LAUSR

Abstract Objective To investigate the correlation between having designated general practitioners (GPs) in residential care homes and the residents’ number of contacts with primary care, number of hospital admissions and mortality. Design A retrospective register-based longitudinal study. Setting Forty-two care homes in Aarhus Municipality, Denmark. Subjects A total of 2376 care home residents in the period from 1 September 2016 to 31 December 2018. Main outcome measures We used two models to calculate the incidence risk ratio (IRR) for primary care contacts, hospital admission or dying. Model 1 compared the residents’ risk time before with their risk time after implementation of the designated GP model. Model 2 included only risk time after implementation and was based on calculations of successful (rate ≥60%) implementation. Results Weighted by time at risk, the proportion of females across the two models ranged from 64% to 68%. The largest group was aged ‘85-94’ years. In Model 1, the mere implementation of the model did not correlate with changes in primary care contacts, hospital admissions, or mortality. Contrarily, in Model 2, residents living in care homes with successful implementation had fewer email contacts (IRR = 0.81, 95%CI: 0.68;0.96), fewer telephone contacts (IRR = 0.78, 95%CI: 0.68;0.90) and fewer hospital admissions (IRR = 0.85, 95%CI: 0.73;0.99), but more home visits (IRR = 1.70, 95%CI: 1.29;2.25) than residents living in care homes with lower implementation rates. Conclusion The designated GP model seems promising, as a high implementation degree of the model correlated with a reduced the number of acute admissions, short-term admissions and readmissions. Future studies should focus on gaining deeper insight into the mechanisms of the designated GP model to further optimize the model. Key points A new care model was introduced in Denmark in 2017, designating dedicated GPs to residential care homes for the elderly. Successful implementation correlated with significantly fewer hospital admissions, specifically for acute admissions, but also with fewer short-term admissions and readmissions. The implementation of the model correlated significantly with fewer e-mail and telephone contacts and with more home visits. Future studies should gain more insight into the mechanisms of the designated GP model to further optimize the model.