The relationship between social care resources and healthcare utilisation by older people in England: an exploratory investigation

Maria Lucia Pace, Dan Liu, Maria Karen Goddard, Rowena Jacobs, Raphael Wittenberg, Gerard McGonigal, Anne Rosemary Mason

Research output: Working paper


Background Since 2010, adult social care spending has fallen significantly in real terms whilst
demand has risen. Reductions in local authority (LA) budgets are expected to have had spill over
effects on the demand for healthcare in the English NHS.
Motivation If older people, including those with dementia, have unmet needs for social care, their
use of healthcare may increase.
Methods We assembled a panel dataset of 150 LAs, aggregating individual-level data where
appropriate. We tested the impact of changes in LA social care resources, which was measured in
two ways: expenditure and workforce. The effects on people aged 65+ were assessed on five
1. Rates of emergency hospital admissions for falls in people with dementia aged 65 and
2. Rates of emergency hospital admissions for fractured neck of femur in people 65 and
3. Extended length of stay in people with dementia, 7 days and over
4. Extended length of stay in people with dementia, 21 days and over
5. Rates of NHS Continuing Healthcare (NHS CHC)
Outcomes (utilisation) data were derived from the Hospital Episode Statistics (1, 2, 3 and 4), the
Public Health Outcomes Framework (2), and publicly available datasets from NHS Digital (5).
Datasets varied in the timeframes available for analysis. Planned analysis of the effects of social care
cuts on delayed transfers of care in mental health trusts, and on deprivation of liberty safeguards
were not undertaken because of data quality concerns.
We tested the effect of two separate explanatory variables: adult social care gross current
expenditure (per capita 65 and over) adjusted by area cost; and adult social care workforce staff (per
capita 18 and over). Workforce measures distinguished LA and independent sector employees and
included professional and non-professional staff providing direct social care. We ran negative
binomial models and linear models, and controlled for a range of confounding factors, including
deprivation, ethnicity, age, unpaid care, LA class and year effects. To account for potential
endogeneity (‘reverse causality’), we also tested the Area Cost Adjustment (ACA) as an instrumental
variable and ran dynamic panel models. Sensitivity analysis explored the effects of the additional
effects of the Better Care Fund.
Results The level of social care expenditure on older people was not significantly related to
emergency admission rates for falls in people with dementia or for fractured neck of femur.
Extended stays of 7 days or longer were significantly and positively related to the level of social care
spend, but this association was no longer significant when additional spend from the Better Care
Fund was taken into account. There was no significant relationship between the level of social care
spend and hospital stays of 21 days or longer or between spend and uptake of NHS CHC.
We also tested the effect of four social care workforce measures. LAs employing higher rates of
social care staff (especially professional staff) had significantly higher levels of NHS CHC, but there
was no significant relationship between LA staffing levels and the remaining four outcomes. LAs with
higher levels of independent social care staffing had significantly lower rates of extended stays, but
there was no association with either emergency admissions or on NHS CHC. The effect of ‘full time’
ii CHE Research Paper 174
unpaid care on outcomes was mixed, with tentative evidence of a protective effect on admissions
for falls, and on extended stays of 21 days or longer.
When the Area Cost Adjustment was used as an instrument in place of expenditure, results were
largely consistent with the main analysis: there were negative effects on NHS CHC but no effect on
any other outcome. The dynamic panel models found a positive relationship between spend and
emergency admissions for falls, but the effect on other outcomes was statistically insignificant.
Conclusions The study found no consistent evidence that reductions in social care budgets led to the
expected rises in hospital admissions, hospital stays or uptake of NHS CHC. However, findings
suggest that public sector staff providing direct social care, particularly professional staff, may be
instrumental in facilitating access to NHS CHC. In addition, the study found tentative evidence that
extended hospital stays are partially offset by social care provision by the independent sector and by
unpaid carers providing intensive care. To test the validity and robustness of these findings, future
research using linked individual-level health and social care data is needed.
Original languageEnglish
Place of PublicationYork, UK
PublisherCentre for Health Economics, University of York
Number of pages37
Publication statusPublished - 4 Nov 2020

Publication series

NameCHE Research Paper
PublisherCentre for Health Economics, University of York


  • Social care
  • Healthcare
  • Dementia
  • Local authority
  • Cost Shifting

Cite this