Matching on health status to estimate the effects of outpatient care and social factors in patients with COPD: a Norwegian registry‑based study

Isolating the effects of outpatient care on hospitalisation in observational data is challenging. We estimate the effects of factors believed to influence hospital admissions in patients with COPD. The factors were related to outpatient care and social factors, motivated by recommendations in COPD guidelines and the literature. We applied linked national and municipal registry data on 25,369 COPD patients in Oslo and Trondheim 2008–2019. Yearly frequency and type of healthcare contacts (respiratory and non-respiratory, across providers), together with comorbidities, were used as proxies for patients’ health status. We constructed samples matched on the full set of proxy variables to estimate the effect of each factor on future COPD hospital admissions. Overall, 9.8% of follow-up years had at least one COPD hospital admission. Findings indicated beneficial effects of several of the factors. Care interaction between general practitioner (GP) and other health providers, follow-up by GP within 30 days after COPD hospital discharge and counselling/psychotherapy by GP showed significant reductions of 2.9, 4.9 and 1.4% points in the rate of next-year COPD hospital admissions, respectively. The effects are noteworthy when extrapolated to the entire Norwegian COPD population. A reduction of 1% point could result in up to 1,500 fewer patients experiencing one or more COPD hospital admissions during a year. In the absence of detailed clinical information, creating samples that are multidimensionally matched on health service use and comorbidities as pragmatic proxies for health status could be a viable approach for estimating the effects of other factors.