Many people who experience severe mental illness want to work, but this group of people still experience high rates of unemployment (Gühne et al., 2021; Centre for Mental Health, 2013). This is in part caused by barriers to employment such as stigma and discrimination (Centre for Mental Health, 2013) and the inflexibility of the current benefits system in the UK. People with lived experience have identified further problems caused by an outdated limit on earnings and rules about permitted work, which can further disincentivise people from returning to work (Appleton et al, 2021). This 2020 blog about benefits and mental health by Andy Bell is a summary of various papers published between 2016-2020, and is a great introduction to the field.
There is also a growing body of evidence to suggest that this relationship is bi-directional, with benefits policies resulting in a negative impact on claimant’s mental health. For example, the use of benefits sanctions can lead to people struggling to pay their bills or buy necessities like food or clothing, which in turn can negatively impact claimants’ mental health (Wright & Patrick, 2019). A recent systematic review which collated evidence of the mental health impact of social security policies found (perhaps unsurprisingly) that policies which lead to more generous benefits are associated with both reduced inequalities and improvements in claimant’s mental health (Simpson et al., 2021). The opposite was also true, as policies which resulted in lower levels of support or had stricter eligibility criteria were associated with a worsening of mental health, and greater inequalities.
This blog summarises a recent paper by Sharon Stevelink and colleagues, which describes the process of linking mental health data from one mental health trust in the UK with benefits records data from the Department of Work and Pensions (DWP).
Data linkage involves linking two or more datasets which contain different information about the same individuals. The data used is from electronic records, and any data that can identify individuals is removed prior to linkage. Mental health data from January 2007 to June 2019 was linked with benefits data from January 2005 to June 2020. The actual linkage took place in late 2020 and was achieved through linking people’s national insurance numbers (held in the DWP dataset) with a pseudonym created for each record in the mental health dataset.
This study focused on those who had been referred to secondary mental health services in the South London and Maudsley NHS Trust; people who were likely to have more severe mental health problems. The study team consulted with a lived experience advisory group before the data linkage took place to discuss and approve their plans for linking the two datasets. This group will also be involved in any future research involving this dataset to discuss ideas for research questions and help with dissemination of the findings.
Participants were classed as receiving a benefit if they had received one of a range of benefits active between January 2005 and June 2020. The DWP dataset also included information on what type of Universal Credit conditionality regime patients were allocated to e.g. searching for work, no work requirements etc.
Combining both the mental health and benefits datasets allowed for missing data in one dataset to be filled in, if complete in the other. The overall linkage rate was calculated by assessing the proportion of pseudonym IDs which were successfully linked to a national insurance number. Further statistical analyses were conducted to explore factors associated with receiving benefits, whilst controlling for variables such as age, sex and ethnicity. The authors also described the benefits use of mental health service users.
The paper reports on data from 239,714 people who had accessed specialist mental health services.
The study team were able to achieve a high level of data linkage between the two datasets, with 92.3% of data successfully linked. Socio-demographic factors which meant data were less likely to be linked included being female, being from an ethnic minority group, being middle aged (defined as age 21-60 years), or only having one postcode available (rather than two or more). People who were over 60 years of age were more likely to have their data successfully linked than other adults.
Contact with DWP
Of those whose data were successfully linked, 83.6% had engaged with the DWP, and of those, 99.8% had received a benefit. Of those service users who received benefits, 85.1% received two or more different benefits.
Most of the patients had a primary psychiatric diagnosis recorded in their electronic healthcare record, the most common being a mood disorder (around 20%) followed by substance abuse disorders (17.5%) and disorders due to physiological conditions, e.g. dementia (17.4%).
The types of benefit received by people was varied by diagnosis. For example, the majority of patients diagnosed with an intellectual disability received disability benefits such as Employment and Support Allowance and Disability Living Allowance, or Income Support benefits such as Income Support and Personal Independence Payment. Similar types of benefits were also received by patients diagnosed with pervasive and specific developmental disorders and patients diagnosed with schizophrenia or psychotic disorders. In contrast, people with substance abuse disorders were more likely to receive unemployment benefits such as Job Seekers’ Allowance.
This paper details the first linkage of a mental health and a benefits dataset. Results show that a high percentage of service users who have been in contact with secondary mental health care services had also received a benefit during the 15-year data linkage period. This novel linked dataset unlocks the potential for future research studies to focus on specific research questions for mental health service users who receive benefits, with the aim of improving future outcomes.
