Abstract
The exact mechanisms relating exposure to ultraviolet (UV) radiation and elevated risk of skin cancer remain the subject of debate. For example, there is disagreement on whether the main risk factor is duration of the exposure, its intensity, or some combination of both. There is also uncertainty regarding the form of the dose-response curve, with many authors believing only exposures exceeding a given (but unknown) threshold are important. This paper explores methods to estimate such thresholds using hierarchical spatial logistic models based on a sample of a cohort of x-ray technologists for whom self-reports of time spent in the sun and numbers of blistering sunburns in childhood are available. A preliminary goal is to explore the temporal pattern of UV exposure and its gradient. Changes would imply that identical exposure self-reports from different calendar years may correspond to differing cancer risks.
Original language | English (US) |
---|---|
Pages (from-to) | 3001-3015 |
Number of pages | 15 |
Journal | Computational Statistics and Data Analysis |
Volume | 53 |
Issue number | 8 |
DOIs | |
State | Published - Jun 15 2009 |
Bibliographical note
Funding Information:The work of the first and last authors was supported in part by NIH grant 1-R01-CA95955-01, while that of the second and third authors was supported by a grant from the University of Minnesota Cancer Center. The USRT Cohort Study is supported by the intramural research program of the National Cancer Institute, Contract N01-CP-31018. The authors are grateful to Ms. Sang Mee Lee for crucial initial programming and data analytic assistance, and to Dr. Sudipto Banerjee for helpful discussions.