TRIANGLE HES Data Application

Anorexia nervosa (AN) is an eating disorder which has both psychiatric and medical features. Patients with the condition tend to show an inability or refusal to maintain a healthy body weight and instead try to keep their weight as low as possible through restricted eating and through other methods such as over-exercise. Patients often demonstrate an intense fear of gaining weight and a distorted perception of their body image. AN can develop into a severe, enduring mental disorder that is associated with increased death rates (Arcelus, Mitchell, Wales, & Nielsen, 2011), substantial physical (Mitchell & Crow, 2006) and psychological complications and adverse social consequences (Hjern, Lindberg, & Lindblad, 2006). Hospital treatment is increasingly used in the management of this stage of illness (Holland, Hall, Yeates, & Goldacre, 2016). There is uncertainty about many of the factors that might contribute to a good outcome during and following hospital care.

Current treatment guidelines suggest the first step of treatment for AN should be a form of outpatient psychotherapy. However, systematic reviews have found the overall outcomes are disappointing, dropout rates are high, and only a third of cases were classified as recovered at two years. Admissions for AN have increased over the last 15 years. UK hospital episode statistics (HES) data shows that between 2001 and 2009, for 8975 individuals with AN, there were 34166 discharges, a number which exponentially rises over time. These admissions are the most prolonged of all adult psychiatric admissions. In a census of child and adolescent services in 1999, 20% of all beds were taken up by AN, 41% of these were in the private sector, albeit funded by the NHS.

Transition in and out of hospital can be challenging. Relapse after discharge lies between 20-50%. A variety of strategies to reduce relapse have been tested in feasibility studies. Guided self-help approaches linking with community support using a variety of digital approaches hold promise in terms of improving patient symptoms and carer burden. Several studies from the researchers at King’s College London (KCL) who will be carrying out this study had a preliminary examination of cost effectiveness and measured bed use as one proxy marker of costs as bed costs contribute heavily to hospital treatment costs.

Recent pilot and feasibility studies using the ECHO ( acronym for “experienced carers helping others”, a carer intervention that informed hybrid intervention for patients and carers) & ECHOMANTRA (intervention combining ECHO with the Maudsley Model of Treatment for Anorexia Nervosa for patients, ECHOMANTRA) is being evaluated in the triangle trial) which are relevant for the current TRAINGLE study, show that self-reported bed use is reduced under these interventions. For example, patients who have had their carers take part in the ECHO intervention reported spending less time in hospital during their admission and fewer readmission rates (Hibbs, Rhind, Leppanen, & Treasure, 2015; Hodsoll et al., 2017; Magill et al., 2016)) and weekly weight gain (Adamson et al., 2019). Taken together these findings suggest that ECHO & ECHOMANTRA (a guided care management with moderate costs) can reduce service costs for a patient with AN.

ECHOMANTRA has been developed by the TRIANGLE research team and there is no commercial purposes associated with the intervention.

NHS Digital Request:

Researchers from KCL are conducting a multicentre randomised controlled trial (RCT) to examine whether the addition of a patient and carer skills-sharing intervention improves long-term patient wellbeing following hospital inpatient treatment for AN. This will be referred to as the HTA TRIANGLE (Health Technology Assessment TRransItion care in Anorexia Nervosa through Guidance onLine from peer and carer Expertise) research study throughout this application. Please see here for further information - https://triangle.slam.nhs.uk/.

HTA TRIANGLE examines the impact of ECHOMANTRA, a digital intervention that aims to augment inpatient care and reduce relapse by providing support for both patients and carers. The core hypothesis is that ECHOMANTRA might improve the transition from inpatient care to community care for adult AN patients and reduce subsequent bed use and mental health readmissions. The HES and Mental Health datasets requested from NHS Digital will be used to better understand bed use following the ECOHMANTRA intervention, as this is one of the secondary outcome measures in this RCT.

The HTA TRIANGLE study began in 2017. It uses a parallel group, superiority randomised control design with allocation of patients and their carers to treatment as usual or treatment as usual plus the guided self-management intervention.

Primary outcomes include the wellbeing of patients and carers. Secondary outcomes include bed use. Self-reported service use data is currently being collected from patients/carers and centres.

