Background Appropriate and timely social care can potentially delay or avoid the need for interventions such as emergency hospitalisation. Routine data provides scope to examine the relationship between emergency hospital admissions in Wales and local authority-supported social care for people aged 65 and over.

Methods This observational study employed mixed methods for data collection and analysis covering the period January 2006 to December 2012. We used interrupted time series analyses to study overall trends in the use of social care and unplanned use of secondary care across Wales. Within one case study area, we used the SAIL database to link health and social care data anonymously, to examine service use at an individual level.

Results At an all-Wales level, we observed no relationship between trends in health care and social care for older people.

In our case study area, we examined data on ‘qualifying emergency admissions’ (QEAs), that is, the first emergency admission to hospital in the study period for a person aged 65 or over. Over the study period, about four-fifths of all the people receiving social care (n=10,864) also had a QEA.

We compared those receiving social care before the QEA (n=3362) with those who received social care only after their QEA (n=7478). Those receiving social care before the QEA were likely to:

▸ have fewer subsequent admissions (1.4 v 2.9, p<0.01).

▸ have a shorter length of stay for all subsequent admissions (41.5 v 66.4 days, p<0.01)

Conclusions Anonymously linking health and social care data for research is feasible and has potential for being taken further. Linked data reveals more about the relationship between health and social care use than unlinked data sets.

Social care appears targeted appropriately at the frailest older people, and seems to have a protective effect in reducing emergency admissions and length of stay following an admission.