The impact of primary care incentive schemes on care home placements for people with dementia

Objectives: the interface between primary care and long-term care is complex. In the case of dementia, this interface may be influenced by incentives offered to GPs as part of the Quality and Outcomes Framework (QOF) to provide an annual review for patients with dementia. The hypothesis is that the annual reviews reduce the likelihood of admission to a care home by supporting the patient to live independently and by addressing carers’ needs for support. Study period: 2006/07 to 2015/16. Outcomes: admissions to a care home. Data: to analyse the impact of annual reviews on care home admissions, we used a linked individuallevel dataset covering primary care, secondary care, and mortality from the ResearchOne database (N=30,216). The data provider supplied data on care home events, which they had derived by linking patient postcodes on SystmOne, the clinical system, to the Care Quality Commission’s database of care home addresses, then checking the first line of the patient’s address for relevant key terms (trigger words). To help interpret our findings, we identified the types of event occurring at three points in time: the day of the annual dementia review, within a week and within a month of the review. We used the Clinical Practice Research datalink (CPRD) (N=5,169) to investigate these events. Methods: a survival analysis model was employed to exploit information on the timing of events, in particular whether QOF-incentivised reviews preceded or followed care home admission. We estimated the hazard of the first care home admission as a function of time-varying annual reviews and comorbidities, and time-invariant demographics and local area characteristics. To identify events around the annual review, we used CRPD to identify visit locations (e.g. face-toface appointment in the clinic - based on the consultation type code) and broad types of care (e.g. examination - based on Read codes). We measured events on the day of the review, and within the subsequent 7 days and 28 days to capture follow-up activity. Results: No association was found between annual reviews and an individual’s hazard (‘likelihood’) of care home placement, even after conditioning on time-varying proxies of severity. Higher likelihood of a care home admission was associated with older age at time of diagnosis, female gender, and with having certain long-term conditions, including atrial fibrillation, epilepsy, and stroke. Patients on the palliative care QOF register were also more likely to have a care home placement. Patients living in areas in the lowest quintile of deprivation (most deprived) had a significantly lower likelihood of care home placement than those living in areas in the least deprived quintile. Compared with residents in the East region, patients in London and the West Midlands were significantly less likely to have a care home admission, and those in the North East, North West and Yorkshire and the Humber were significantly more likely to be admitted. Most reviews took place in the GP surgery (85%) or at home (13%). During the review, GPs examined patients’ physical and mental health, ordered tests and made referrals. Carer details were rarely recorded during the review (0.7% of all review events). There was substantial evidence of follow-up activity by the GP practice over the week and month following the review. Conclusions: Our study found no evidence of an association between the QOF annual dementia review and the likelihood of care home placement. Our analysis utilised a relatively new primary care dataset containing large numbers of individuals with dementia. The outcome measure was robust, and our model incorporated the timing of annual reviews, events and morbidities. However, this was an analysis of observational, not randomised, data and we were unable to control for some confounding factors. The annual review was assessed using a binary measure, so we could not test whether the quality of the review influenced the risk of care home admission. Therefore, findings should not be interpreted as definitive evidence of the absence of a relationship between the annual dementia review and the likelihood of care home placement