In Reply Dr Paul and colleagues question our decision to exclude observational studies that did not perform statistical adjustment for illness severity or comorbidity when assessing the association between time to initiate antibiotic therapy and mortality. All studies meeting eligibility for this question were observational in design and had a high likelihood of confounding by indication (ie, patients who were more severely ill would be more likely to have antibiotic therapy initiated rapidly). Therefore, we considered it imperative to focus the review on studies with adjustment for such confounding.
Lee JS, Giesler DL, Fine MJ. Management of Community-Acquired Pneumonia—Reply. JAMA. 2016;316(2):222-223. doi:10.1001/jama.2016.5028