NHS Digital Data Release Register - reformatted

University Of Essex

Project 1 — DARS-NIC-07360-K4R9R

Opt outs honoured: N

Sensitive: Non Sensitive, and Sensitive

When: 2016/09 — 2016/11.

Repeats: One-Off

Legal basis: Health and Social Care Act 2012, Approved researcher accreditation under section 39(4)(i) and 39(5) of the Statistical Registration Service Act 2007

Categories: Anonymised - ICO code compliant


  • Hospital Episode Statistics Accident and Emergency
  • Mental Health Minimum Data Set
  • Office for National Statistics Mortality Data (linkable to HES)


Community Treatment Orders (CTOs) were introduced to the Mental Health Act in 2007 in England and 2008 in Wales allowing patients detained in hospital to continue treatment in the community under specific conditions. To be eligible for a CTO, a patient must be detained in the hospital under the Mental Health Act on a section that allows for compulsory treatment. Around 20,000 patients have been placed on CTOs since 2008. CTOs were introduced to reduce the ‘revolving door’ phenomenon (frequent re-admission to hospital by small number of people with severe and persistent mental illness) and to provide treatment in the least restrictive setting. In practice CTOs tend to be applied at the end of an admission, following complete remission, as a means of optimising treatment adherence in those likely to default or disengage from care. Patients and carers perceive CTOs as coercive and mainly concerned with medication adherence. The use of CTOs remains controversial as its applications seems to run counter to the notions of ‘preventative’ and ‘least restriction’ for which CTOs were introduced to the mental health care system. The Oxford Community Treatment Order Evaluation Trail (OCTET) compared discharged from hospital on a CTO with discharge from the hospital under section 17 of the Mental Health Act. No significant differences between the groups were found. Swartz et al (1999) randomized 264 severe and persistent mentally ill patients to either CTO or voluntary care following discharge. The results showed that schizophrenia patients on CTO for longer than 180 days had fewer hospital admissions and spend less time in the hospital then the controls receiving voluntary care. Limitations of these trial studies include small sample size, sample selection (exclusion of specific patients, refusal of participation by specific patients). In addition, the relative brief duration of follow-up might be too short to observe beneficial outcomes. Observational studies evaluating CTOs have been undertaken in diverse settings, at different times, using different methods. Results from these studies are inconsistent. Discrepant findings about CTO outcomes highlight the importance of local context. Understanding the effects of local context may also be the key to explaining persistent ethnic inequalities in CTO use. This lack of evidence for the effectiveness of CTOs is problematic for policy makers, service commissioners and providers and for those who receive mental health care. CTOs require stringent governance and therefore place significant demands on the already stretched NHS mental health budgets. CTOs also reduce satisfaction with care and increase stigma. The Care Quality Commission and the House of Commons Health Committee have called for research into the ‘value of CTOs in different clinical and social circumstances’. The House of Commons Health Committee has noted:’… that the evidence base for this policy remains sparse…’ and recommended ‘…,a fuller analysis of the value of a CTO in different clinical situations’. The aims of this research project are: 1. To explore spatial and secular variation in the use of Community Treatment Orders (CTOs) in England over a four-year period (2011/12 to 2014/15), including variation between people, places and services and over time in the likelihood of patients on Sections 3 and 37 being placed on CTOs, and in the frequency and time to subsequent events including recall, revocation, readmission and discharge from CTO. 2. To describe and model associations between CTO use and outcomes, namely re-admission, time spent in hospital, time spent in intensive psychiatric care or forensic units, episodes of seclusion and restraint, community mental health service contacts, criminal justice system contacts, Accident & Emergency Department attendances and deaths, after adjusting for propensity to require compulsory treatment. 3. To test the hypotheses that outcomes associated with CTOs (including the re-admission and the adverse outcomes listed in (3) above) vary between people and places. These models will be used to explore which patients, if any, benefit from CTOs and the types of places and service context where CTOs might have the greatest impact on patient and service-level outcomes. 4.To model the additional health care costs associated with CTOs (including administrative and regulatory costs) and the impacts of increasing or decreasing their use using a range of projections. Note: The term ‘Mental Health (MH)’ data is used to refer to the Health and Social Care Information Centre’s Mental Health Minimum Data Set (MHMDS) for the years from 2011/12 to 2014/15 and Mental Health and Learning Difficulties Minimum Data set for the years since 2014. MH data will be used to address these questions. A cohort of mental health users will be selected from MH data and these records will be linked with corresponding Hospital Episode Statistics Accident and Emergency and ONS Mortality records. Using Lower Super Output Area (LSOA), ONS staff will link HSCIC MH data to the publicly available data, i.e. Indices of Multiple Deprivation (Department for Communities and Local Government) and aggregate LSOA Census 2011 (ONS) data. This is to further explore the spatial variation in CTO use in England. Effectiveness of CTO will be measured in relation to compulsory treatment. The exposed group will be patients discharged from the hospital on CTOs, the unexposed group will be patients discharged from hospital treatment orders but not subject to CTOs. The two groups will be matched on time of discharge, service provider, and propensity score. The HSCIC data will be used to determine the effectiveness of CTOs compared to compulsory treatment. Primarily this project will lead to submission of academic papers describing spatial and secular variations in the use of CTOs in England, the effectiveness of CTOs, and healthcare costs associated with CTOs. Results from the project will be presented at internal and external conferences. Findings from the research will be used to engage with public/ third sector organisations or mental health care providers during meetings or conferences. The Mental Health Foundation is a key partner in this research project. With their help two expert reference groups will be recruited: one with carers/mental health care users and one with mental health care professionals. These reference groups will meet to discuss the study aims and methods as well as comment on the released non-disclosive results and support dissemination of findings.

