Liaison Psychiatry Service Configurations and Referral Patterns and their Effects on outcomes: Sub Study An Evaluation of Cost-effectiveness and Efficiency Using Routine NHS Data

The purpose of this updated application is to extend the data sharing agreement relating to this data for a further year (until 30/06/2020) to enable analysis to be finalised and prepared for publication. No additional data is being requested as part of this extension request.

Liaison psychiatry (LP) involves the provision of mental health services in non-psychiatric settings, in this case in general hospitals. Liaison services exist because there are higher rates of most mental health problems in general hospitals than there are in the general population. The two main questions asked of them are: do they improve outcomes for the people referred to them and if so can they do so in a cost-effective way? It has even been claimed that LP saves the NHS money by reducing inappropriate use of expensive general hospital stays and treatment for people who are best helped in other ways. The challenge in answering these questions is that liaison services vary greatly in how they are set up, in the sort of referrals they see, and in how they deliver care. Liaison psychiatry for the elderly does not cover the same ground as that for working age adults. So asking whether liaison services are cost effective is not like asking whether a cardiac surgery service is cost-effective but more like (on a smaller scale) asking if general practice is.

A major challenge in assessing the cost-effectiveness of liaison psychiatry services resides in the variability in their case-mix and also in how they are configured. There is also considerable diversity in external (demographic and other service) factors that influence outcomes for liaison psychiatry. LP-MAESTRO is a project at the University of Leeds, which has been funded by the National Institute for Health Research and Health Services and Delivery Research (NIHR HS&DR) as part of a commissioned call (13/58 Organisation, quality and cost-effectiveness of psychiatric liaison services in acute settings), to develop an approach based upon economic modelling and using routinely-collected NHS data, that allows the University to evaluate the cost-effectiveness and efficiency of particular configurations of liaison psychiatry service for specified target populations.

The overall aim of LP-MAESTRO is to evaluate the cost-effectiveness and efficiency of particular configurations of liaison psychiatry service for specified target populations. To do this, an innovative approach based upon linking routinely collected patient-level data and using economic modelling with the resulting aggregated data will be developed and evaluated.

Under the General Data Protection Regulation (GDPR), a legal basis is required for processing of personal data. The legal basis under which data is processed within the project Liaison Psychiatry Measurement and Evaluation of Service Types, Referral Patterns and Outcomes (LP-MAESTRO) is:

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

Processing of data concerning health (and other ‘special categories’ of data) requires the fulfilment of an additional condition under GDPR. The specific condition fulfilled by the processing of such data within LP-MAESTRO is:

•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.”

The overall aim is to evaluate the cost-effectiveness and efficiency of particular configurations of liaison psychiatry service for specified target populations. The purposes described in this agreement amount to a sub-study of the LP-MAESTRO project.

LP-MAESTRO is a collaboration between academics and practitioners at the University of Leeds, University of Manchester and Imperial College, Peninsula Schools of Medicine & Dentistry, The Centre for Mental Health, The Royal College of Psychiatrists and Leeds York Partnership Foundation Trust. Leeds University are the Data Controllers for this work described under this agreement - the other organisations noted in the collaboration do not have any influence over the outputs generated. The other collaborators listed provide contributions to the main LP-MAESTRO project DARS-NIC-77953-C4M3T-v1.3. These organisations have no direct role in the processing of the data received under this agreement for the sub-study of LP-MAESTRO

Leeds University will have rights to the data disseminated for this project to relate to the processing of the data. Only staff at the University of Leeds will be involved in the processing of the data received from NHS Digital.

LP-MAESTRO has three work packages as a whole. The current work as described in this agreement was originally intended as pre-protocol preparation for Work Stream 2 of the LP-MAESTRO Project. However, delays in data release mean it has been subsumed into that work stream, demonstrating the value but also the significant drawbacks of using single datasets like HES to attempt to evaluate liaison services.

The work packages within this agreement do not directly relate to the work being undertaken in Workstream 2 which has its own agreement in place - DARS-NIC-77953-C4M3T-v1.3

Workstream 2 – Phase 1 (WS2P1) focuses on 11 acute hospitals in England that were identified within Workstream 1 of LP-MAESTRO and the Liaison Psychiatry Survey England 2015 as not having a Liaison Psychiatry service.

Workstream 2 – Phase 2 (WSP21) focuses on hospitals in England that were identified within Workstream 1 of LP-MAESTRO and the Liaison Psychiatry Survey England 2015 as having a particular configuration of Liaison Psychiatry service. To enable pathways to be constructed for patients attending these hospitals that consider the interaction of patients with the Liaison Psychiatry services in these hospitals, data is additionally

required from the NHS trusts that run these services.

The work-packages in this agreement relate to the preparation for work-stream 2.

Record-level data supplied under this agreement will only be held, processed and stored by the University of Leeds. Other collaborators and partners mentioned above will only have access to aggregated outputs with small numbers suppressed in line with the HES analysis guide.

