Choosing among generalized linear models applied to medical data

Abstract

SUMMARYWhen testing for a treatment effect or a difference among groups, the distributional assumptions made about the response variable can have a critical impact on the conclusions drawn. For example, controversy has arisen over transformations of the response (Keene). An alternative approach is to use some member of the family of generalized linear models. However, this raises the issue of selecting the appropriate member, a problem of testing non-nested hypotheses. Standard model selection criteria, such as the Akaike information criterion (AIC), can be used to resolve problems. These procedures for comparing generalized linear models are applied to checking for difference in T cell counts between two disease groups. We conclude that appropriate model selection criteria should be specified in the protocol for any study, including clinical trials, in order that optimal inferences can be drawn about treatment differences.

Why eyewitnesses fail

“…Promising new approaches in this area use generalized linear models based on manifold variables (e.g., viewing conditions, perceptual/cognitive states) that have known effects on visual uncertainty. Such models have long been used in basic research on sensory and memory function and they may be used to estimate the likelihood of a correct identification given a specific set of conditions, in a manner analogous to their use in medical practice (48,49).…”

Major Genetic Susceptibility for Venous Thromboembolism in Men: A Study of Danish Twins

“…The model with the lowest AIC represents the best balance of goodness of fit and parsimony. 23 For comparison among nested models, we used the chi-square difference test. The difference in chi-square of the models is itself distributed as a chi-square statistic, with the degrees of freedom equal to the difference in the degrees of freedom of the models being compared.…”

Heterogeneity in vaccination coverage explains the size and occurrence of measles epidemics in German surveillance data

“…The approach of Held et al [10] is extended to allow for the inclusion of covariates and applied to the measles data using vaccination coverage as an explanatory variable. Different formulations of the proposed model are compared based on Akaike's Information Criterion (AIC [13]). A simulation study is performed in order to further investigate the ability of AIC to identify the underlying true model.…”