A BIOMEDICAL APPLICATION OF LATENT CLASS MODELS WITH RANDOM EFFECTS

Traditional latent class modelling has been used in many biomedical se ttings. Unfortunately, many of these applications assume that the diag nostic tests are independent given the true disease status, an assumpt ion that is often violated in practice. Qu, Tan and Kutner developed g eneral latent class models with random effects to model the conditiona l dependence among multiple diagnostic tests. In this paper latent cla ss modelling with random effects is used to estimate the sensitivity a nd specificity of six screening tests for detecting Chlamydia trachoma tis in endo-cervical specimens from women attending family planning cl inics.