Visit complexity, diagnostic uncertainty, and antibiotic prescribing for acute cough in primary care: a retrospective study

Background: Guidelines and performance measures recommend avoiding antibiotics for acute cough/acute bronchitis and presume visits are straightforward with simple diagnostic decision-making. We evaluated clinician-assigned diagnoses, diagnostic uncertainty, and antibiotic prescribing for acute cough visits in primary care. Methods: We conducted a retrospective analysis of acute cough visits - cough lasting <= 21 days in adults 18-64 years old without chronic lung disease - in a primary care practice from March 2011 through June 2012. Results: Of 56,301 visits, 962 (2%) were for acute cough. Clinicians diagnosed patients with 1, 2, or >= 3 cough-related diagnoses in 54%, 35%, and 11% of visits, respectively. The most common principal diagnoses were upper respiratory infection (46%), sinusitis (10%), acute bronchitis (9%), and pneumonia (8%). Clinicians prescribed antibiotics in 22% of all visits: 65% of visits with antibiotic-appropriate diagnoses and 4% of visits with non-antibiotic-appropriate diagnoses. Clinicians expressed diagnostic uncertainty in 16% of all visits: 43% of visits with antibiotic-appropriate diagnoses and 5% of visits with non-antibiotic-appropriate diagnoses. Clinicians expressed uncertainty more often when prescribing antibiotics than when not prescribing antibiotics (30% vs. 12%; p < 0.001). As the number of visit diagnoses increased from 1 to 2 to >= 3, clinicians were more likely to express diagnostic uncertainty (5%, 25%, 40%, respectively; p < 0.001) and prescribe antibiotics (16%, 25%, 41%, respectively; p < 0.001). Conclusions: Acute cough may be more complex and have more diagnostic uncertainty than guidelines and performance measures presume. Efforts to reduce antibiotic prescribing for acute cough should address diagnostic complexity and uncertainty that clinicians face.