Statistical techniques applied to environmental health sciences

National Institute Of Environmental Health Sciences

Investigators

Linked publications & trials

Abstract

We have recently contributed to data analysis for several projects, primarily involving electronic health records data from the UNC sleep laboratory (see ES101765, Leping Li, PI). We are also collaborating with a group of scientists in the DTT on analysis of data involving laboratory studies of combined exposures and mixtures (see ES103373, Cynthia Rider, LI) We have used UNC health records from in-laboratory overnight sleep studies that included electroencephalography (EEG) to compare the power spectra between patients with depressive disorder and controls and to assess differences in spindle wave density between patients with type 2 diabetes and controls. We have characterized sex differences in the age trajectories of oxygen saturation and carbon dioxide levels as well as sex differences in the age trajectories of supine positional obstructive sleep apnea, a condition where apnea symptoms are marked when sleeping on one’s back and mild or non-existent when sleeping on one’s side. In addition, we aided the development of a procedure to detect movement and lead-popping artifacts in EEG traces. With investigators in DTT, we are refining methods to predict the dose-response function for a mixture of defined chemical components under various additivity assumptions using dose-response data for each of the component chemicals individually. Further, we are developing methods to compare the predicted curves derived under additivity to dose-response data generated for the mixture itself. We will apply these methods to mixtures of polyaromatic compounds using generated in DTT laboratories.