This article is only available in the PDF format. Download the PDF to view the article, as well as its associated figures and tables.
We agree with Dr Black-welder that the possibility of a type II error is large when comparative trials of antimicrobial agents are performed in fewer than 100 patients. For this reason we were cautious with our conclusions stating that "we do not recommend the shorter course of antibiotics for all patients with meningitis." Specific situations were presented that we believed required conventional antimicrobial schedules for treatment of meningitis. Although correct from a statistical analysis standpoint, Dr Blackwelder's point is impractical because it would take five years or longer to enroll approximately 500 patients into a comparative therapeutic trial. Even then, we might not satisfy the statistician's concerns about legitimacy of conclusions and recommendations.
McCracken GH, Nelson JD, Chrane DF. Comparing Antibiotics: Statistical Rigor vs Practicality-Reply. JAMA. 1985;254(23):3311-3312. doi:10.1001/jama.1985.03360230041015