Predictors of antibiotic prescription in respiratory tract infections by ambulatory care practitioners

The use of antibiotics in viral respiratory infections (common cold, acute rhinosinusitis and acute bronchitis) promotes the emergence and spread of resistant bacteria. Studies have found that antibiotics are prescribed for 50-70% of respiratory tract infections, despite the fact that most of them have a viral etiology. The objective of the study was to determine predictors of antibiotic use in acute respiratory infections. It was conducted as a cross-sectional study on physicians' practices for antibiotic use. The subjects were internists and otolaryngologists of adult patients in an ambulatory setting in Buenos Aires. The instrument was a questionnaire with 20 clinical vignettes that included relevant variables for making decisions regarding antibiotic use in acute respiratory infections. The vignettes were constructed with a fractional factorial design with nine clinical variables. The absolute and relative weight of each clinical variable that predicted antibiotic use were calculated for each individual practitioner using multiple linear regression. The predictors with the greatest absolute weight in the decision to prescribe antibiotics were nasal discharge and cough (24% and 21% of total weight). The correlation between predictors and individual physician answers was high (r(2) = 0.73). The mean probability and the rate of antibiotic prescription were both about 50%. Predictors of antibiotic use for acute respiratory infections among ambulatory physicians in this sample differ from internationally accepted guidelines. The likelihood of prescribing antibiotics for these illnesses is high. Wider implementation of management guidelines for acute respiratory infections could improve cost effective antibiotic use and decrease the development of antibiotic resistance.