The UK’s leading statisticians have recognised linking data from different trusted sources as an innovative way of ‘enhancing our understanding of society, driving policy change for greater public good and minimising respondent burden’. 

The Social and Economic Predictors of severe Mental Disorders (SEP-MD) study takes up this call. Launched in 2018, our initiative is led by Principal Investigator Dr Jayati Das-Munshi and supported by SLAM-Trust BRC Clinical Data Linkage Service and the ESRC. It is now also part of the ESRC Centre for Society and Mental Health cohorts platform. Our project connects mental health hospital records with Census data to explore the social and economic dimensions of mental health conditions. But before we get to our findings, it’s worth asking…

What’s so special about data linkage?

Data linkage has rarely been used on a large scale but has tremendous potential to give rich detail on people’s social circumstances, which is otherwise missing from health records. This is important because people’s social experiences play a fundamental role in mental health.

Large scale routine health records can shed light on conditions which are less well captured in national surveys or smaller survey-based studies. Conditions like schizophrenia or bipolar disorders, which are normally designated as 'severe mental illness' (SMI), are relatively rare in national surveys and often not captured in adequate numbers to enable inferences. Routine health records from secondary mental health Trusts do hold information about people living with these conditions at sufficient scale; the SEP-MD linkage, for example, included 20,000-30,000 people with SMI, many more than would be included in a national survey.

Previous studies led by Dr. Jayati Das-Munshi confirm that ‘people with severe mental illness have higher mortality rates, culminating in about 20 years of lost life compared with that of the general population’. These appalling findings contribute to our understanding of the ‘social gradient’ of health, famously demonstrated in Marmot’s Whitehall studies. These showed that people in more deprived areas, or considered to be of lower ‘social class’, are more likely to experience adverse health outcomes.

Studies like ours aim to unpack these deadly inequities, and data linkage allows us to analyse overlapping layers of disadvantage usually hidden from view.

Why is the project is so innovative?

The linkage, which was supported by the former Administrative Data Research Centre (ADRC) and the Economic and Social Research Council (ESRC) and the South London and Maudsley Trust Clinical Data Linkage Service (SLaM CDLS), is innovative in several ways.

First, it focuses specifically on mental health, drawing from the records of South London and Maudsley (SLaM) NHS Trust, one of Europe’s largest secondary mental health service providers. These records, available through the Clinical Records Interactive Search (CRIS), have had all identifiable information removed and have been ethically approved for researchers to access. This means that our data analysis can inform research and service improvements while keeping the identities of all services users protected and confidential.

Second, the project draws from a dense and diverse catchment area; only 38.7% of SLaM service users reported their ethnicity as ‘White’ in 2020-21. Many other surveys and cohort datasets are unable to produce sufficient data on ethnicity to draw significant conclusions. The SEP-MD study, however, includes enough records from patients identifying with minority ethnic groups to shed light on ethnic inequalities in mental health. SLaM’s catchment also includes several of the most deprived neighbourhoods in England, representing one extreme of the ‘social gradient’.

Third, the SEP-MD linkage makes use of the momentum generated by the pandemic. COVID-19 exposed stark ethnic and racial inequalities across health and social care, and the government’s pandemic response increased public awareness of the importance of national-level data collection. The King’s-ONS linkage was completed in early 2020, and enables a perspective on inequities prior to the pandemic. In the future we may consider refreshing the linkage to provide a perspective on changes since 2020.

Lastly, SEP-MD’s focus on social and economic predictors — based on data supplied by the Census on factors like employment — aligns with the interests and priorities of many patients and carers. From clinical experience and through consultations with mental health service user groups, the King’s researchers learned that employment, joblessness and household-level need – among other factors – shaped experiences of illness and access to care. Data linkage allows us to improve our understanding of these social conditions through quantitative metrics, while making sure these findings stay relevant to those affected by mental illness.

Stay in the loop

Our data analysts – PhD candidate Rosie Hildersley and postdoctoral associate Dr Lucasz Cybulski – are now hard at work cleaning and describing the linked dataset and exploring what data is missing. We are documenting each step of the process to make sure our study is transparent and reproducible. More updates soon!

Email milena.wuerth@kcl.ac.uk for opportunities to collaborate.

Follow the Centre on Twitter (@kcsamh) or visit kcl.ac.uk/research/the-sep-md-data-linkage-study for more information.