Investigating the impact of interventions on social care: a quantitative approach using data linkage and joint modelling
End of project summary
Main messages
Background
The intention of the fellowship was to increase the capacity of quantitative expertise in social care research. This would be enabled through the use of anonymised linked data within the Secure Anonymized Information Linkage (SAIL) databank. The COVID-19 pandemic started during the fellowship, which re-priotised some research activities towards care homes and the spread of COVID-19.
Methods
The fellowship enabled the learning and use of advanced statistical techniques including multilevel regression modelling, survival analysis, joint modelling and geo-spatial modelling. Many collaborations were created between universities, government, the NHS, and third-sector, with many created to supply evidence to help with the COVID-19 pandemic. The fellowship also included many reproducible research methods within the SAIL databank, including the linkage of anonymised care home addresses to individuals.
Results
Throughout the fellowship statistical techniques were applied to a variety of research areas. This included geriatric medicine, as well as multiple applications to the COVID-19 pandemic. Of note, this included work with Care & Repair Cymru, as well as analyses on the impact of the pandemic on care homes in Wales, and high-fidelity geospatial modelling of the spread of COVID-19 across Wales.
The work with Care & Repair Cymru investigated the impact of home interventions on the chance of a fall for older people in Wales. This was compared against all individuals in Wales aged 60+ who did not receive a service from Care & Repair Cymru. The analysis used multilevel modelling to analyse longitudinal data for over half a million older people in Wales. The results showed that although individuals receiving a service from Care & Repair Cymru were more likely to have a fall, that after receiving an intervention the chance of having a fall reduced relative to those without a Care & Repair service.
Work on care homes in Wales included collaborations with public health wales, UK government, the NHS, and care home representatives. The analyses included the change in mortality in care homes between 2016-2020, analysing nosocomial seeding of COVID-19 within care homes, and the effect of vaccination status on the likelihood of testing positive for COVID-19. The results of these analyses were presented to Welsh government and the carehome subgroup for the UK Scientific Advisory Group for Emergencies (SAGE).
The analyses of the spread of COVID-19 within Wales was achieved with a collaboration between Swansea University, Lancaster University, and Health Data Research UK. The resulting analyses investigated the spread of COVID-19 over space and time using COVID-19 testing data. The results were presented to Welsh government and were subsequently used by the first minister and televised.
Conclusion
The fellowship enabled the learning and application of advanced statistical techniques. This was then transferred to multiple applications including priority areas to aid government with decision-making with respect to COVID-19.