Expected Benefits:

Date: • Improved insight in the use of CTOs. CTOs have been used far more extensively than was anticipated. Analyses will show when, where, and for whom CTOs are being used. Understanding the large spatial and secular variation in the use of CTOs can help inform mental health policy and support commissioning decisions on productive monitoring procedures for CTOs. • Developing evidence on the effectiveness of CTOs. Understanding patient and outcome characteristics can help identify those inpatient detained under hospital treatment orders that would benefit most from CTO upon release from the hospital. • Understanding the cost-effectiveness of CTOs. Identifying positive outcomes and the reduced health care costs associated with the use of CTOs will help direct mental health care related funds to where they are most needed. • Findings that can be translated into actionable conclusions for service users. Involvement of mental health care users, carers, and professionals will ensure a study focus on outcomes and associations relevant to mental health care users. The findings can then be translated into patient-approved interventions. Involvement of the Mental Health Foundations also facilitates dissemination of study findings. • Findings with direct policy relevance. The research team will ensure that these findings are shared at the earliest opportunity with the Care Quality Commission, the Department of Health, the Chief Medical Officer, the Royal College of Psychiatrists, the NHS Confederation, and NICE. • Guidance to the National Institute for Health Research on strategies for future policy and practice with regard to CTO use in England. Results might inform future interventions to reduce compulsory admission and variation in CTO use. The research team will work on facilitating the adoption of these patient beneficial interventions within the NHS. The project is expected to start around March 2016 and expected to be completed in March 2018. Funding for this project has been requested for a period of 2-years.


The expected outputs from this study will include papers on: • The use of CTOs in England including variations in rates of CTO use at different spatial levels • The hypotheses that CTOs are associated with reduced (mental health) inpatient care, reduced interaction with accident and emergency services, reduced mortality, and increased community mental health service contacts and these associations vary between places after controlling for patient characteristics. • The health care resources employed by CTOs for mental health treatment In addition to mandatory publication in the NIHR Journals Library as required by the project’s funder, the findings will be submitted to high impact journals. Presentations of interim results will be made to relevant national and international conferences. The project (i.e. production and dissemination of outputs) is expected to be completed in September 2018. Funding for this project has been awarded for a period of 2-years from March 2016. The project has an absolute end date of September 2020. The intention is to maximise the potential impact and societal benefits of the research through: 1. Presentation of findings and contributions at national and international scientific conferences 2. Workshops for key stakeholders in each of the three regions from which the clinical academic applicants are drawn: London, the Midlands, and the North East. 3. Discussion of the implications for health services with health care providers and patient support groups (Mental Health Foundation) 4. Engagement with other academics – discussing avenues for future research if the research suggests differential use of CTOs in England 5. Deposition of academic papers in accordance with open access protocol to ensure that published papers are not withheld from a wider audience via journal pay-walls. 6. Development of short plain-language summaries of findings targeted towards non-academic professional audiences (e.g. patient support groups, local authorities, health care professionals, policy makers). 7. Work with Trusts, Innovation Hub, University Partners, and AHSN to ensure dissemination of results and adoption of proposed interventions locally, regionally, and nationally.


