Premium
Estimation of Epidemiologic Interaction of Etiologic Factors in Biomedical Studies
Author(s) -
Rao B. Raja
Publication year - 1990
Publication title -
biometrical journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.108
H-Index - 63
eISSN - 1521-4036
pISSN - 0323-3847
DOI - 10.1002/bimj.4710320510
Subject(s) - statistics , logistic regression , odds ratio , mathematics , population , risk factor , econometrics , demography , medicine , environmental health , sociology
In the present paper, a multiple logistic model has been proposed to estimate the exposure‐specific incidence rates of several dichotomous etiologic factors from case‐control data. The relative probability of falling in a given exposure level with respect to the unexposed population for cases is expressed as a logistic function of the exposure levels (1 if exposed, 0 if not) of these factors. The odds ratio of the exposure category of cases relative to the controls comes out to be an exponential function of the exposure levels. It is shown how the maximum likelihood estimates of the logistic parameters can be obtained. The term ‘marginal effect’ of a factor is introduced. This is the effect of a factor A in the presence of one or more of other factors. The difference S A between the marginal effect and the individual effect of a factor A is the increase in the effect of A when it interacts with other factors. A simple relationship has been established between the Hogan et al. measure and the increases S A , S B , … of two or more factors. Several numerical examples illustrate the results.