The data requested under this agreement includes the Hospital Episode Statistics (HES) inpatient episodes for patients (aged 16 years or over) who have attended two specific hospitals in Birmingham over a 5 year period (2007-2011):

o City Hospital Birmingham

o Queen Elizabeth Hospital Birmingham (University Hospitals)

For each episode, the following data items were requested and have been received:

o Patient pseudonym

o Duration of spell

o Diagnostic codes

o Treatment speciality

o Provider code

o Date of admission

o Method of admission

The episodes in the two hospitals have been compared in terms of:

o Length of spell

o Readmission within 12 months from discharge

Figures have been compared (for different patient subgroups) across the hospitals using standard statistical methods.

Special statistical techniques (multilevel negative binomial model) were used to deal with the two challenges:

- [1] length of stay are very unevenly distributed with some people staying only one day and some staying many weeks,

- [2] a single patient may account for several inpatient episodes, complicating interpretation.

This has enabled the estimation of whole hospital effects over time (for the five-year study period) and comparison between hospitals. The data have allowed the study to answer an important question about liaison psychiatry services and their effect on lengths of stay in a general hospital; this was a key part of the logic of the project and is therefore useful to the overall research programme. The analysis indicated a number of factors of relevance in understanding the project as a whole: the inaccuracy of mental health coding in HES (mainly under-reporting); striking differences between hospitals in terms of coding rates and of lengths of inpatient stay (LOS); a time trend for LOS with a linear reduction over a 5 year period – this pattern is extremely important in understanding claims that liaison services reduce hospital lengths of stay; lack of an obvious effect of introduction of a Rapid Assessment Interface and Discharge (RAID) liaison psychiatry service on hospital-wide lengths of stay once the background trend is taken into account.

Data will not be used for any other purpose that is not defined in this agreement. No attempts will be made to re identify individuals with data provided under this agreement.

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

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.

Data is stored securely in an infrastructure provided by the Leeds Institute for Clinical Trials Research (LICTR). Access to the data is restricted to specific members of the research team authorised by the Principal Investigator and granted access by the administrator(s) of the LICTR infrastructure.

Data relating to this project is backed up to the University of Leeds Central Backup Service on University of Leeds owned and managed servers, located within University of Leeds managed data centres. For disaster recovery purposes this data is mirrored to University of Leeds owned and managed servers located in the TFTA Data Centre at University of York. At no point does the University of York or its staff have access to view or process this data. UoY are renting secure rack space for servers to the University of Leeds. The servers themselves, their storage, access and management are all maintained exclusively by UoL staff. Therefore University of York are not considered to be Data Processors.

Data was initially processed to derive the analytical variables of interest. Special statistical techniques (multilevel negative binomial model) were then used to deal with the two challenges: [1] length of stay are very unevenly distributed with some people staying only one day and some staying many weeks, [2] a single patient may account for several inpatient episodes, complicating interpretation. This has enabled the estimation of whole hospital effects over time (for the five-year study period) and comparison between hospitals. The data supplied has been (and will be) solely used for the purpose outlined within this agreement. Data will be securely destroyed prior to the end date of the data sharing agreement (subject to any further extensions).

This data captured the admissions details for two specific hospitals in Birmingham over a 5 year period to facilitate their comparison.

Only substantive employees of University of Leeds will have access to the data. Project collaborators at the Royal College of Psychiatrists and the Centre for Mental Health do not process any of the data disseminated they only have access to aggregated outputs with small numbers suppressed in line with the HES analysis guide.

The work packages within this agreement do not directly relate to the work being undertaken in Workstream 2 which has its own agreement in place - DARS-NIC-77953-C4M3T-v1.3

Workstream 2 – Phase 1 (WS2P1) focuses on 11 acute hospitals in England that were identified within Workstream 1 of LP-MAESTRO and the Liaison Psychiatry Survey England 2015 as not having a Liaison Psychiatry service.

Workstream 2 – Phase 2 (WSP21) focuses on hospitals in England that were identified within Workstream 1 of LP-MAESTRO and the Liaison Psychiatry Survey England 2015 as having a particular configuration of Liaison Psychiatry service. To enable pathways to be constructed for patients attending these hospitals that consider the interaction of patients with the Liaison Psychiatry services in these hospitals, data is additionally

required from the NHS trusts that run these services.

The work-packages in this agreement relate to the preparation for work-stream 2.

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 the University of Leeds and will not be accessed or processed by any other third parties not mentioned in this agreement.

The data so far have allowed the study to answer an important question about liaison psychiatry services and their effect on lengths of stay in a general hospital; this was a key part of the logic of the project and is therefore useful to the overall research programme. Preliminary results have been presented at a Rapid Assessment, Intervention and Discharge (RAID) network conference in Chester in April 2017, but the main dissemination activity is as below.

