Disease, temporal and sociodemographic influences on initial treatment of community-acquired pneumonia in Manitoba, Canada.

Reimbursement restrictions on newer antibiotics, common to many drug plans, may result in unnecessary hospitalisation when patients are unable to pay 'out of pocket' for recommended antibiotics. We examined the effect of income, among other subject characteristics, on the likelihood of hospitalisation or receipt of restricted antibiotics for initial treatment of community-acquired pneumonia (CAP). A retrospective cross-sectional review of healthcare claims from the province of Manitoba, Canada, from 1 May 1996 to 1 March 2002 was conducted. Of 36969 subjects with a diagnosis of CAP, 13.6% were initially hospitalised and 86.4% were treated as outpatients. Independent predictors of initial hospitalisation included: age (for every 10-year increase) (odds ratio (OR)=1.42); male gender (OR=1.22); urban residence (OR=0.52); presentation to the emergency department (OR=5.14); and level of co-morbidity (high versus low) (OR=1.66). The effect of income level on hospitalisation was modified by co-morbidity status. Among subjects lacking CAP-specific co-morbidities, the probability of initial hospitalisation was greater in the lowest versus the highest income quintile (OR=1.87). Among outpatients, restricted antibiotics were widely received and differences in use by income were modest.