In Reply: We agree in principle with Dr Kaczorowski that P values should not be the sole driving force in determining true associations. However, his concern about inadequate power may not be that major. Although only 22.3% of the sample in our study were 65 years or older, this comprised more than 18 000 individuals at baseline. Because the precision of our estimates depends in large part on the absolute sample size, the CIs around our estimates of effect were fairly narrow, indicating reasonably precise estimates even after adjusting for potential confounders.1 Moreover, the estimates suggest that these data are compatible with at most a weak association. Given the potential residual or uncontrolled confounding, even these modest associations are suspect.
Tseng HF, Jacobsen SJ. Pneumococcal Vaccination and Cardiovascular Events in Men—Reply. JAMA. 2010;304(7):742-743. doi:10.1001/jama.2010.1146