nonalcoholic fatty liver disease-Meta-analytic assessment of prevalence, incidence, and outcomes. Hepatology. 2016;64:73-84.

[4] Vuppalanchi R, Siddiqui MS, Van Natta ML, Hallinan E, Brandman D, Kowdley K, Neuschwander-Tetri BA, Loomba R, Dasarathy S, Abdelmalek M, Doo E, Tonascia JA, Kleiner DE, Sanyal AJ, Chalasani N; NASH Clinical Research Network. Performance characteristics of vibration-controlled transient elastography for evaluation of nonalcoholic fatty liver disease. Hepatology. 2018; 67(1):134-144. doi: 10.1002/hep.29489.

[5] Singh S, Allen AM, Wang Z, Prokop LJ, Murad MH, Loomba R. Fibrosis progression in nonalcoholic fatty liver vs nonalcoholic steatohepatitis: a systematic review and meta-analysis of paired-biopsy studies. Clin Gastroenterol Hepatol. 2015; 13(4):643-54.e1-9. doi: 10.1016/j.cgh.2014.04.014

METHODOLOGY

1. The University of Birmingham (UoB) holds the BALLETS study data set. As per the Confidentiality Advisory Group (CAG) section 251 advisory guidance and reference 20CAG0071, UoB will send an identifiable data set (NHS Number and Date of Birth) to UHB.

2. UHB will then extract an identifiable study cohort of 1,237 individuals (Study ID, NHS Number, Date of Birth, First Name, Last Name) and send to NHS Digital via the Secure Electronic File Transfer Service (SEFT). The cohort of 1,237 individuals treated in England are filtered by admissions and attendances with a diagnosis (ICD-10) code K70-K77 or B15-B17 or the treatment specialty was liver related, 2007/08 to 2021/22.

3. NHS Digital will use the cohort to link to HES and Mortality data, and create:

- a set of HES pseudonymised record level extracts and sent via SEFT to UHB.

- an automated record level Mortality report including identifiable data and send via SEFT to UHB.

4. UHB will then link NHS Digital data extract to the BALLETS study data using the Study ID, and then analyse the data as per the study protocol. UHB will then aggregate and suppress any small numbers and send the anonymous summary and results of the analysis to University of Birmingham. Anonymised data will be suppressed in line with the HES analysis guide.

UHB will link only the record-level identifiable data from NHS Digital to the BALLETS study data. No other linkages are permitted. The University of Birmingham will not attempt to re-identify individuals from the output they receive from UHB.

Data Minimisation

The application was submitted with the methodology that University Hospitals Birmingham NHS Foundation Trust will provide a cohort (disseminated to NHS Digital under Section 251) to NHS Digital which will link to HES OP and APC data, and NHS Digital will return the HES identifiable record-level data extracts (Study ID plus NHS Number). The Study protocol, Ethics approval and section 251 approvals all agree that this methodology is acceptable.

However, in order to further minimise the data requested and pseudonymise the HES data, NHS Digital will remove the NHS number and return only the study ID.

This linkage will enable UHB to access the baseline information required for analysis (such as the BMI, age, sex test results of the patients and relevant dates.

All processing of record-level data received from NHS Digital will be performed by substantive employees of UHB. When the analysis is complete, aggregated results with small numbers suppressed from the analysis will be shared with UoB and included in a jointly written peer-reviewed publication. The aggregated results will have small numbers supressed in line with the current version of the HES Analysis Guide.

When UHB receives the requested data from NHS Digital, it will be linked with the data set for the original BALLETS study. The association between each outcome and the presence (and degree) of Fatty Liver in the BALLETS study will be investigated using appropriate statistical methodology (e.g. logistic regression) adjustments will be made for baseline risk factors including BMI, alcohol intake, age, sex and ALT. A survival analysis will also be undertaken on all cause, and liver-related mortality by UHB. The results of these analyses will be shared with collaborators from UoB (ensuring that small numbers have been supressed).

All employees at UHB must undergo annual information governance training, which includes passing the subsequent test.

The data are both processed and stored on a secure SQL* server which sits within the Trust's IT framework, and is not available for querying by any unauthorised external parties or tools. Passwords are strictly controlled by UHB IT services, with access permissions for each user administered on the authorisation of UoB through an identified senior member of staff at UHB. The physical servers are located on site at UHB in a locked storage room where entry in and out of the rooms is logged and controlled by senior managers.

*SQL is a domain-specific language used in programming and designed for managing data held in a relational database management system, or for stream processing in a relational data stream management system.

Statistical data analysis will be carried out via UHB owned remote device connected to the UHB network either directly in person or remotely, using an appropriate statistical package (either R or STATA). To remotely access the devices requires a secure 2-factor authenticator (VPN) and users are then able to securely access the secure SQL server on the Trust’s IT framework. All data analysis will be conducted within the confines of the Trust’s secure server, and will not be downloaded to remote devices for storage or processing.

UHB use datacentre-as-a-service (co-location) from Crown Hosting, a HM Government-approved datacentre facility, supplied by ARK Data Centres. The server and storage equipment is Trust owned and operated via a privately dedicated link to UHB. IT Services span the local network to the servers and storage hosted at ARK Data Centres, accessing data using the same protocols that UHB do with servers and storage at their other location. Therefore, ARK Data Centres do not access data held under this agreement as they only supply the building. Therefore, any access to the data held under this agreement would be considered a breach of the agreement. This includes granting of access to the database[s] containing the data.

