Context affects physicians' antimicrobial prescribing decisions, study finds

Context can play a role in whether physicians decide to prescribe antimicrobials, according to a new study.

Researchers performed a 5.5-year retrospective cohort study in a Midwestern primary care network to determine whether contextual factors were associated with decisions to prescribe antimicrobials for febrile respiratory illness. Specifically, they wanted to test whether the publicity and educational efforts surrounding the H1N1 influenza pandemic in the spring and fall of 2009 made physicians more likely to consider a viral cause of febrile respiratory illness, and less likely to prescribe an antibiotic, than during a typical flu season. The study also examined whether physicians' own recent experiences would have an impact on prescribing decisions. All patients who presented with febrile respiratory illness during influenza seasons between 2006 and 2011 at 26 outpatient practices were included in the study. The results appeared in the Aug. 7 Annals of Internal Medicine.

Overall, the study examined 28,301 unique patient encounters for febrile respiratory illness among 69 physicians. In 12,795 cases (45.2%), patients were prescribed an antibiotic. During the seasonal influenza period, an antibiotic was prescribed in 47.5% of patient encounters, while during the pandemic influenza period, an antibiotic was prescribed in 39.2% of patient encounters (P<0.001). The authors performed multivariable adjustment for patient and physician characteristics and found that antibiotic prescribing remained lower in the pandemic period than in the seasonal period (odds ratio, 0.72). Physicians were also less likely to prescribe an antibiotic if they had seen more patients with febrile respiratory illness in the previous week; the odds ratios were 0.93 for two to three patients seen, 0.84 for four to six patients seen, 0.71 for seven to 11 patients seen and 0.57 for at least 12 patients seen, compared with the reference range of zero to one patient seen.

The authors acknowledged that their study was retrospective and focused on only one geographic area. However, they concluded that epidemiologic variables and the number of recent febrile respiratory illness cases seen affected whether a physician was likely to prescribe an antimicrobial for a patient presenting with this condition. “Even controlling for physician characteristics and the patients' presentation, the epidemiologic context of the visit, and the physician's recent experience with [febrile respiratory illness] (personal context) seemed to strongly influence prescribing of therapy,” the authors wrote. “Our work suggests that tools that increase physicians' awareness of context may help to drive much-needed improvements in prescribing.”

The authors of an accompanying editorial pointed out that the study was not able to answer some important questions, including the kind of epidemiologic knowledge clinicians accessed and how it affected their clinical reasoning and their encounters with patients. “Answers to these questions are certainly beyond the scope of [the current study's] methods, but they would be invaluable in helping us understand exactly where and how to intervene when designing strategies to reduce unnecessary prescriptions,” the editorialists wrote. “To change practices, we need to understand not only which contextual factors shape them but how and why.”

In a related study in another journal, Canadian researchers found that patients with acute respiratory infections who consulted about treatment with a physician trained in shared decision making were less likely to choose antibiotics than those receiving usual care. The study was published online July 30 by CMAJ.