Strengths and limitations
This study reported a very high rate of data linkage of 92.3%. One particular strength of the approach used by the research team was the involvement of a lived experience advisory group in planning the data linkage. The involvement of people with lived experience is important to ensure that research is conducted in a way that does not exclude the needs of service users or carers and maximises the usefulness of research findings to the population they represent.
However, there were also some limitations reported regarding data availability. For example, the data sets did not allow the study team to distinguish between different reasons for people stopping benefits such as returning to work or disengaging from the benefits system. The authors also reported that this issue may disproportionately affect disadvantaged groups such as people who are homeless or refugees, as they are more likely to have disengaged from the benefits system but not be in work. There were also some inequalities in whether linkage was achieved, with women, people from an ethnic minority group, or those who were middle aged, less likely to have had successful data linkage, which may lead to biases in future research studies using the linked data. The current data linkage also only includes service users who have accessed secondary mental health services, there is no information for people who may have received other forms of mental health care. Finally, the mental health data was only from one mental health trust in one area of the UK. It is therefore unclear whether similar high rates of benefits contacts amongst people accessing secondary mental health services occurs in other areas of the country.
Implications for practice
As the relationship between poor mental health and receiving benefits is thought to be bi-directional, it is important to understand more about how people with mental illness can be supported to return to work, whilst minimising the psychological harms caused by certain aspects of the benefits system.
Given the high rates of involvement with the benefits system amongst this population, clinicians should be aware of their clients’ potential engagement with the benefits system and any associated distress caused by systemic issues.
Implications for research and policy
This novel data linkage paves the way for further research in this area. For example, having access to a large dataset containing 15 years’ worth of linked benefits and secondary mental health service use data allows for research exploring the longitudinal relationships between mental health and benefit receipt. Future research using this dataset could also explore the impact of the introduction of certain benefit policies (such as Universal Credit) on the mental health of service users. Expanding the data linkage to include the wider population could also allow for comparisons in mental health outcomes (e.g. suicides) between benefit recipients and those not in contact with the welfare system.
In light of the current cost of living crisis disproportionately affecting those on lower incomes, policy makers should examine the available data to explore how the benefits system can be amended to improve the mental health of those in receipt of social security benefits.
Statement of interests
Stevelink, S.A., Phillips, A., Broadbent, M., Boyd, A., Dorrington, S., Jewell, A., Leal, R., Bakolis, I., Madan, I., Hotopf, M. and Fear, N.T., (2023). Linking electronic mental healthcare and benefits records in South London: design, procedure and descriptive outcomes. BMJ Open, 13(2), p.e067136.
Gühne, U., Pabst, A., Löbner, M. et al. Employment status and desire for work in severe mental illness: results from an observational, cross-sectional study. (2021) Social Psychiatry & Psychiatric Epidemiology 56, 1657–1667 https://doi.org/10.1007/s00127-021-02088-8
Centre for Mental Health (2013) Briefing 47. Barriers to Employment
Appleton, R., Barnett, P., Allman., F & Lloyd-Evans, B. (2022) IMPACTS OF THE SOCIAL SECURITY SYSTEM ON CLAIMANTS’ MENTAL HEALTH AND WELLBEING, AND HOW MIGHT HARMS BE MITIGATED: NIHR Mental Health Policy Research Unit Report.
Wright, S., & Patrick, R. (2019). Welfare Conditionality in Lived Experience: Aggregating Qualitative Longitudinal Research. Social Policy and Society, 18(4), 597-613. doi:10.1017/S1474746419000204
Simpson, J., Albani, V., Bell, Z., Bambra, C. and Brown, H., (2021). Effects of social security policy reforms on mental health and inequalities: a systematic review of observational studies in high-income countries. Social Science & Medicine, 272, p.113717.
- Photo by Tamanna Rumee on Unsplash
- Photo by Kelly Sikkema on Unsplash https://unsplash.com/photos/3-Tc_5LROrM
- Photo by Joshua Sortino on Unsplash https://unsplash.com/photos/LqKhnDzSF-8
- Photo by Priscilla Du Preez on Unsplash https://unsplash.com/photos/F9DFuJoS9EU
- Photo by Marten Bjork on Unsplash https://unsplash.com/photos/6dW3xyQvcYE
- Photo by Nick Fewings on Unsplash https://unsplash.com/photos/zF_pTLx_Dkg