The HSCIC will create a cohort of Mental Health patient records based on the following criteria: Patients that have or had a MH spell of care open within the period from April 2011 to March 2015 that includes: a) any associated Mental Health Act Event Episode/s where the legal status is coded as one of a list of supplied codes indicating detention under specific sections of the Mental Health Act or the Criminal Procedure (Insanity) Act 1964 as amended by the Criminal Procedures (Insanity and Unfitness to Plead) Act 1991, or; b) any associated Supervised Community Treatment Episode The HSCIC will then generate a file of pseudonymised Mental Health patient identifiers. Using this file, the HSCIC will determine the relevant HES ID using a bridging file. They will then extract the relevant HES and ONS Mortality records for those patients identified within the Mental Health cohort. Two bridging files will be created i.e. a HES/ONS bridging file and a Mental Health/HES bridging file. HSCIC will extract the remaining Mental Health data records for the cohort. From the ONS mortality data, the field 'Date of Death' will be replaced with the derived month and year of death only. The Mental Health data is classed as sensitive. All of the linked data is classed as pseudonymised. The multiple files will be transferred to the secure environment at ONS VML in Titchfield where ONS staff will use the LSOA to link the HSCIC data to publicly available data (Indices of Multiple Deprivation and aggregate LSOA Census 2011 data) as described above. All data linked to HSCIC data are non-identifiable and publicly available. Data will be used only for the purposes indicated on the approved ADRN project application and only by ADRN accredited researchers named in this application. Researchers have to undergo ADRN training before accessing any data or intermediate outputs and will only access the data in the secure environment. ONS will ensure that any researchers accessing data have accredited researcher status. All researchers have been guaranteed by someone who can sign up to the breaches and penalties policy by their institution and have received safe data handling training. Data will not be used for commercial purposes. The secure environment at ONS VML Titchfield uses state-of-the-art secure information technology and procedures which provide physical, hardware and software security. Data will be processed (i.e. linked and assembled) in a separate (virtual) secure area than the area the researchers will be accessing the data. All staff handling the data are required to have security clearance at a level appropriate for the data they are going to handle as per the ADRN Secure Environment Policy (ADRN032). All named individuals in this agreement are employed by the ONS, University of Southampton, the University of Portsmouth, University of Warwick, Ulster University or the Northern Ireland Statistics and Research Agency (NISRA). The researchers named in the application will be granted access to the VML ONS for this specific research project as part of the ADRN services and infrastructure. The research team will visit ONS VML Titchfield on location to access the linked de-identified dataset in a dedicated secure room. They will be able to use installed software on their allocated PC (thin client) to analyse the data and produce outputs (e.g. tables), but they cannot take anything in or out of the room with them (including mobile, phones, memory sticks or even pen and paper). They will also be watched by the secure environment staff via a CCTV and all their activity when logged in on the PC will be monitored. They will not be able to copy, download or disseminate the data in any way – if they do there will be strict penalties for any breach or misuse of data, following the ADRN breaches and penalties policy (ADRN003). Due to geographical distance, one instance of remote access to the VML from the secure environment of the Administrative Data Research Centre of Northern Ireland (ADRC-NI) will be needed for one of the researchers (from Ulster University) working on the project, however all data will remain on the ONS VML Titchfield server as per named users accessing on location. ADRC-NI secure environment is located in the Northern Ireland Statistics and Research Agency (NISRA) and the user will undergo training in accordance with the ADRN policy, adhering to the operating procedures of both NISRA and VML safe settings, which will include site specific training and BPSS security. Statistical outputs will only be cleared and released by VML Titchfield. ONS will act as joint data controller, as theirs will be the location where data will flow to and be stored and ONS will have full control of access of the data. As such, they have provided all the necessary security assurances (current IG toolkit and DPA registration). The listed data users' organisations will not need to separately demonstrate compliance with security assurances as they will be performing data processing activities solely in ONS' controlled environment and will not be storing or transferring data in any way. They are therefore covered by the security assurances of ONS. ONS is ultimately responsible for managing secure access. Data will be only stored and processed within ONS VML Titchfield. For all users, outputs will be statistical tables and output of statistical models reviewed by ONS to ensure they are non-disclosive and suitable to be made publicly available. Once cleared for release, ONS then supply the outputs to the named researchers via secure encrypted email. All outputs will be aggregated with small numbers suppressed in line with the HES analysis guide. After the end of the project, secure environment staff in ONS Titchfield will store the dataset for as long as required and authorised to do so and will destroy the data, as per the data retention and destruction policy when appropriate, securely and in line with HSCIC requirements. Once analysis is completed, the research team will produce a two-page plain English summary for ADRN and publish their results. Finally, note that non-disclosive outputs and results will be discussed among the whole research team before final results are compiled for dissemination. These results will also be used to populate the health care cost analyses. As health cost modelling is based on the non-disclosive results, these analyses can be performed outside the secure facility of the ADRN.