Research Findings

Methods Studies - Statistical Analysis

Spatiotemporal and spatial threshold models for relating UV exposures and skin cancer in the central United States (2009)
 Statistical models were explored for estimating the dose-response relationship between ultraviolet radiation exposure measures derived from the NASA Total Ozone Mapping Spectrometer (TOMS) database and non-melanoma skin cancer incidence in a sample of U.S. radiologic technologists. Models took into account time spent outdoors in the sun, skin sensitivity characteristics, sun protection measures, severe sunburn history, and other covariates.
An application of a weighting method to adjust for nonresponse in standardized incidence ratio analysis of cohort studies (2005)
 Cohort studies often conduct periodic surveys to determine disease outcomes and update exposure and other information. Ignoring individuals who do not repond to one or more surveys in analyses can lead to biased results if responders and non-responders differ according to important characteristics associated with outcomes of interest. We describe in this paper a simple weighting method that can be used to adjust for potential non-response bias in studies that have baseline information available on everyone.