Methods: Retrospective prescribing decisions for emergency department patients with a URI diagnosis over a 17-month time period were combined with survey data on categorical gists that underlie clinician prescribing decisions in a multivariate regression model.
Results: Antibiotic prescribing was highly variable (figure). A total of 94 clinicians completed the survey (61% completion rate). The categorical gists relating to side-effects and risk-taking captured the majority (66%) of variance in the survey results, while satisfying patients was not regarded as significant. Analysis >4,000 patient URI encounters found that variability in antibiotic prescribing rate was most significantly explained by categorical gists of the prescribers. In particular, greater agreement with statements capturing the “why not take a risk” categorical gist had increased rates of prescribing. Those who did not agree with the gist relating to side effects also had higher prescribing rates.
Conclusion: Our results provide evidence that clinicians’ prescribing decisions are not principally motivated by a desire to satisfy patients, but by a strategic bias assuming that the possibility of effective treatment trumps risk of antibiotic-associated side effects. Greater regard for the significance of side effects ameliorates this bias. Interventions that emphasize the non-negligible possibility of serious side-effects of antibiotic use may thus be more effective than current approaches using education to reduce inappropriate prescribing.
L. May, None
M. Saheed, None
V. Reyna, None
D. Broniatowski, None