Achievement of Causal Inference in the Social Medicine in Japan

• Published in: Nippon Eiseigaku Zasshi (Japanese Journal of Hygiene) Vol. 55; no. 2; pp. 462 - 473
• Main Authors: TSUDA, Toshihide, BABAZONO, Akira, MINO, Yoshio, YAMAMOTO, Eiji, MIYAI, Masaya, SHIGEMI, Jun
• Format: Journal Article
• Language: English; Japanese
• Published: Japan The Japanese Society for Hygiene 01.07.2000
• Subjects: black box; causal criteria; causal inference; Causality; Humans; Japan; manifestational criteria; mechanism; Models, Statistical; non-specific disease; Social Medicine; specific disease; Japan
• ISSN: 0021-5082; 1882-6482
• DOI: 10.1265/jjh.55.462

As a condition to achieving an agreement of recognition on the causal relationship in medicine, we firstly explained Hume's problem and counterfactual model. We, however, emphasized that we believe in the existence of causality on medical issues in our daily lives. Therefore, we illustrated conditions when we usually believe in causality. On the other hand, we criticized two well-known key phrases, “lack of mechanism in epidemiology” and “black box in epidemiology”, which have often been used in Japan for skeptic viewpoints against epidemiologic methods even if epidemiology is often used to elucidate a causal effect in medicine in the world. We emphasized that a priori determinations of levels for inference of mechanism is necessary. And, the level and feature of mechanism should be defined in concrete expressions. After explanation of these basic concepts, we mentioned a classic view on specific diseases and non-specific diseases which have not been sufficiently discussed enough yet in Japan. As an example, we used the statements in the Japanese Compensation Law for the Health Effect by Environmental Pollution. In Japan, the classification of these diseases has been confused with that between manifestational criteria of diseases and causal criteria of them. We described the basic concepts to illustrate the causal relationship between non-specific disease and its exposure by using attached figures. Actually, we cannot recognize disease occurrence as a specific disease for several reasons. We indicated that we can recognize the magnitude of effect by causal relationships in medicine as a quantitative continuous variable.