Adjusting for measurement error to assess health effects of variability inbiomarkers

Longitudinal studies of health effects often relate individuals' biomarker levels to disease progression. Repeated measurements also provide an opport unity to assess within-individual biomarker variability, and it is reasonab le to postulate that this measure might provide additional information abou t a particular outcome variable, Given the existing precedent for applicati on of adjustment methods to account for measurement error in subject-specif ic average levels of a covariate, this concept motivates the application of such methods to incorporate variability as well. In this paper, we investi gate the nature of the relationship between the decline of CD4 cell count i nduced by infection with human immunodeficiency virus, and CD4 level and va riability prior to infection. We first describe the distribution of repeate d CD4 measurements prior to infection using a model that accounts both for random average levels and random subject-specific variance components. Base d on this model, we define true unobservable random variables that correspo nd to prior level and stability. We perform a linear regression analysis, u sing these latent variables as covariates, by means of a full maximum likel ihood approach. We compare the resulting parameter estimates with those bas ed on regressions employing sample-based estimates of pre-infection levels and variances, and empirical Bayes estimates of these quantities. Although the final inferences are similar to those based on the unadjusted analysis, we find that the magnitude of association with prior level decreases, whil e that with prior stability increases. Stratified analyses indicate that sm oking status affects the relationship between prior CD4 level and initial C D4 decline. We point out advantages associated with the maximum likelihood approach in this particular application. Copyright (C) 1999 John Wiley & So ns, Ltd.