RESUMO
We aimed to evaluate the effect of pain education on opioid prescribing by early-career general practitioners. A brief training workshop was delivered to general practice registrars of a single regional training provider. The workshop significantly reduced "hypothetical" opioid prescribing (in response to paper-based vignettes) in an earlier evaluation. The effect of the training on "actual" prescribing was evaluated using a nonequivalent control group design nested within the Registrar Clinical Encounters in Training (ReCEnT) cohort study: 4 other regional training providers were controls. In ReCEnT, registrars record detailed data (including prescribing) during 60 consecutive consultations, on 3 occasions. Analysis was at the level of individual problem managed, with the primary outcome factor being prescription of an opioid analgesic and the secondary outcome being opioid initiation. Between 2010 and 2015, 168,528 problems were recorded by 849 registrars. Of these, 71% were recorded by registrars in the nontraining group. Eighty-two percentages were before training. Opioid analgesics were prescribed in 4382 (2.5%, 95% confidence interval [CI]: 2.40-2.63) problems, with 1665 of these (0.97%, 95% CI: 0.91-1.04) representing a new prescription. There was no relationship between the training and total prescribing after training (interaction odds ratio: 1.01; 95% CI: 0.75-1.35; P value 0.96). There was some evidence of a reduction in initial opioid prescriptions in the training group (interaction odds ratio: 0.74; 95% CI: 0.48-1.16; P value 0.19). This brief training package failed to increase overall opioid cessation. The inconsistency of these actual prescribing results with "hypothetical" prescribing behavior suggests that reducing opioid prescribing in chronic noncancer pain requires more than changing knowledge and attitudes.