摘要：OBJECTIVES. Death rates for community-acquired pneumonia based on relatively small-scale, published studies tend to exceed 15% to 20%. This study reexamined these estimates by using very large, population-based databases. METHODS. Death rates from 1993 associated with community-acquired pneumonia were reexamined with hospital discharge data from all of Washington, Illinois, and Florida. RESULTS. These death rates were substantially lower (7.0%, 8.1%, and 9.7%, respectively) than what appears in the literature. Significant risk factors for dying were being 65 years of age or older (odds ratio [OR] = 2.9), being positive for human immunodeficiency virus (OR = 2.9), and having a high severity of illness (OR = 7.1). CONCLUSION. Sampling bias associated with selection for hospital admissions explain the discrepancy between previous and this study's results.