Researchers at KCL aim to supplement and validate this data from sources such as Hospital Episode Statistics (HES) and the Mental Health Services Dataset (MHSDS) to minimise the amount of missing data.

The multicentre RCT includes 370 dyads (patient and supporter (non-professional carer in the study e.g. family member, friend of the patient); 370 patient-carer pairs were recruited into the study and these pairs are referred to as "dyads"), the study includes females (n=346) and males (n=24). KCL researchers are following these patients for up to 18 months. Recruitment of this cohort finished in July 2020, and the 18-month follow up is due to be completed by March / April 2022.

There is a monetary incentive for this study. The monetary incentive for the study was up to £60 for each participant depending on how many assessments they completed (£10 per time point, and six time points over the course of the trial).The monetary incentive was approved by Camberwell St Giles Research Ethics Committee. Participants who received the ECHOMANTRA intervention during the study are able to continue accessing the self-help components of the intervention after they have participated for eighteen months at no cost. Participants who were allocated to the control group during the study are also given access to the self-help components of the intervention after they have completed eighteen months of their participation at no cost.

HES and MHSDS data will be used to obtain bed use for each participant enrolled in the TRIANGLE study, at 12 months, and up to 18 months post-randomisation to compare it with bed use for the year before the individual was recruited into the study. Bed days used in the year before recruitment will be compared with bed days used 18 months post recruitment.

This work will also contribute to the European Value of Treatment Project for Anorexia Nervosa (funded by European Psychiatric Association and the European Brain Council) which began in January 2019. This project aims to map care pathways and health economic data related to AN across 12 European countries. The TRIANGLE study will be included as a case study demonstrating a new pathway of care. The European Value of Treatment Project will not have access to NHS Digital data.

COHORT

Data subjects are individuals aged 16 or over, who were admitted to an intensive specialist unit (inpatient or day-patient), with a primary diagnosis of AN F50 (AN) or atypical or subclinical AN, with a BMI of ≤18.5 kg/m2.

The use of a younger cohort provides the opportunity examine whether there are subgroups who may respond differently to the intervention i.e., younger patients with a shorter duration of illness, compared to older patients with more severe and enduring AN.

At the time of consent to the trial, patients did not have another severe mental or chronic physical illness needing treatment, nor were pregnant.

Patients have been randomised on a 1:1 ratio to:

• Active Intervention: ECHOMANTRA + Treatment as usual

• Comparison intervention: Treatment as usual

All participants provided consent for the collection of data for 30 months, starting from 12 months prior to their date of randomisation into the trial and to 18 months following their randomisation date. This time window for data collection will allow for a longer-term analysis of whether the digital intervention reduces service use over time.

The sole Data Controller who also process NHS Digital data are King’s College London (KCL). Data processors have completed mandatory training in Data Protection Regulation and Good Clinical Practice training, and Research Confidentiality training provided by the Medical Research Council.

Researchers who are employed full-time at The London School of Economics and Political Sciences (Health Economist for the trial and Research Fellow in Health & Social Care) will be working jointly with the KCL researchers on the output of academic papers for TRIANGLE involving service use data collected both within the trial and through the HES and MHSDS. Both these London School of Economics and Political Sciences researchers have KCL visiting researcher appointments which included assumed data processing responsibility by KCL and approved by NHS Digital.