UHB will undertake processing activities on behalf of UoB. UHB may provide expertise and guidance about the processing activities, but the ultimate decision-making lies with UoB. UHB will undertake no processing activities in relation to this application without the explicit approval of UoB.

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.

UHB will assist UoB on a paper including the aggregated, small number supressed (in line with the HES analysis guide) results of the data analysis with a view to submitting to peer-reviewed journals. The study team feel that the results of this study will be a good fit for the BMJ, and that is where they aim to first submit. Some of the results may be presented at local meetings.

Local meetings include regular seminar series organised and run by both the University of Birmingham and University Hospitals Birmingham NHS Foundation Trust. These include:

• Monthly seminar series organised by Institute of Applied Health Research, University of Birmingham. These seminars are available to staff in the University of Birmingham and advertised through University of Birmingham communications channels.

• Seminar series organised by University Hospitals Birmingham NHS Foundation Trust. These health seminars are open to members of the University Hospitals Birmingham NHS Foundation Trust (patients, carers and staff) and are free to attend.

• Seminars at the Institute of Translational Medicine. Seminars@ITM are cross-discipline seminars designed to showcase the research conducted across the Birmingham Health Partners campus (University Hospitals Birmingham NHS Foundation Trust, University of Birmingham, Birmingham Women’s and Children’s NHS Foundation Trust)

This agreement for this current study on the BALLETS project is a standalone follow-up to the original BALLETS study. The study team have involved members of the public from the outset. (By involvement, the study team refer to the NIHR definition: doing research ‘with’ or ‘by’ members of the public rather than ‘to’, ‘about’ or ‘for’ them).

To clarify, the study team have not involved people who participated in the original BALLETS project in this study. This is a standalone project and they recruited a new group of public members to discuss their views and perspectives on the long-term follow up of patients that consented to participate in the original study. Their insights confirmed that the BALLET study’s aims were relevant to patients and the study methods were acceptable from an ethical perspective.

Members from the Patient Involvement discussion groups were recruited from the following local groups:

• People involved in shaping local NHS Services (e.g. members of local Patient Participation Groups, public members of Clinical Commissioning Groups, members of Hospital Trust patient groups).

• People involved in research through local NIHR Centres (e.g. public advisors in NIHR West Midlands Biomedical Research Centre (WM BRC), Applied Research Centre West Midlands (ARC WM), Clinical Research Network West Midlands (CRN WM).

• Local Patient Research Ambassadors, a group convened by the West Midlands Clinical Research Network.

• People attending local patient groups (e.g. Liver and GI group, 1000 elders).

Members of the discussion group agreed to be kept informed of progress of the project and expressed a willingness to be involved in discussing the research findings, highlighting messages most important to them, and helping share the message with relevant communities.

Ways in which the study team could indirectly share the results of the study with the initial participants of the study were discussed.

Public contributors agreed that it would be important to share the research findings with people that consented to take part in the original study. It was agreed that the key findings need to be shared sensitively (informative but not ‘alarmist’) and in accessible formats. Several strategies were identified as ways to share the data in a meaningful way:

• Providing information about the BALLETS project on a website. It was agreed that the NIHR ARC WM could host information about the project.

• Providing information to GP Practices that were recruited to be part of the original study.

• Sharing findings through relevant charities.

Public contributors involved in the discussion group wanted to be involved in future discussions about sharing findings from the study future discussion group once data analysis has been completed have been agreed upon.

The Study team will work with their public contributors to co-produce materials that are in formats that are accessible. All outputs of data will be aggregated with small number suppression applied as per the HES analysis guide.

Once the study team at UoB have analysed the data, they hope to work with their contributors to identify relevant local and national organisations (including charities) to share their findings with. This would most likely include: British Liver Trust and Liver Research Foundation.

The study team hope to complete the analysis within six months of receiving data from NHS Digital. However, as an insurance against delay, the study team seek permission to hold the data for up to 18 months from receipt of the data from NHS Digital. This will enable them to complete their purpose of following up the patients from the BALLETS study to find out if they have had liver disease sufficient to warrant hospitalisation or to have been a cause of death, and ensure that the study team are able to attend to any queries from peer reviewers when submitting the results.

As outlined previously, the study team believe that this research is of importance to the general public as NAFLD is a condition thought to affect at least one in four adults in the UK. They believe it is important that these patients, when first identified in primary care are given realistic and reliable information about their future risk of severe liver disease and survival. When fully informed about the potential risks patients may be more proactive in their treatment or monitoring plans. They hope that results of this research will provide the evidence to determine whether routine assessments of the patient’s liver well-being is justified. As a secondary benefit of this work the study team to hope be able to provide reliable information about the use of hospital liver services that patients with NAFLD will have, allowing hospitals to more accurately estimate the availability of these services if the prevalence of NAFLD continues to rise.

The study team hope the findings of the study will be of great value to current health and social care services. Since the BALLETS patients underwent a 'battery' of tests both when they entered the study and again after two years, the study team are able to characterise patients in detail, and are in a unique position to provide information that they hope be of great interest to patients and the public.

This is a standalone project which will not be used as part of post-graduate student’s studies.

The previous study used data that was collected specifically for the study and did not link to any NHS Digital data.