A shared–decision-making model governing physician-patient communication is receiving increasing emphasis in clinical medicine. Applying a patient-centered approach may be of benefit in managing acute otitis media because of the varied beliefs about whether antibiotics are necessary. Acute otitis media is a major cause of high antibiotic prescription rates in the United States and increasing use of broad-spectrum antibiotics. Evaluation of benefit, cost, and antibiotic resistance has led to expert recommendations of watchful waiting and delayed, “safety-net” antibiotic prescriptions as a strategic option in nonsevere cases. Merenstein and colleagues compared physician communication styles to determine whether they affected parents' demand for antibiotics or satisfaction with the visit.

Participants were parents in a research network who each had been the primary caregiver of a child younger than five years. Block groups of participants were randomized to receive a cross-sectional survey containing one of three vignettes representing a continuum of communication styles from a paternalistic to a shared-decision approach. The survey also included questions about demographics, the relationship with the child's physician, knowledge of antibiotic treatment, and interest in shared decision making. The three vignettes presented the same scenario and three management options (treating, not treating, or waiting). In the first model (paternalistic/antibiotic), the physician recommended antibiotic use, in the second (shared–decision-making group) the physician allowed the parent to decide, and in the third the physician recommended acetaminophen and gave a prescription that could be filled if the child did not improve. Participants were asked to imagine being the parent in the vignette and to rate the visit in terms of their overall satisfaction, their satisfaction with the amount of information given, and their likelihood of using antibiotics for acute otitis media.

Of 883 parents who agreed to participate, 466 were eligible and were allocated randomly to one of the three groups. Characteristics among groups were similar except for a higher percentage of ethnic minorities in the shared–decision-making group. The majority of parents were currently taking care of a child younger than five years and had been to a physician with their child in the previous year. All groups showed a high preference for shared decision making.

Satisfaction with the visit was rated lowest in the paternalistic/antibiotic group, in which 19 percent of parents expressed disagreement with a statement of satisfaction, compared with 8 percent in the acetaminophen group and 2 percent in the shared–decision-making group. Although 93 percent of parents initially thought an antibiotic was required for an ear infection, only 14 percent thought so after reading the vignettes; however, 27 percent of parents in the paternalistic/antibiotic group said they would want immediate antibiotics, compared with 7 percent in the other groups. The paternalistic/antibiotic group was 4.9 times as likely to use antibiotics as the acetaminophen group, and this likelihood was even greater after adjusting for potential confounders.

The authors conclude that communication style has a powerful effect on the likelihood that antibiotics will be used and can significantly reduce the belief that antibiotics are necessary for treatment of acute otitis media. Previous studies have shown the benefits of educational interventions. The results of this study show, in addition, that shared decision making leads to greater parent satisfaction and a reduced desire for antibiotics.

editor's note: This study shows that patients prefer a shared–decision-making approach; however, whether “safety-net” prescriptions reduce antibiotic use because of the negotiation inherent in this approach or for other reasons needs further study.¹ Similarly, it is unclear whether antibiotic use would have been different if the paternalistic approach had focused on avoiding antibiotics instead of recommending them. —c.w.