Online Program Home
My Program

Abstract Details

Activity Number: 82 - Statistical Methods for Disease Prevention and Prediction
Type: Contributed
Date/Time: Sunday, July 28, 2019 : 4:00 PM to 5:50 PM
Sponsor: Section on Statistics in Epidemiology
Abstract #301671
Title: Determinants of Inter-Individual Variation in Nevus Counts Among Children
Author(s): Jaya M Satagopan* and Ariel Chernofsky and Qin Zhou and Stephen W Dusza and Allan Halpern and Irene Orlow
Companies: Memorial Sloan Kettering Cancer Center and Boston University and Memorial Sloan Kettering Cancer Center and Memorial Sloan Kettering Cancer Center and Memorial Sloan Kettering Cancer Center and Memorial Sloan Kettering Cancer Center
Keywords: coefficient of variation; Gini index; Bayesian shrinkage estimation; proportion of explained variation
Abstract:

We examined determinants of variation in childhood nevus counts using data from two age cohorts in the study of nevi in children (SONIC). We measured inter-individual variation using the squared coefficient of variation and the Gini index and visualized the former by plotting the cumulative probability distributions of nevi under two sampling schemes. This plot provides vital information about the predictive performance of risk factors. A monotonic transformation of this plot gives the Lorenz curve and, hence, the Gini index. Thus, the two statistics are equivalent. In SONIC, the observed squared coefficients of variation in nevi were 1.5 (Gini = 0.52) and 0.74 (Gini = 0.45) for children aged 10 and 13 years, respectively. The estimated values using genetic and demographic factors were 0.5 (Gini = 0.38) and 0.39 (Gini =0.34) for children aged 10 and 13 years, respectively. Adding sun exposure increased these estimates by only a small amount. Thus, nevus development in children is under strong genetic control. Plots showed that additional risk factors for nevi remain to be identified to better predict nevi, especially in younger children.


Authors who are presenting talks have a * after their name.

Back to the full JSM 2019 program