Strength of Association. Report 1. Graduations of Relative Risk

Purpose: To summarize data on graduation of the effect size on the base of Hill’s first causality criterion ‘Strength of association’ on relative risk parameters (RR). Material and methods: Survey of published sources: monographs, handbooks, papers, educational material on statistics in various disciplines (including on-line), etc. (128 references; of which about 30 handbooks on epidemiology, carcinogenesis and medical statistics). Results: For the RR value, the collected data summary (1980–2018) implies non-homogeneity in concepts. The most common references are to the Monson scale (two editions of the monograph on the epidemiology of occupational exposures Monson R.R., 1980; 1990). In our opinion, the optimal graduation can be developed on the basis of this scale, and it should include both the range of no effect (RR = 0.9–1.2) and the weak (RR = 1.2–1.5, or 0.7–0.9), moderate (RR = 1.5–3.0, or 0.4–0.7), strong (RR = 3.0–10.0; or 0.1–0.4) and very strong’(RR = 10.0–40.0; or 0.0–0.1) ranges. Examples of epidemiological effects with overwhelming strength of association are presented (RR > 40.0). For the effects of thalidomide, RR reached thousands, for diethylstilbestrol, conditional infinity, and when irradiated in childhood, the frequency of some cancers increased tens and even hundreds of times. The juristic aspects of compensation payment based on RR are briefly reviewed. According to the Daubert rule (Daubert ruling, Daubert standard) on the 1993 precedent in the United States, risks are recognized only at RR > 2.0, when the probability of causality is more than 50 %. Conclusions: To estimate the RR value, one should use the most common and officially established Monson scale, albeit with an expansion in the range of dramatic or overhelming risks. This study can be used as a reference guide on the graduations of effect size on RR (OR) for a wide variety of observed disciplines.