[Principles and methods of influenza epidemiology: with special reference to field evaluation of vaccine efficacy]

In Western countries, prevention of influenza has been a major public health concern, promoting the vaccination program to high-risk individuals including the elderly. On the other hand, in Japan, there has been no systematic approach to such a selective vaccination on the social basis. This is due to the deep-rooted skepticism on vaccine efficacy. A number of epidemiologic studies have so far reported conflicting results in this country. We investigated the principles and methods of influenza epidemiology focusing on the field evaluation of vaccine efficacy. General and methodologic problems in vaccine field trials with naturally occurring influenza include; unpredictability of the time of its occurrence; antigenic differences between the vaccine strains and epidemic viruses; preexistence of already-immune individuals; indirect effect of hard immunity by vaccination on nonvaccinees; possible difference in the virus exposure between compared groups, particularly when epidemic scale is small; possible misdiagnosis of cases, if not laboratory-confirmed. To cope with these, the following measures are essential in conducting epidemiologic study on influenza. 1) Influenza epidemics show differential occurrence by time and place. Therefore, much attention should be pain when analyzing the pooled data obtained from various study samples at different locations. 2) It should be the first step of a research to consider whether the outbreak of acute respiratory illnesses observed among subjects is caused by influenza virus exposures. 3) In general, failings to detect vaccine efficacy are attributable to the dilution of outcome with noninfluenzal illnesses; cases defined by clinical symptoms include substantial number of acute respiratory illnesses other than influenza. To minimize this nondifferential misclassification, three methods are thought to be important; confining observation period during the peak epidemic; applying a strict criteria to measure the outcome, preferably with laboratory examination; and conducting the study in the season with large-scale epidemic. 4) It is also important to take into account the preexistence of already-immune individuals. "Antibody efficacy" is a keen index to assess vaccine effectiveness in this instance. Besides, further research is required to clarify the individual characteristics related to influenza attack, which affect the validity of analytic epidemiologic study on vaccine efficacy by yielding bias or confounding effect.