It is becoming increasingly important for local councils and health services to take early action to support people to remain independent and stay in their own homes. Our researchers explored whether statistical models can be used to predict an individual person’s future need for intensive social care so they can be supported to remain independent.
One potential way of helping local councils and health services to deal with rising demand for care and support is through the application of predictive models, which use patterns in routine electronic data to identify which individuals in a population are at greatest risk of needing future high cost social care. These models have an important role to play in helping to target preventative care. Similar models have been used for some time in the NHS to look at the chances that somebody will be admitted or readmitted to hospital in the next year.
Current methods of assessing risk of care home admission rely largely on face-to-face assessments. This project explored the feasibility of using routine electronic data linked pseudonymously (with a false name to ensure anonymity), and specifically investigated whether it was possible to predict admission to a care home and high social care costs. This project has shown that it is possible to build similar models that can identify those individuals most at risk of starting intensive social care. The new models are thought to be the first of their type in the world.
This study could help identify those who will need social care support and prompt earlier action to keep people independent and in their own homes
We know from the literature that variables which can influence care home admission include: age, sex, ethnicity, deprivation, morbidity, health service use, drugs prescribed, as well as patterns of social care needs and usage. As these span both health and social care records, a critical first step in this project has been to link health and social care data at an individual level whilst ensuring confidentiality.
This was a significant challenge and we worked with five sites to link up health and social care records for over 180,000 older people. We used these records to see if information about previous health and social care needs and use could help us to predict who will use intensive social care in the next year.
We have now developed a number of prototype predictive models for social care. Overall, our models have relatively good predictive power (i.e. acceptable positive predictive values) but their ability to identify high risk people across the population is relatively poor (i.e. low sensitivity).
The report resulting from this project: Predicting social care costs: a feasibility study, by Dr Martin Bardsley and others, discusses the method and results of this work and we have produced a summary document too. The team has also published some of the results in a paper, published in Age and Ageing (January, 2011).
Aside from the predictive models we developed, this work has generated important lessons about the potential of linked health and social care data to support policy analysis and to guide the planning and commissioning of services.
The Trust has a number of other projects looking at different ways to use such linked data including studies of: social care at the end of life; the lifetime costs of long term care; and evaluation of hospital avoidance schemes.
This project was funded by the Care Services Efficiency Delivery (CSED) unit at the Department of Health.