Basic principle of population-based cohort study to evaluate influenza vaccine effectiveness among elderly Japanese.

Influenza vaccines minimize the risk of influenza-related morbidity, complication, and death in elderly people. Although evaluating vaccine effectiveness (VE) is important for promoting immunization programs and coping with influenza epidemics, it is difficult to evaluate its effectiveness in Japan, where no frameworks to use large databases, such as a vaccination registry and health maintenance organization datasets, are available. Therefore, another analytic epidemiological investigations to evaluate VE in Japan are required. Herein, we describe the basic principles of a cohort study, which might be the most comprehensive, but expensive, study design. It is particularly important to be aware of the potential bias and confounding factors that should be minimized in the study design and analysis. We focus on "laboratory-confirmed influenza" and "influenza-like illness", and discuss why it is important to follow up with equal intensity, and how to control for bias; problems that often arise in population-based observational cohort studies.