Bayesian Age-Period-Cohort Model of Lung Cancer Mortality

• Activity Number: 136
• Type: Contributed
• Date/Time: Monday, August 10, 2015 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Bhikhari P. Tharu* and Ram C. Kafle and Chris P. Tsokos
• Companies: University of South Florida and Sam Houston State University and University of South Florida
• Keywords: Histogram smoothing ; logit parameters ; multivariate

Abstract

In this study, we aim to develop an age- period- cohort model to study the mortality trends of a particular disease in a population. We have considered that the observed number of deaths follows Poisson probability distribution. We model the rate of deaths as a function of age, period, and cohort. We propose the histogram smoothing technique to smooth the data. We apply the model to study the mortality due to lung cancer (LC) patients for the periods 1970-2010 in USA. The data set is obtained from the Surveillance Epidemiology and End Results (SEER) of National Cancer Institute (NCI). We noticed that mortality rate due to LC is decreasing overall periods and fairly stable for age group 37 and lower.