This presentation is part of: Oral Concurrent Session A - Judgment and Decision Making
PRACTITIONERS’ INSIGHT INTO THEIR OWN DECISIONS ABOUT PRESCRIBING ANTIBIOTICS IN RESPIRATORY INFECTIONS
Robert S. Wigton, MD, MS1, Carol A. Darr, BA2, Bonnie A. Leeman-Castillo, MS2, and Ralph Gonzales, MD, MSPH3. (1) Univ of Nebraska Medical Center College of Medicine, Internal Medicine, Omaha, NE, (2) University of Colorado at Denver, Health & Behavioral Sciences Program, Westminster, CO, (3) University of California, San Francisco, Medicine, San Francisco, CA
Purpose: Understanding how physicians decide whether to prescribe antibiotics is a key step in reducing antibiotic use in respiratory infections (RI). It is not known, however, whether practitioners understand their own use of clinical information in making this decision. Because recent studies have suggested, contrary to prior reports, that decision makers can recognize their own policies, we hypothesized that primary care practitioners (PCP) would be able to distinguish their own decision policy for prescribing antibiotics in RI from that of other PCPs. Methods: Each of 81community PCPs reviewed 20 case vignettes of patients with RI constructed using 8 variables in a fractional factorial design. For each case, they decided whether to prescribe antibiotics. For each PCP, we determined the weight they gave each clinical cue in deciding about antibiotics (the decision policy). We used cluster analysis to define 9 groups of practitioners with similar policies and chose a policy closest to the mean cue weights of the group to be the group archetype. We showed histograms of the 9 archetypes to each PCP with a histogram of their own policy substituted for the archetype of the cluster to whic they belonged. We asked them to select their own policy from the display. Results: Of 81 PCPs who completed the profiles, 54 participated (67%). When asked to identify their own from the 9 policies displayed, 9 of the 54 participants (11%) correctly identified their own, a rate no better than chance (p = 0.93). Conclusions: PCPs could not distinguish their own policy for prescribing antibiotics from archetypes of the 8 other approaches taken by 81 PCPs. These other approaches differed significantly in the pattern of weighting. This indicates that PCPs were not sufficiently aware of how they use clinical variables in deciding about antibiotics – an important consideration in changing prescribing behavior. Thus, approaches to improve prescribing may need to include not only information about the optimal use of clinical information in making the decision for antibiotics but also making practitioners aware of their own decision processes.
