M-PHiL Study - Mental and Physical Health in Lambeth

The MRC 2010 review of mental health research underlined the importance of understanding the links between physical illness and poor mental health and noted the difficulties faced by some patients in obtaining satisfactory treatment when mental and physical conditions co-occur. The report called for an improvement in access to anonymised data based on NHS contact "for example, by data linkage across cohorts".
We plan to link Lambeth Data Net (LDN) established by Lambeth Primary Care Trust (PCT) containing the primary healthcare records from patients from all but two GP practices in the London Borough of Lambeth (population ~ 300 000) with CRIS, the Case Register Interactive Search system containing the entire electronic patient record of the South London & Maudsley Trust, the sole provider of NHS mental health services to patients in Lambeth (n=180000 of whom 36 000 live in Lambeth). The Caldicott Guardians for SLAM and for Lambeth PCT have both already approved the linkage plans. Each dataset has been linked before - LDN with primary care databases in east and south-west London and CRIS with the Thames Cancer Registry and Hospital Episode Statistics data. The study will be based in the Biomedical Research Centre for Mental Health, recently given a national leadership role to develop e-health for research in recognition of the advances made by CRIS.
The aim of the study is to generate a unique resource to improve research into the course and outcomes of mental ill health. Specifically our objectives are: [1] to create a one-off link between two established datasets containing information on the health of an entire London borough [2] to demonstrate the utility of this new dataset by using it to examine in detail why some patients with severe mental illness have excellent physical health and some very poor physical health [3] to develop the methods to be able to update the linkage so that the impact of any interventions we might develop from the proposed work can be assessed in real time.
Addressing health inequalities is a major area of interest both nationally and locally within London. Patients with severe mental illness such as schizophrenia have worse physical health than the general population- in south London we have shown that patients with severe mental illness die up to 17 years early. Whilst we know that on average patients with severe mental illness have more risk factors for poor physical health & they seem less likely to be offered the best medical care in hospital, we know little about which individual factors (e.g. ethnicity), disease factors (e.g. medication) and systemic factors (e.g. single-handed GP) are associated with good and bad physical health. Creating from scratch a dataset detailed enough to answer these questions would be prohibitively expensive but by linking two datasets comprising routinely collected data on the whole population (~36 000 in both datasets) we have sufficient data on sufficient individuals to address all these important issues.
In meeting the challenges thrown down by the 2010 MRC Review head on, our proposal has the potential to deliver both clinically and academically. Our study, linking the complete primary care and mental health records of one ethnically diverse London borough, is population-based and uses datasets specifically designed to enhance ease of access and which have themselves already produced major research findings. Our study will produce a methodological template for other groups wishing to exploit the possibilities of data linkage, a linked dataset that can be interrogated to address multiple issues of physical and mental comorbidity locally and beyond, and will provide a structure to monitor any interventions arising from such analyses almost in real time.

SLAM Information Governance group Clinical Data Linking Service (CDLS) will act as the Trusted Third Party. A bespoke dataset will be pulled within LDN and transferred to CDLS via the NHS 3N network. The BRC-ID, a pseudonymised identifier, links to patient identifiers in a secure database outside CRIS accessible to CDLS. Using identifiers present in both datasets CDLS will attach appropriate BRC-IDs to LDN cases known to CRIS. A column containing an anonymised ID that applies to all LDN cases will be created ("LDNID"). All identifying data in LDN will then be destroyed. This "new" dataset will be stored securely on a CDLS server behind the SLAM firewall. For approved projects researchers will send CRIS data to CDLS who will pull in LDN data linked by the BRC-ID to create a project-specific dataset, each case having a project-specific ID. The BRC-ID then removed, the final dataset becomes fully anonymised. [1] CRIS patients with an SMI diagnosis & a Lambeth GP will be identified and compared to LDN patients on the SMI case register. We will describe factors associated with having care from only one service or joint care. [2] A cohort of SMI patients known to both services in 2009 will be established. Physical healthcare received 2009-12(e.g. diabetes, smoking cessation, screening for cervical cancer) will be compared with other LDN patients. Individual, disease-related and systemic factors associated with both good and poor physical healthcare will be identified. [3]New cases of SMI on CRIS will be examined for information on physical illness and prescribed medications. This will be compared to that in LDN. LDN and CRIS data will be used to assess the effect of incomplete information on physical health in CRIS. Application: Our results will generate potential interventions, both clinical and at policy level. Our collaboration means these could be quickly implemented locally. Repeated linkages will allow the results to be observed almost in 'real time'.

We have identified that clinicians caring for patients with severe mental illness, patients with severe mental illness, local commissioners and service managers and policy makers at the local, regional and national level might benefit from our research.
Clinicians
Clinical staff caring for patients with severe mental illness (SMI) can benefit from our work by increasing their awareness of the association between SMI and poor physical health. This study will contribute to the development of specific tools to assist clinicians (both primary and secondary) in the management of the physical health of SMI patients. We will ensure our colleagues locally and within the broader Kings Health Partners are made aware of the award of the grant and our results via the intranet and other internal communications. We will publish our findings in peer-reviewed journals and at conferences. We will work with the IT department at SLAM who developed CRIS to adapt the IT:clinician interface to make the recording of physical health data more straightforward.
Patients with severe mental illness
Patients with severe mental illness will benefit from our research by being cared for by clinical staff with better quality more detailed information on the association between their mental and physical health. Our findings will lead to changes in the way services support or relate to patients with SMI. Patients will be able to benefit from having care provided within a system that is better set up to care for their physical health.
Local Commissioners and Health Service Managers
Individuals and organisations responsible for the design commissioning and management of services for the severely mentally ill will benefit from having high quality, population-based, locally derived data with which they can assess the nature and the extent of the problem of poor physical health in this population. Preliminary discussions have already revealed the high value local commissioners in Lambeth would place on knowing the results of our first analysis (what proportion of patients with severe mental illness receive care from both GP and mental health services). They will also benefit from being able to monitor the impact of service innovations in this vulnerable group. Our team includes local health professionals with a detailed knowledge of the organisation of healthcare in Lambeth and close links to local commissioners. They are in a position to lead on the targeted dissemination of our results. We will work with commissioners to devise measures of physical healthcare in SMI patients which might become clinical commissioning goals - new CQUINS for SMI patients.
Policy-makers within London and at national level
We are aware that the team implementing No Health Without Mental Health at the Department of Health (DH) are interested in both the physical health - mental health interface and especially in the facility to link datasets across primary and secondary care. We believe policy makers will welcome the establishment of the linkage and will use us as a resource. Our population-based data on both physical and mental health will be of use in policy terms as the No Health Without Mental Health strategy is rolled out. Policymakers have highlighted improving the health of SMI patients as a key target. We believe they will benefit from being aware of our methodology, from the results of our initial analyses and from the results of any further interventional or longitudinal work. Many members of our research team have established track records in working with policy makers. For example Henderson has spent time on secondment at the Department for Work and Pensions, and Ashworth has been an advisor at the DH. We will use established relationships to alert, in particular, policy makers at the Department of Health of our study and we will make ourselves available to assist non-clinicians in their interpretation.