Control selection and confounding factors: A lesson from a Japanese case-control study to examine acellular pertussis vaccine effectiveness

Abstract When using a case-control study design to examine vaccine effectiveness, both the selection of control subjects and the consideration of potential confounders must be the important issues to ensure accurate results. In this report, we described our experience from a case-control study conducted to evaluate the effectiveness of acellular pertussis vaccine combined with diphtheria-tetanus toxoids (DTaP vaccine). Newly diagnosed pertussis cases and age- and sex-matched friend-controls were enrolled, and the history of DTaP vaccination was compared between groups. Logistic regression models were used to calculate odds ratios (ORs) and 95% confidence intervals (CIs) of vaccination for development of pertussis. After adjustment for potential confounders, four doses of DTaP vaccination showed a lower OR for pediatrician-diagnosed pertussis (OR = 0.11, 95% CI, 0.01–0.99). In addition, the decreasing OR of four doses vaccination was more pronounced for laboratory-confirmed pertussis (OR = 0.07, 95%CI, 0.01–0.82). Besides, positive association with pertussis was observed in subjects with a history of steroid treatment (OR = 5.67) and those with a recent contact with a lasting cough (OR = 4.12). When using a case-control study to evaluate the effectiveness of vaccines, particularly those for uncommon infectious diseases such as pertussis, the use of friend-controls may be optimal due to the fact that they shared a similar experience for exposure to the pathogen as the cases. In addition, to assess vaccine effectiveness as accurately as possible, the effects of confounding should be adequately controlled with a matching or analysis technique.