〈総説〉疫学の影響の指標としてのオッズ比

概要

論文の詳細を見る
The epidemiologic theories on odds ratio as an effect measure are reviewed to be applied for its autual studies. First, under consideration in the "fixed cohort model", odds ratio is approximated to be relative risk in a traditional case-control study. Next, when the "dynamic population model" is assumed, odds ratio is considered to be an estimate of incidence rate ratio in both the cases where controls are directly sampled from base in one type or indirectly. Odds ratio is also applied not only for case-control studies, but also for other study designs. Cause specific death data, which were used in the PMR (Proportionate Mortaliyt Ratio) analysis, can also be analyzed using odds ratio for estimating incidence rate ratio. In a cross-sectional study, odds ratio is also used for estimating incidence rate ratio under particular conditions. By summarizing these applications, odds ratio is clarified to be a quite useful indicator to assess the magnitude of effect and a good estimator of incidence rate ratio. Finally. we classify several epidemiologic designs into three types according to two sampling sources ; base and outcomes, where the concepts of primary base and secondary base are also described.
九州大学の論文