Covariates dichotomous

Let us look at the typical case in more detail. Suppose we have k dichotomous indicators. The basic strategy of SSMAXCOV is to take a pair (any pair) of indicators as output variables and combine the remaining (k - 2) indicators in one scale. In other words, the input variable is the sum of scores on all indicators, with the exception of the two that were taken as output variables. Then, covariances of the output indicators are computed and plotted for each... 

Since the model is user supplied, the pred block is used and only the NONMEM-required variables id (subject identifier), dv (the dependent variable or dichotomous endpoint), mdv (the missing dependent variable data item), and the explanatory variables of interest (the predictors, or independent variables) are required. An example NM-TRAN control stream for the base model (described below) estimating the probability of experiencing rash, prior to the inclusion of any potential covariate (predictor) effects, is shown in Appendix 24.1. In this case, the base model to be specified in NONMEM is described in the Eq. (24.1) and (24.2) ... 

Covariate interactions, if biologically plausible, can also be estimated and tested for signihcance as they would with other linear or mixed effects models. To determine whether an interaction between covariates might be present, the empirical logit plots described in Section 24.3.1 can be examined. To assess the possibility for continuous-dichotomous predictor interactions, the empirical logit can be plotted versus each continuous predictor with separate symbols and lines for each level... 

When the covariate is categorical and dichotomous, three possible models are typically used additive, fractional change, and exponential. To illustrate these models, it will be assumed that clearance is dependent on age in a linear manner and that clearance is also dependent on the sex of the subject. In an additive model, the effect of the dichotomous covariate is modeled as... 

When the covariate is of two or more categories, pharmacokineticists often first try to convert the variable to a dichotomous variable, such as converting race into Caucasian versus others, and then use standard dichotomous modeling techniques to build the model. This practice is not entirely appropriate because categorization results in loss of information. 

Sometimes a continuous variable is categorized and then the categorical variable is included in the model. For example, if age were the covariate and there is a question about the clearance of a drug in the elderly, then age could be transformed to a dichotomous covariate (Age group) where subjects younger than 65 years take on a value of 0 and those older than or equal to 65 years a value of L Clearance could then be modeled as... 

Sometimes, rather than treating the covariate as a continuous variable, the covariate will be categorized and the categorical variable will be used in the model instead of the original value. For example, de Maat et al. (2002) in a population analysis of nevirapine categorized baseline laboratory markers of hepatic function into dichotomous covariates. Patients were coded as V if their laboratory value was 1.5 times higher than the upper limit of normal and 0 otherwise. Patients with an aspartate aminotransferase 1.5 times higher than normal had a 13% decrease in nevirapine clearance than patients with normal values. 

Koch GG, Tangen CM, Jung JW, Amara lA (1998) Issues for covariance analysis of dichotomous and ordered categorical data from randomized clinical trials and non-parametric strategies for addressing them. Statistics in Medicine 17 1863-1892. 

The present study now considers whether gestational age (or the dichotomous variable, preterm delivery) is related to lead level. This, too, is an important outcome measure and is a covariate in some studies of later... 

---