Sir.—In spite of the study design limitations (possible biases due to a higher exposure rate for the control children and biases due to the lack of double blinding), the findings in the recent article by Heikkinen et al1 suggest that influenza vaccination may be effective in the prevention of acute otitis media in children. The results, however, contain a statistical error that affects the conclusions.
The study correctly reports a significantly higher proportion of cases of influenza A and acute otitis media (AOM) among control children compared with vaccinated children. The authors report diagnoses of influenza A in five (2.67%) of 187 vaccinated and 29 (15.51%) of 187 control children. As the authors report, this difference is statistically significant (χ2=18.63; P=.00001),2 and represents an 83% protective efficacy against influenza A: protective efficacy=1 − (failure rate of the vaccine/failure rate of no treatment)=1−(2.67/15.51)=0.83.3 In