Methodologic issues in the evaluation of vaccine effectiveness: measles vaccine at 12 vs. 15 months

Abstract

Initial studies of vaccine effectiveness are usually controlled clinical trials. Subsequent serologic and epidemiologic investigations cannot be as carefully controlled, and methodologic rigor must be considered when evaluating their results. The authors reviewed the methodologic features of greatest importance in the design of field investigations of vaccine efficacy and examined how different methodologic approaches may have influenced the results of studies of measles vaccine efficacy at 12 vs. 15 months of age. Among the serologic studies, time since vaccination was not controlled, so that children vaccinated several years ago at 12 months of age may have been compared with children vaccinated much more recently at later ages. Most studies included children vaccinated at several years of age as controls rather than children vaccinated only at the pivotal ages of 13-15 months. Finally, studies using a single clinic as their vaccine source found smaller differences in efficacy than studies using multiple vaccine sources, suggesting that variation in handling and storage may affect the results. Among the epidemiologic studies, no single methodologic feature clearly influenced whether there was significantly improved vaccine efficacy with increasing age at vaccination. However, studies differed with respect to intensity of effort to detect cases unreported through the usual channels; definition of what constituted a case of measles; whether the effect of midoutbreak clinics on overall estimates of vaccine efficacy was considered; and inclusion of revaccinated children in the analysis. These methodologic variations can make it extremely difficult to evaluate the size and clinical importance of relatively small differences in vaccine efficacy. Greater attention to the methodologic features discussed here will be needed in future studies of vaccine efficacy.