A multi-state transition model among older adults receiving home care services: A population-based cohort study

Background Older adults with loss of autonomy experience multiple care transitions, some of which can be burdensome. While home care services (HCS) are thought to support aging in place, little is known about the patterns of care transitions over time among this complex population. The objective is to describe transitions and quantify probabilities of transitioning, using multi-state Markov models in a population-based cohort from Quebec (Canada). Methods This is a retrospective observational cohort study, selecting recipients of HCS within a program for older adults with loss of autonomy. We estimated transition probabilities with continuous-time Markov models across the following states: A) home, B) hospital, C) at home waiting for long-term care (LTC) placement, D) LTC, and E) death. Sex, age, type of residence, and comorbidities were included in models as covariables. Results Among the 3424 HCS recipients, 65.7% were females, while males were younger on average and had more comorbidities. Several factors were significantly associated with specific transitions, such as higher rates of hospital admissions for males and higher rates of LTC transfers upon hospital discharge for males and patients with dementia. HCS were also strikingly lower in intensity, and with fewer services provided by healthcare professionals, for patients waiting at home for LTC placement, being wait-listed itself associated with dementia and residing in private senior homes. Conclusions This study highlights gaps, and hence potential areas for improvement, in HCS for older adults with loss of autonomy in Quebec and shows how multi-state models can enhance understanding of care transitions in complex populations.