In a matched case control study the matched control is intended to reduce bias associated with a confounding variable. If the study is overmatched then the estimate of risk or its statistical significance may be biased.
Matching is part of the study design. It assumes that the confounding variable:
(1) is independent of the risk factor being investigated.
(2) is associated with the outcome
In over-matching the confounding variable:
(1) may not be independent of the risk factor
(2) may not be associated with the outcome (irrelevant)
Any matched case control study should be evaluated for the possibility of overmatching. Overmatching should be looked for if a matched case control study:shows findings discrepant with other well-done studies or if bias is suspected. The problem with the confounding variable may not be immediately obvious.
Overmatching reduces the reported statistical significance if a matched analysis is used. If a matched analysis shows statistical significance then overmatching may not be a problem.
To read more or access our algorithms and calculators, please log in or register.
