Background: Community-acquired pneumonia (CAP) is the most common infectious diagnosis necessitating adult hospitalization in the United States (US). Timely diagnosis of CAP is important to improve patient outcomes. However, overdiagnosis of CAP, or treatment of CAP despite inadequate signs or symptoms of CAP, may also pose a significant threat to patient safety. Potential harm of overdiagnosis includes the downstream implications of antibiotic overuse (e.g., antimicrobial resistance and antibiotic associated side effects) as well as missed or delayed treatment of alternative diagnoses. Despite these risks, little is known about the prevalence, risk factors, and outcomes related to overdiagnosis of CAP.
Methods: We performed a retrospective cohort study of hospitalized patients treated for pneumonia at 48 hospitals in Michigan between July 2018 and March 2020. Treatment intended for CAP was defined as any antibiotic use on day 1 or 2 of hospitalization in a patient without an alternative infection who had a discharge diagnosis of pneumonia. Patient data (e.g., signs, symptoms, laboratory/radiographic data, antibiotic treatment, 30-day post-discharge outcomes) were collected by trained abstractors from the medical record and 30-day post-discharge telephone calls. Patients treated for pneumonia who did not meet US-guideline recommended diagnostic criteria for CAP were considered overdiagnosed. The proportion of patients overdiagnosed with CAP and variation by hospital were characterized using descriptive statistics. Logit generalized estimating equations with stepwise selection were used to identify risk factors for overdiagnosis of CAP. Finally, generalized estimating equation models adjusted for patient characteristics were used to assess patient outcomes associated with each day of unnecessary antibiotic use.
Results: A total of 18,450 patients received antibiotics for CAP, of which 12.4% (n=2,286) did not meet diagnostic criteria and thus were considered overdiagnosed. There was wide variation in the proportion of patients overdiagnosed with CAP by hospital with over two-thirds of hospitals overdiagnosing >10% of patients treated for CAP. On multivariable analysis, predictors of overdiagnosis included dementia, acutely altered mental status without dementia, Hispanic ethnicity, Black race, Charlson comorbidity index >1, age >65, concurrent COPD exacerbation, and presence of diabetes (Figure 1). After adjustment, each unnecessary day of antibiotics in patients overdiagnosed with CAP increased the odds of 30-day post-discharge patient reported adverse events (aOR 1.05, 95% CI 1.01-1.08, p=0.01). There were no differences in other outcomes (Table 1).
Conclusions: Overdiagnosis of CAP is common, varies widely across hospitals, and is associated with increased odds of antibiotic-associated adverse events. Patients at the highest risk for overdiagnosis include those with dementia or altered mental status, Hispanic ethnicity, Black race, and higher Charlson comorbidity index. Given the prevalence of hospitalizations for CAP, improvements in diagnosis and treatment of CAP are imperative to improve antibiotic stewardship and patient outcomes.