Identifying predictors and measuring variation in hospitalisation for older adult care home residents: a retrospective cohort study using routinely collected data

Background

There over 250 000 care home residents in England and they account for a disproportionate number of hospital admissions. Since 2016, enhanced services for care homes have been developed to reduce admissions but there is regional variation in provision. We used a nationally representative cohort of care home residents to identify predictors of admissions and measure regional risk adjusted variation.

Methods

Using primary care data from a third of practices in England, we derived a cohort of over 40 000 care home residents linked to the national regulator’s register of care homes between 01/01/2023 and 31/12/2024. We estimated a three-level negative binomial regression model including resident, postcode district and commissioner (Integrated Care Board, ICB) level factors which predicted 6-monthly count of admissions. Using this model, we estimated funnel-plots to measure case-mix risk adjusted variation in admission rates at postcode district and commissioner geographical levels.

Results

Significant predictors of admission count included age, gender, frailty, presence of feeding tube, long-term catheterisation, polypharmacy, anticholinergic burden, sedative load, care home rating and proportion of care home jobs filled. Considerable variation in risk adjusted admission rates at commissioner (ICB) level was demonstrated. ICB outliers ranged between having 0.6 and 2.2 times as many admissions as would be expected based on case-mix.

Conclusion

This is one of the largest studies of care home residents in England. We identified potentially modifiable risk factors for admissions and significant variation in ICB level admission rates. This variation could be the result of differences in enhanced services provision and requires further research.