Clinical sites involved in the recruitment of cases

1. SOUTH LONDON AND MAUDSLEY NHS FOUNDATION TRUST

2. CHESHIRE AND WIRRAL PARTNERSHIP NHS FOUNDATION TRUST

3. MIDLANDS PARTNERSHIP NHS FOUNDATION TRUST

4. AVON AND WILTSHIRE MENTAL HEALTH PARTNERSHIP NHS TRUST

5. DORSET HEALTHCARE UNIVERSITY NHS FOUNDATION TRUST

6. CENTRAL AND NORTH WEST LONDON NHS FOUNDATION TRUST

7. BARNET, ENFIELD AND HARINGEY MENTAL HEALTH NHS TRUST

8. LEICESTERSHIRE PARTNERSHIP NHS TRUST

9. CUMBRIA, NORTHUMBERLAND, TYNE AND WEAR NHS FOUNDATION TRUST

10. SOUTH WEST LONDON AND ST GEORGE’S MENTAL HEALTH NHS TRUST

11. ESSEX PARTNERSHIP UNIVERSITY TRSUT

12. GLOUCESTERSHIRE HOSPITALS NHS FOUNDATION TRUST

13. BERKSHIRE HEALTHCARE NHS FOUNDATION TRUST

14. OXFORD HEALTH NHS FOUNDATION TRUST

15. SURREY AND BORDERS PARTNERSHIP NHS FOUNDATION TRUST

16. DEVON PARTNERSHIP NHS TRUST

17. TEES ESK AND WEAR VALLEY NHS FOUNDATION TRUST

18. CAMBRIDGE AND PETERBOROUGH NHS FOUNDATION TRUST

19. THE PRIORY HOSPITAL ROEHAMPTON

20. THE PRIORY HOSPITAL SOUTHAMPTON

21. THE PRIORY HOSPITAL ALTRINCHAM

22. THE PRIORY GROUP CHEADLE ROYAL

23. THE PRIORY HOSPITAL HAYES GROVE

24. ELLERN MEDE RIDGEWAY

25. ELLERN MEDE BARNET

26. THE CARDINAL CLINIC

27. ORRI EATING DISORDER CLINIC

28. NEW MARKET HOUSE

This research is funded by the National Institute for Health Research (NIHR) under its Health Technology Assessment Programme (Grant Reference Number: 14/68/09).

Data processing will only be carried out by substantive employees who have been appropriately trained in data protection and confidentiality.

The data will be stored on the “King’s College Military Health Research” (KCMHR) secure network and will be accessible from a single nominated machine within the eating disorders research unit (system access will be granted by KCMHR). The person(s) will be granted access to the data once they have completed the required training and have signed the KCMHR confidentiality agreement. King’s College London performs weekly secure offsite backups stored at a data centre in Slough. This procedure follows Cyber Security Essentials guidelines. All data stored on the KCMHR network is encrypted prior to these back-ups taking place. Only KCMHR IT management have access to the encryption keys.

King’s College London will supply the following information about each patient:

Study ID

NHS Number

Gender

Date of Birth

Date each participant was consented into the trial

CONSENT_1 - Earliest date patient information will be requested

CONSENT_2 - Latest date patient information will be requested

DATA FLOW

1. KCL researchers will create a file containing personal identifiers (i.e., study ID, date of birth, Gender, postcode) for consented TRIANGLE participants.

2. This file will contain a TRIANGLE Study ID, which is a personal identification number used by KCL researchers to identify consented TRIANGLE participants.

3. TRIANGLE study IDs will then be pseudonymised in a separate file to protect the confidentiality of the data; this pseudo-ID will be created for the purpose of sending data to NHS Digital to undertake linkage.

4. The TRIANGLE Study ID and pseudo-ID will be stored on an encrypted device within the KCMHR secure server.

5. KCL researchers will create a file with personal identifiers, the pseudo-ID, removing the original TRIANGLE Study ID and remainder of the cohort data.

6. This dataset will be sent securely by KCL researchers to NHS Digital.

7. NHS Digital will link the personal identifiers (i.e., study ID, date of birth, Gender, postcode) to HES and MHSDS data for the consented cohort.

8. KCL researchers will receive a pseudonymised linked dataset containing the requested HES, ECDS and MSDS variables with only a pseudonymised ID to ensure KCL researchers cannot identify patient records.

9. KCL will link this dataset containing HES and MHSDS data to the TRIANGLE pseudonymised cohort data using pseudo-ID as a linkage key, resulting in an analysis dataset containing pseudonymised HES, MHSDS and TRIANGLE cohort data.

10. The raw data will be stored on the KCMHR secure server till the DSA end date. Once the DSA ends KCL will securely destroy NHS Digital data and provide NHS Digital with a data destruction certificate.

There will be no subsequent flows of data.

DATA MINIMISATION

The following steps are taken to mitigate any risk of reidentification where data is linked:

• The pseudo-ID and TRIANGLE Study ID are stored on encrypted device in the KCMHR secure server.

• NHS Digital will remove personal identifiers from the HES and MHS dataset and send the pseudonymised dataset (including pseudonymised ID) securely back to KCL.

