Changes in early-career family physicians' antibiotic prescribing for upper respiratory tract infection and acute bronchitis: a multicentre longitudinal study

Background: Inappropriate antibiotic prescription and subsequent antibacterial resistance are major threats to health worldwide. Objectives: We aimed to establish whether early-career 'apprenticeship-model' experience in family practice influences antibiotic prescribing for respiratory tract infections and to also establish other associations of antibiotic prescribing changes during this early-career experience. Methods: A longitudinal analysis (2010-2014) of a cohort study of Australian GP registrars' (vocational trainees') consultations. Registrars from five regional training programs recorded data from 60 consecutive consultations, once each 6-month training Term, including the diagnoses managed and medications prescribed. The outcomes were whether an antibiotic was prescribed for the diagnoses 'upper respiratory tract infection (URTI)' and 'acute bronchitis/bronchiolitis'. Generalized linear mixed modelling was used to account for repeated measures on registrars and to include the time component: 'Term'. Results: A total of 856 registrars recorded 108759 consultations, including 8715 'URTI' diagnoses (5.15% of diagnoses) and 2110 'acute bronchitis/bronchiolitis' diagnoses (1.25%). Antibiotics were prescribed in 16.3% [95% confidence interval (CI) 14.9-17.8] of URTI and 72.2% (95% CI 69.6-74.6) of acute bronchitis/bronchiolitis diagnoses. Moving from an earlier to later term did not significantly influence registrars' antibiotic prescribing for URTI [adjusted odds ratio (OR) 0.95; 95% CI 0.87, 1.04, P = 0.27] or acute bronchitis/bronchiolitis [OR 1.01 (95% CI 0.90-1.14), P = 0.86]. Significant associations of antibiotic prescribing for URTIs were the registrar being non-Australian educated, greater patient age, practices not privately billing patients, pathology being ordered, longer consultation duration and the registrar seeking in-consultation information or advice (including from their supervisor). Conclusions: Early-career experience/training failed to produce rational antibiotic prescribing for URTI and acute bronchitis/bronchiolitis. Our findings suggest that prescribing interventions could target the registrar-supervisor dyad