There are a number of audiences for the outputs of this research – clinical service providers in liaison psychiatry and general hospital managers, both groups who wish to plan the most effective provision of psychiatric services; commissioners of healthcare in Clinical Commissioning Groups (CCG's) and their associated Commissioning Support Units; data scientists who are interested in the challenges of data-linkage and the information it can yield.

For clinicians and academics in liaison psychiatry and health services researchers the study will produce standard academic publications – the NIHR report, papers and conference presentations. The study will disseminate these outputs via academic and clinical publications . The Centre for Mental Health will produce two publications related to the findings – these are disseminated widely to non-academic audiences.

For non-academics in management the study will produce a synopsis of findings. These will be prepared in paper format and in the form of mini-tutorials that the study will make available by links to the website.

The study will signpost these outputs through articles in newsletters and via professional discussion forums such as the liaison psychiatry jisc-mail group (JISC-Mail provides email discussion lists for education and research) and economic associations’ mailing lists.

To ensure that the research reaches as wide an audience as possible the study plan to develop a series of interactive webinars that will be free to view. After the live webinar, the presentation recording will be available free to download, hosted on the university leading the research and the project webpage and will be publicised via mailing lists and social media such as Twitter.

At the end of the project the study will hold a specialist conference the aim of which will be to review current knowledge in the underpinning areas of whole-service evaluations of cost-effectives as relevant to liaison services and the use of data linkage to evaluate services that cut across sectors, and the specific content area - what the study have learned about cost-effectiveness in liaison psychiatry.

An Academic paper to a health services research or mental health journal will be submitted by the end of 2017 (The previous version of this agreement requested an extension to the agreement end date in order to facilitate the required analysis and enable generation of research outputs. The date for this output has not changed from the previous version of the agreement).

Use of the data will be included in the final report of main project. The end date of the project was previously 31st December 2017. However, a request is currently under consideration by the funder of the main project (NIHR HS&DR) to extend the project for a further 12 months (30th June 2020).

A discussion paper will be produced for mental health service commissioners.

The Birmingham findings will be presented as described in this section, as part of the wider LP MAESTRO findings

A paper published in a medical academic journal will be read mainly by clinicians in liaison psychiatry and by applied health researchers, and this will influence national debates about the nature of liaison provision, which take place in professional forums such as the Royal College of Psychiatrists. Mental Health Service commissioners influence services by the way they fund structures and require certain outcomes; the project will inform the approach to service evaluation when services are commissioned with the aim of reducing hospital lengths of stay. The project has influenced the overall approach in LP-MAESTRO by indicating the need to have patient-level data to evaluate the impact of liaison services.

All outputs will consist of aggregate data with small numbers suppressed in line with the HES Analysis Guide. No record level data will be shared with any third party.

The results of this application will be used to determine the cost-effectiveness and efficiency of a particular liaison psychiatry service configuration. They are planned for later this year with two groups in Royal College of Psychiatrists and Centre for Mental Health.

Benefits include:

- Illustration of a value of a particular approach (interrupted time series - which is analysis using a quantitative, statistical method in which multiple (sometimes as many as 40 or 50) repeated observations are made at regular intervals before and after intervention) in evaluating impact of new services.

- Recommendation to commissioners on an approach to evaluating services using routine data.

Project collaborators in Royal College of Psychiatrists and the Centre for Mental Health are conducting interviews and focus groups with service commissioners, to identify the best form and content for materials generated by the project when the aim of those materials is to inform decision-making. These are planned for later this year with two groups in Royal College of Psychiatrists and Centre for Mental Health.

In Work Stream 3 of LP-MAESTRO (a further stage) the study will then generate those materials. The impact on health and social care of any one piece of research is necessarily limited by the impact of competing claims – including policy imperatives that are based upon values rather than evidence. Nonetheless, this work should be influential because it is only the second attempt in the UK to evaluate the effect of liaison psychiatry services on acute hospital lengths of stay.

The study has developed groups of committed and informed members of the public who are all people with experience of long term physical illness and awareness of mental health issues. They are drawn from patient organisations such as Diabetes UK, Breatheasy and Heartline. and have experience of membership of steering committees, development of participant information resources, research websites, patient interventions and dissemination of research findings. These organisations provide representation for patients with physical health problems. Such patients may experience mental health problems during the course of a hospital admission and require care from a liaison psychiatry service.

The data so far have allowed the study to answer an important question about liaison psychiatry services and their effect on lengths of stay in a general hospital; this was a key part of the logic of the project and is therefore useful to the overall research programme. Preliminary results have been presented at a Rapid Assessment, Intervention and Discharge (RAID) network conference in Chester in April 2017.

The purpose of this updated application is to extend the data sharing agreement relating to this data for a further year (until 30/06/2020) to enable analysis to be finalised and prepared for publication.