Once the HES and MHSDS data has been received by the data recipient at King’s College London, there will be no requirement/attempt to re-identify individuals.

Researchers from KCL will be using 5-character provider codes and site treatment codes to identify if care has been provided in the independent sector.

Researchers from KCL will provide NHS Digital with the consent date variables (this is the date each participant was recruited into the trial), to enable accurate windows of data to be provided for each participant. For all records researchers from KCL request data from each participant 12 months prior from their consent date and 18 months after their consent date.

Minimum datasets requested:

Researchers from KCL have requested four datasets:

Hospital Episode Statistics Admitted Patient Care,

Hospital Episode Statistics Accident & Emergency,

Emergency Care Data Set (as HES A&E ceased collection after 19/20) and the

Mental Health Services Data Set.

It is not relevant for the data to be narrowed further by clinical factors as the aim is to assess total length of stay in hospital for general admissions for any health problem. All patients’ episodes are required as researchers need to assess readmission rates. Maternity episodes are required, but the unborn child and neonatal records are not necessary, so these have not been requested.

Minimum patient records requested:

370 patient records have been requested, as that is the number of patients who have consented overall to participate in the research trial. This cannot be minimised further as a power calculation was used to inform researchers that approximately 370 patients were required to be able to statistically detect differences in total number of mental health contacts between the two study groups in the trial.

Minimum study time periods requested: 31/07/2016 – 20/01/2022

The period requested only spans the 30-month period for each participant in the trial, according to the consent that they have provided. Participants have provided consent for 30 months of their data to be used by researchers at KCL beginning one year prior to the date of randomisation using consent forms and an online method on the study website. The first patient in the overall cohort was randomised at 31/07/2017 – so data will be requested for participants starting from 31/07/2016 onwards, depending on each participant’s randomisation date. The last overall cohort participant was randomised at 20/07/2020 – so data will be requested up until 20/01/2022, as this date is 18 months post-randomisation. This period is necessary for researchers to assess the effectiveness of the intervention, as well as assess length of stay in hospital in the year prior to admission and eighteen months following randomisation into the trial.

Cohorts / Linkages:

Mental health service data will be linked to HES and ECDS data by NHS Digital. As the cohort were recruited into the trial from specialist eating disorder units it is very likely that there will be an associated mental health record for cohort patients. Therefore, the request to access and link both these data sets would be a logical and beneficial process to meet the outcomes of the trial. Furthermore, researchers at KCL could use these data relating to instances of service use with mental health services to calculate more accurately the costs associated with community-based MH provision for this cohort.

The overall aim of this research study is to inform clinical practice within the NHS. The proposed dissemination plan is multi-faceted and will target various audiences at different levels, acknowledging the wide range of stakeholder groups with an interest in the TRIANGLE intervention:

The study protocol (Cardi et al., 2017) was published by the study team in the European Eating Disorders Review in 2017 and is accessible here: https://onlinelibrary.wiley.com/doi/abs/10.1002/erv.2542. Qualitative findings have been published (Clark Bryan et al., 2020) and indicate the valued nature of remote support from the perspective of the TRIANGLE patient and carer cohort, during conditions of increased isolation and reduced access to eating disorder services (i.e. https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/erv.2762).

The research team will submit further scientific reports to peer reviewed journals for publication. Early secondary level academic papers from the baseline data are planned for submission in 2022, reporting on the comparison of the baseline variables and service use / carers of adult patients with severe AN in the early vs the late phase of AN- on service use.

The proposed late trial phase academic papers include:

(1) A Randomised Controlled Trial comparing a joint family intervention (Transition Care in Anorexia Nervosa Through Guidance Online from Peer and Carer Expertise (TRIANGLE)) with treatment as usual. 12 months outcome, cost effectiveness (to help address the primary outcome).

(2) A Randomised Controlled Trial comparing a joint family intervention (Transition Care in Anorexia Nervosa Through Guidance Online from Peer and Carer Expertise (TRIANGLE)) with treatment as usual. 18 months outcome, cost effectiveness (or perhaps a combination or Paper 1 and Paper 2).

(3) Predictors of service use. 12- & 18-months outcomes.

(4) Trajectories of illness following inpatient care.

