Methodological considerations in case-control studies to evaluate BCG vaccine effectiveness

Several case-control studies evaluating the effectiveness of BCG vaccine in the last decade have presented contradictory results like previous prospective studies. Methodological differences could explain some of the case-control study results. This study explores the possibility that contradictory results could be imputed to the choice of different series of controls. Three controls were compared for each case of tuberculous meningitis: neighbourhood, hospital and household. BCG effectiveness estimates were 86.8%, 92.0% and 29.5%, respectively. The data indicated an interaction between BCG vaccine status and tuberculous focus. This could have influenced the lower effectiveness estimates found when cases were compared with household controls. The paper discusses aspects related to case-control studies applied to evaluate BCG effectiveness such as: incubation period and sufficient time since vaccination to allow development of an immune response; the presence of a tuberculous focus among the groups of cases and controls and the interaction between focus and BCG vaccination; recall bias; and optimum selection of controls in case-control studies in the context of infectious diseases.