LONG TERM OUTCOMES AND HEALTH INEQUALITIES IN BIPOLAR AFFECTIVE DISORDER WITHIN A UK PRIMARY CARE COHORT (1995-2012)

Bipolar affective disorder (BPAD), or manic depression, is recognised as one of the commonest causes of worldwide disability in adults of working age. It is a lifelong illness, which often starts early in adulthood, and causes sufferers to experience prolonged episodes of depression or elation. Currently the illness affects more than 1% of the population and costs the UK £6 billion/year. It is recognised that research into BPAD is underfunded relative to other severe mental illness (SMI).

This study aims to answer two questions about BPAD: 1) Can long-term health outcomes be improved by appropriate medication? 2) Does social deprivation or where you live lead to reduced access to treatment and worse outcomes?

This study is important because we know that people with BPAD are frequently admitted to hospital, suffer more physical and mental illness, more commonly die by suicide, and are more likely to die early compared to the rest of the population. Episodes of illness can result in loss of jobs, relationships and housing. However the true extent of these problems isn't fully understood and needs to be clarified.

Stabilising the mood of people with BPAD with medication should, in the long-term, have a positive impact not only on their lives, but also on their families, use of NHS services, and society as a whole. There are a number of medications recommended as appropriate treatment to manage the condition in the long-term (including lithium, valproate and olanzapine). The evidence supporting the use of these drugs is based on randomised controlled trials (RCTs), which aim to compare directly the effectiveness of drugs with each other, or with a placebo. However, these trials often have short follow-up times (of a matter of months whilst in reality patients take these medications for years), involve small numbers of patients, and are very selective about the types of patients they include (perhaps excluding "typical" people with BPAD, who also have other health problems). So it remains unclear if the medications really do reduce these negative outcomes.

It is also not known if a patient's age, sex, ethnicity, social deprivation, urbanicity and location affect their access to treatment and outcomes. Any inequalities in health treatment are unfair and detrimental to society, but they are avoidable and identifying them would allow targeted interventions to be developed.

The project will use information from The Health Improvement Network (THIN), a database of General practice (GP) consultations. It contains records of diagnosis and medication prescribed to over 10 million patients. There are over 20,000 patients with BPAD in the database (the largest group of BPAD patients ever examined), with an average follow-up time of almost eight years. The database is updated automatically and anonymously when a GP makes an entry, therefore the project does not require input from patients (whose mental health may be affected by involvement in such a study).

I will examine how often patients experience hospitalisation, contact with emergency psychiatric teams (Crisis Teams or Home Treatment teams), how many self harm, die by suicide and die prematurely. I will see if these outcomes are more common in particular groups; by treatment and by social deprivation.

The reason studies such as this have not been possible in the past is that UK databases were not of sufficient quality, and did not include information on admissions. This is the first time GP and hospital data will be combined to examine BPAD. Also there was a worry that, by not randomly allocating patients to a treatment (like RCTs), results would be biased. In this study a number of statistical methods will be used to minimise possible bias.

The project will give a comprehensive understanding of the effects of, and outcomes in BPAD, and will add to the evidence on appropriate treatment with an approach that has not been attempted before.

Aims
To examine i) long-term serious adverse outcomes in bipolar disorder (BPAD), and ii) whether socio-demographic gradients exist in access to treatment and clinical outcome

Objectives
1. To determine the long term prognosis of patients with BPAD, by identifying rates of serious adverse outcomes: i) hospitalisation and crisis team intervention, ii) self-harm, iii) suicide and iv) all cause mortality
2. To determine if proportion of time in treatment, and treatment type, predict rates of serious adverse outcome
3. To determine whether social gradients exist in: i)The receipt of appropriate psychotropic medication, ii) The rate of serious adverse outcomes
4. To describe time trends in any such associations, in relation to changes to GP contract and financial incentives for care for Severe Mental Illness (the Quality and Outcomes Framework)

Methodology
THIN is one of the largest sources of primary care data in the UK, with information on over 20,000 patients with BPAD, it contains records of each patient's medical conditions, symptoms, and medications. The study will compare individuals with BPAD by time in treatment and treatment type. Outcomes will be serious adverse events. Social gradients in access to treatment and frequency of negative outcomes will be examined. Statistical techniques will be used to reduce confounding by indication and other biases related to unmeasured covariates.

Scientific & medical opportunities
An understanding of long term outcomes and existence of health inequalities in BPAD is lacking. Large-scale observational studies will be needed to fill this gap. The project will lead to identifying demographics of at risk groups, treatment that offers an improved quality of life for those with BPAD and a decreased negative impact on society. The argument for use of primary care data for outcomes/health inequalities research in other databases and areas of medicine will be strengthened.

All patients diagnosed with bipolar disorder, (conservative estimates suggest around 1% of the population) are potential beneficiaries of the project, which aims to add to the body of evidence in an under researched area of mental health. The question still remains "what is the most effective long-term treatment for bipolar disorder?" Patients could benefit from this research immediately if results were appropriately disseminated to clinicians and became part of management guidelines. If there is a difference between the study medications a clear discussion of the risk and benefits of each could be had between clinician and patient. Ultimately patients could benefit from prevention of relapse; fewer hospitalisations, reduced self harm and suicide, improved social functioning and enhanced quality of life. This would then have an impact on the contribution of these individuals to society in terms of employment, creativity and reduced use of NHS services. The societal impact of bipolar disorder, and any variation by social status, is unclear and an understanding of this will support service improvement within the UK and internationally. If particular groups are identified as beeing high risk for serious adverse outcomes then targeted interventions could be developed. Service providers could benefit from reduced costs, if a certain medication or intervention is able to reduce hospitalisation and other service use. There would therefore be a reduced cost to the taxpayer. A full analysis of the cost implications of each treatment will be completed, but is considered to be beyond the scope of the current proposal. It may be that the research findings are integrated into updates of existing treatment guidelines and policies (e.g. NICE, British Association of Psychopharmacology) therefore producing an operational change in prescribing and management. This will be pushed for via UCLs links with the Centre for Outcomes Research and Evaluation.

A focus group of service users with a diagnosis of bipolar disorder (established through the pre-existing UCL Service Users Research Forum (SURF) and MDF the Bipolar Organisation) will be invited to comment, initially on the study design and outcome measures, and eventually on the research findings. The group will be consulted at each stage of the research process. It is hoped that this will lead on to future research into health inequality in BPAD and factors effecting medication adherence, as it is recognised that this is another major influence in overall treatment effectiveness. SURF has already been consulted on the design of the project proposal.