Premium
Bias Adjustment with Polychotomous Logistic Regression in Matched Case‐Control Studies with Two Control Groups
Author(s) -
Becher K.H.,
Jöckel K.H.
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.4710320706
Subject(s) - logistic regression , statistics , representativeness heuristic , control (management) , econometrics , mathematics , population , estimation , group (periodic table) , regression analysis , regression , computer science , economics , medicine , artificial intelligence , environmental health , chemistry , management , organic chemistry
Relative risk estimation in case‐control studies is based on the premise that the control group represents the underlying population. Often more than one control group is collected in order to minimize the possibility of accepting a false result. In this paper it is assumed that a case is matched individually to two different controls and that one control group may lack representativeness with respect to some risk factors. It is discussed whether this group may be used for relative risk estimation and a polychotomous logistic regression model is suggested in which these differences between the control groups are taken into account. A practical method for model search is given. Data of two case‐control studies are used to demonstrate the method. In a simulation study its efficiency is investigated. Some computations illustrate the effect of ignoring a bias in one control group.