Abstract
We have frequently implemented crossover studies to evaluate new therapeutic interventions for genital herpes simplex virus infection. The outcome measured to assess the efficacy of interventions on herpes disease severity is the viral shedding rate, defined as the frequency of detection of HSV on the genital skin and mucosa. We performed a simulation study to ascertain whether our standard model, which we have used previously, was appropriately considering all the necessary features of the shedding data to provide correct inference. We simulated shedding data under our standard, validated assumptions and assessed the ability of 5 different models to reproduce the parameters used and assess model performance. Our standard Poisson model, which includes a random per-person intercept to account for overdispersion, provided surprising findings: excess type I error (up to 28%) and poor coverage (~70%). A Poisson model with data-derived empirical variance structure, however, provided optimal power with type I error controlled at 5% and near 95% coverage. Explanations include that HSV detection frequency is doubly overdispersed: 1) person-level differences in shedding frequency must be accounted for by a random intercept and 2) the episodic nature of shedding induces additional extra-Poisson variance in person-level shedding frequency that must be accounted for using an empirical variance structure. These data support using the model with empirical variance structure for future crossover studies of HSV shedding.
Disciplines
Applied Statistics | Bioinformatics | Biometry | Biostatistics | Categorical Data Analysis | Clinical Epidemiology | Clinical Trials | Computational Biology | Design of Experiments and Sample Surveys | Disease Modeling | Epidemiology | Genetics | Genomics | Health Services Research | Institutional and Historical | Laboratory and Basic Science Research | Longitudinal Data Analysis and Time Series | Medical Specialties | Microarrays | Multivariate Analysis | Probability | Statistical Methodology | Statistical Models | Statistical Theory | Survival Analysis | Vital and Health Statistics
Suggested Citation
Magaret, Amalia, "Models for HSV shedding must account for two levels of overdispersion" (January 2016). UW Biostatistics Working Paper Series. Working Paper 410.
https://biostats.bepress.com/uwbiostat/paper410
Included in
Applied Statistics Commons, Bioinformatics Commons, Biometry Commons, Biostatistics Commons, Categorical Data Analysis Commons, Clinical Epidemiology Commons, Clinical Trials Commons, Computational Biology Commons, Design of Experiments and Sample Surveys Commons, Disease Modeling Commons, Epidemiology Commons, Genetics Commons, Genomics Commons, Health Services Research Commons, Institutional and Historical Commons, Laboratory and Basic Science Research Commons, Longitudinal Data Analysis and Time Series Commons, Medical Specialties Commons, Microarrays Commons, Multivariate Analysis Commons, Probability Commons, Statistical Methodology Commons, Statistical Models Commons, Statistical Theory Commons, Survival Analysis Commons, Vital and Health Statistics Commons