Am Fam Physician. 2003;68(2):370-371
Many physicians prescribe antibiotics unnecessarily, most notably for acute respiratory tract infections, but also for conditions such as otitis media or sinusitis, in which antibiotic benefit is likely to be marginal. Little is known, however, about the factors involved in choosing which type of antibiotic to use. Steinman and colleagues conducted a study to determine which factors were associated with the prescription of broad-spectrum antibiotics for nonpneumonic acute respiratory tract infections.
Using data from the National Ambulatory Medical Care Survey, the authors analyzed 1,981 patient visits for nonpneumonic acute respiratory tract infection, which included diagnoses such as the common cold, sinusitis, otitis media, and bronchitis. A total of 1,257 visits involved the use of an antibiotic. The authors weighted visits to make them nationally representative and adjusted for individual physician clustering. Broad-spectrum antibiotics were classified to include the quinolones, amoxicillin-clavulanate potassium, second- and third-generation cephalosporins, azithromycin, and clarithromycin.
Among patients with acute respiratory tract infections, the most common diagnoses were sinusitis (24 percent of patient visits), the common cold (24 percent), and acute bronchitis (23 percent). Other less common diagnoses included otitis media, pharyngitis, laryngitis, tracheitis, or a combination of these. Antibiotics were prescribed to 63 percent of patients with acute respiratory tract infection. Of these, 54 percent were given a broad-spectrum antibiotic. Azithromycin and clarithromycin accounted for 21 percent of all antibiotic prescriptions.
Multivariate analysis showed that the strongest predictors of broad-spectrum antibiotic prescribing were national region and physician specialty, which were additive; 76 percent of all broad-spectrum antibiotic prescriptions were written by internists in the Northeast and South. Independent predictors of lower rates of broad-spectrum antibiotic choice included black patient race, membership in a health maintenance organization, and lack of patient health insurance for the visit.
In summary, the authors found that internists and physicians in the Northeast and South are particularly likely to prescribe broad-spectrum antibiotics. In many cases, antibiotics are not indicated at all, while in others, the benefit of broad-spectrum agents over narrow-spectrum agents is doubtful at best. The authors conclude that efforts should be made to target cultural factors within physician specialties and geographic regions, particularly through the use of opinion leaders, clinical champions, and academic detailing, all of which have successfully modified physician prescribing behavior.