Additionally, KCL aim to present their findings at hosted events such as eating disorder, psychiatric, student counselling and primary care conferences and to special interest groups within the professional bodies of the multiple disciplines involved in the care of this group of patients (nurses, psychology, occupational therapy, dieticians, social workers, etc.). These plans are hoped to be in place from 2023.

Data sharing with potential collaborators (UK and International) will be encouraged in an anonymised format, aggregated with small numbers suppressed. The researchers also aim to present the findings via online events such as presentations, webinars, online conferences etc. So far, during the study, the digital intervention has been presented to the Consultant Psychiatrist and his team from the Anxiety Disorders Residential Unit at the Bethlem Hospital (part of South London and Maudsley NHS Foundation Trust in the UK) who aim to develop a transition tool.

The carer component of the intervention has been used to facilitate carer groups within SLAM eating disorder services for adults via Zoom. The intervention is also being adapted to be tested in child and adolescent services in collaboration with the Clinical Lead for the Eating Disorder Service for Young People (Northwest Boroughs Health Care NHS Foundation Trust) and Honorary Clinical Lecturer, University of Liverpool.

Policy:

A full report with the executive summary will be sent to all NHS commissioning agencies. The co-applicant LG, CEO of the Beating Eating Disorders (BEAT) charity, will use this to assist in dialogues with policy makers, including MPs.

Patients and public:

Patients and the public are a core part of the research study and have been involved at all stages of the research process so far. BEAT will coordinate the patient and public involvement aspect of dissemination. Dissemination strategies will include emails updates to participating patients, carers and hospitals, presentations/blogs on eating disorder charities and carer and user websites, and communication channels (FEAST; Families Empowered and Supporting Treatment of Eating Disorders, BEAT, Student Minds, etc.) at yearly carers and users’ workshops, media articles, discussion forums, website postings, and schools’ training events. Training will be conducted through presentations, discussions, small group teaching and large group seminars. A summary of findings will be made available on the websites of the key charities and the Psychological Medicine website at the Institute of Psychiatry. All participants will also be sent newsletters and will have free admission to the carers conference.

Media:

The Press Offices at South London and the Maudsley NHS Foundation Trust and Kings College London will co-ordinate dissemination to the media. Researchers at KCL aim to present the results on radio stations Woman's Hour and All in the Mind, via academic webpages and via social media platforms such as Twitter (please see @kingsedresearch). The study team will also inform patient participation groups, newspapers, and neighbourhood organisations to disseminate findings in an anonymous form, aggregated with small numbers suppressed.

• KCL hope the outcomes of this research study to benefit patients and the NHS on multiple levels. Primarily, the study team hope to demonstrate providing online resources and support to patients and carers during the transition from inpatient to outpatient care can improve patient and carer wellbeing in the long-term. KCL also hope the support will improve patient eating disorder symptoms, weight, quality of life, importance, and confidence to change, work and social functioning and carer skills. KCL hope to demonstrate whether providing this support can impact on the number of days spent in hospital following randomisation. KCL hope the findings will help inform health and care decisions across the UK to address the problem of rising intensive readmissions for AN.

• The programme under evaluation is an intervention that has been developed by patients and carers, focusing on the impact of providing extra support online during the transition from intensive treatment back into the community. The intervention includes written materials, video-clips, and interactive online group forums which are delivered during the period of transition from intensive care to the community. The videos include face-to-face interviews with patients about experiences of inpatient treatment, transitions from hospital to home, and the roles that carers played in their recovery journeys. In these videos, patients provided recovery narratives, explicit behavioural change tips and anxiety management techniques, in core domains (cognitive, emotional, social and nutritional), which are frequently used as examples in the online group sessions. Videos include extensive interviews with an experienced carer (both personally and professionally) and carer skills videos which were scripted based on real-life experiences and produced in collaboration with the SUCCEED foundation.

Additionally, researchers at KCL have a patient and carer member on the Trial Steering Committee who have been heavily involved in the development of the intervention and progress of the trial and have attended regular meetings with the team to discuss participant recruitment strategies, participant communications, trial safety reporting, preparing progress reports to the funder, newsletters for participating patients, carers, and clinical centres, and assisting with dissemination plans. This patient member helped prepare email communications sent to the cohort, in order to invite them to re-consent for the collection of a longer period of HES and MHSDS data.

