Table 4.

Results from discrete time survival analyses in five models of increasing complexity

CharacteristicModel 1Model 2Model 3Model 4Model 5
OR (with 95% CI) for selected covariates as potential modifiers of the estimated annual hazard (incidence) of AD. Analysis based on 3,281 study members, 122 of which were incident cases of AD. In all, the study members with complete AD diagnoses and covariate information contributed 10,458 person-years to the analysis. Models were centered at mean values of 76 years of age and 13 years of education. Improvement in fit of each model, compared with its predecessor, is indicated in the bottom row by the likelihood ratio (LR) χ2 statistic with degrees of freedom (df) equal to the difference in the number of parameters in the models compared.
Age, y1.1681.1791.2511.3101.264
Education, y1.0571.0461.0481.0501.046
One ε41.9771.9232.1682.165
Two ε48.5538.55215.75614.894
Age-by-one ε40.9870.989
Age-by-two ε40.8370.841
LR χ233.427.9210.553.69
(df 2, p < 0.001)(df 1, p < 0.01)(df 2, p < 0.01)(df 1, p = 0.055)