This study aims to identify the determinants of diabetes and to explore variation in diabetes outcomes for people with severe mental illness (schizophrenia, schizoaffective disorder and bipolar disorder) in order to develop potential healthcare interventions that can be tested further.

The study utilises a mixed methods design comprising analysis of patient primary care records and interviews with patients living with comorbid SMI and diabetes, family carers and healthcare professionals involved in commissioning or delivering healthcare for this population. This entry on the Protocol Registration and Results System describes only the quantitative Work Package of the study in detail i.e. analysis of patient primary care records.

People living with severe mental illness (SMI) have a lower life expectancy (by around 20 years) than the general population, often dying of preventable or manageable illnesses (Woodhead et al., 2014). Diabetes contributes significantly to this health inequality, being 2-3 times more likely in people with SMI (Stubbs et al., 2015). This study aims to understand the determinants of type 2 diabetes and variation in diabetes outcomes for people with SMI in order to identify potential healthcare interventions that can be tested further.

Research questions:

1. What are the socio-demographic and illness-related risk factors associated with

   1. developing type 2 diabetes in people with SMI?
   2. variation in diabetes and mental health outcomes in people with SMI and type 2 diabetes?

2. How do physical and mental health outcomes differ between people with SMI and type 2 diabetes,

   1. compared to people with SMI without diabetes?
   2. compared to people with type 2 diabetes but no SMI?

3. What factors influence access to, and receipt of, diabetes care for people with SMI, and how are diabetes healthcare interventions experienced by people with SMI?

4. How and at what cost is type 2 diabetes monitored and managed in people with SMI compared to those without SMI?

5. What healthcare interventions (e.g. medication, referrals and care pathways) are associated with better diabetes outcomes for people with SMI and type 2 diabetes?

The study has the following objectives:

1. In people with SMI, to identify which socio-demographic, illness, family history and lifestyle factors are associated with the development of type 2 diabetes.
2. In people with SMI and type 2 diabetes, to identify which socio-demographic, illness, family history and lifestyle factors are associated with variation in diabetes and mental health outcomes.
3. In people with SMI, to compare healthcare interventions, physical and mental health outcomes in those with type 2 diabetes with those without diabetes.
4. In people with type 2 diabetes, to compare healthcare interventions, physical and mental health outcomes in those with SMI and those without SMI.
5. To understand the factors that influence access to, and receipt of, diabetes care for people with SMI, and explore the experience of diabetes healthcare by people with SMI.
6. To compare diabetes care provision for people with and without SMI, and estimate costs for these.
7. To identify which healthcare interventions (e.g. medication, referrals and care pathways) may be associated with better diabetes outcomes for people with SMI and type 2 diabetes.

Study design:

The study will utilise a convergent triangulation mixed methods design comprising a quantitative longitudinal observational study of individual patient records of adults with diagnosed SMI and diabetes held in the Clinical Practice Research Datalink (CPRD, a computerised database of anonymised primary care medical records drawn from a proportion of general practices in the UK) with qualitative interviews with patients, their family carers and healthcare professionals who commission or deliver services to this comorbid population. This entry on the Protocol Registration and Results System describes only the quantitative work stream of the study i.e. the interrogation of patient primary care records. The key healthcare outcomes to be examined are listed; however, a definitive list of outcomes will be agreed through a consensus process with project team members.

The complete study is organised in three work packages:

1. Work Package One, which involves ongoing consultation with experts (service users and family members belonging to a Patient and Public Involvement (PPI) panel, healthcare staff and researchers) to refine questions and analyses, a synthesis of evidence on SMI and diabetes risks and outcomes, integrating the qualitative and quantitative studies, conducting co-design workshops to translate study findings into recommendations, and optimising the dissemination of findings.

2. Work Package Two, which investigates the role of risk factors thought to influence i) development of type 2 diabetes in people with SMI (objective 1), and ii) variation in health outcomes in people with comorbid SMI and type 2 diabetes (objective 2). Further analyses involve a comparison of diabetes healthcare and outcomes for people with SMI and type 2 diabetes with outcomes for people with either condition alone (objectives 3 and 4), an examination of variations in type 2 diabetes screening, monitoring and management and estimation of the costs for these (objective 6); and an exploration of the role of these interventions in contributing to health outcomes (objective 7). These investigations will utilise the CPRD database, which is linked to other datasets, including Hospital Episode Statistics (HES) data, HES-Office of National Statistics mortality data, and Index of Multiple Deprivation (IMD) data.

   Three CPRD datasets will be analysed which will consist of linked records from the study period of 01 April 2000 to 31 March 2016. The first of these, Dataset A, will contain records of a cohort of adult patients with a diagnosis of SMI. Dataset B is constructed from Dataset A, and comprises those in Dataset A who also have a diagnosis of diabetes. Finally, Dataset C is a cohort of patients with diabetes but no record of SMI who are age and gender-matched to Dataset B on a ratio of 4:1.

3. Work Package Three, which explores the experience of diabetes management and healthcare from three perspectives: patients, family members or friends who support them and healthcare professionals who commission or deliver services.

Dissemination

Findings will be translated into service recommendations through two multi-stakeholder co-design workshops. The first of these will identify further avenues of inquiry for the final stages of analysis and will develop draft recommendations to improve the organisation and delivery of healthcare for this comorbid population. The second workshop will further develop the recommendations based on the final analyses and will design interventions or care pathways, assessing their potential acceptability and feasibility for future evaluation and implementation.

A dissemination event will be held, presenting materials in a range of formats to maximise opportunities for knowledge exchange. Study outputs will include a final report for the funding body, scientific papers submitted to international peer-reviewed journals, short articles in practitioner journals, executive summaries published on websites, conference presentations, and other materials disseminated through clinical, academic and voluntary sector networks.

Inclusion Criteria:

• Patients aged 18 and over
• Patients with diagnoses of severe mental illness, severe mental illness and type 2 diabetes, or with diabetes but no diagnosis of severe mental illness.
• Patients with research-quality data.

Exclusion Criteria:

• Patients under the age of 18
• Patients without research quality data.
• Patients who were diagnosed with Type 1 diabetes following a diagnosis of Type 2 diabetes.
• Patients who have a code for a diabetic complication (e.g. diabetic retinopathy) but no diagnostic code for diabetes.

- Dataset A: patients (with research-quality data) who received a diagnosis of severe mental illness (schizophrenia, schizoaffective disorder and bipolar disorder) at the age of 18 years or above, and who have at least one event of SMI in the study period (01/04/2000-31/03/2016)

Dataset B: patients drawn from Dataset A who also have a diagnosis of type 2 diabetes mellitus (received on/after the age of 18).

Dataset C: patients with diabetes but no diagnosis of severe mental illness matched to Dataset B on a ratio of 4:1 on the basis of age (+-2 years), gender and GP practice.