• Online support can be delivered via anonymous groups by post-graduate psychology students (as opposed to clinical psychologists) which could save money for the NHS.

• Combining the HES and MHSDS data with cohort data to explore what variables are associated with longer hospital stays (for example, if levels of depression might be linked to more days in inpatient units) is likely to help inform further treatments and early intervention services to better support patients.

• The outputs produced by the research team and co-applicants hope to achieve the stated purpose by reporting the health economics analysis, stating whether the digital intervention, when added to treatment as usual, was associated with fewer days in hospital In the 18 months following randomisation, compared to treatment as usual alone. These results are hoped to inform the decision to provide a digital intervention to patients and families to bridge the transition from hospital to home. This change may impact on all patients admitted to intensive treatment services and their families in the future.

• Furthermore, the specialised clinical commissioning group has shown interest in optimising the care pathway for this stage of illness and reducing bed usage if possible. Reduction in bed usage would presumably have improved outcomes, requiring fewer days of intensive support over time.

• Finally, this information will also be used to inform a European, value of treatment project, funded by European Brain Council and European Psychiatric Association (https://www.braincouncil.eu/projects/the-value-of-treatment/) which hopes to examine the cost effectiveness of treatment for various brain disorders.

Reference List

Adamson, J., Cardi, V., Kan, C., Harrison, A., Macdonald, P., & Treasure, J. (2019). Evaluation of a novel transition support intervention in an adult eating disorders service: ECHOMANTRA. International Review of Psychiatry, 31(4), 382-390.

Arcelus, J., Mitchell, A. J., Wales, J., & Nielsen, S. (2011). Mortality rates in patients with anorexia nervosa and other eating disorders. A meta-analysis of 36 studies. Arch Gen Psychiatry, 68(7), 724-731. doi:10.1001/archgenpsychiatry.2011.74

Cardi, V., Ambwani, S., Robinson, E., Albano, G., MacDonald, P., Aya, V., . . . Treasure, J. (2017). Transition Care in Anorexia Nervosa Through Guidance Online from Peer and Carer Expertise (TRIANGLE): Study Protocol for a Randomised Controlled Trial. Eur Eat Disord Rev, 25(6), 512-523. doi:10.1002/erv.2542

Clark Bryan, D., Macdonald, P., Ambwani, S., Cardi, V., Rowlands, K., Willmott, D., & Treasure, J. (2020). Exploring the ways in which COVID‐19 and lockdown has affected the lives of adult patients with anorexia nervosa and their carers. European Eating Disorders Review, 28(6), 826-835.

Hibbs, R., Rhind, C., Leppanen, J., & Treasure, J. (2015). Interventions for caregivers of someone with an eating disorder: a meta-analysis. Int J Eat Disord, 48(4), 349-361. doi:10.1002/eat.22298

Hjern, A., Lindberg, L., & Lindblad, F. (2006). Outcome and prognostic factors for adolescent female in-patients with anorexia nervosa: 9-to 4-year follow-up. The British Journal of Psychiatry, 189(5), 428-432.

Hodsoll, J., Rhind, C., Micali, N., Hibbs, R., Goddard, E., Nazar, B. P., . . . Todd, G. (2017). A pilot, multicentre pragmatic randomised trial to explore the impact of carer skills training on carer and patient behaviours: Testing the cognitive interpersonal model in adolescent anorexia nervosa. European Eating Disorders Review, 25(6), 551-561.

Holland, J., Hall, N., Yeates, D. G., & Goldacre, M. (2016). Trends in hospital admission rates for anorexia nervosa in Oxford (1968–2011) and England (1990–2011): Database studies. Journal of the Royal Society of Medicine, 109(2), 59-66.

Magill, N., Rhind, C., Hibbs, R., Goddard, E., Macdonald, P., Arcelus, J., . . . Landau, S. (2016). Two‐year follow‐up of a pragmatic randomised controlled trial examining the effect of adding a carer's skill training intervention in inpatients with anorexia nervosa. European Eating Disorders Review, 24(2), 122-130.

Mitchell, J. E., & Crow, S. (2006). Medical complications of anorexia nervosa and bulimia nervosa. Current Opinion in Psychiatry, 19(4), 438-443.

Yielded Benefits is not a requirement for new applications.