The Role of Body Mass Index at Diagnosis on Black-White Disparities in Colorectal Cancer Survival: a Density Regression Mediation Approach

• Activity Number: 402 - HPSS Student Paper Competition Winners: Statistics Advancing Policy
• Type: Topic Contributed
• Date/Time: Tuesday, July 31, 2018 : 2:00 PM to 3:50 PM
• Sponsor: Health Policy Statistics Section
• Author(s): Katrina Devick* and Linda Valeri and Jarvis Chen and Alejandro Jara and Marie-Abele Bind and Brent A. Coull
• Companies: Harvard TH Chan School of Public Health and McLean Hospital, Harvard Medical School and Harvard TH Chan School of Public Health and Pontificia Universidad Catolica de Chile and Harvard University and Harvard TH Chan School of Public Health
• Keywords: causal inference; cancer health disparities; nonparametric Bayesian ; accelerated failure time model

Abstract

The study of racial/ethnic inequalities in health is important to reduce the uneven burden of disease. In the case of colorectal cancer (CRC), disparities in survival among non-Hispanic Whites and non-Hispanic Blacks are well documented, and mechanisms leading to these disparities need to be studied formally. Body mass index (BMI) is a well-established risk factor for developing CRC, and recent literature shows BMI at diagnosis of CRC is associated with survival. Since BMI varies by racial/ethnic group, a question that arises is whether disparities in BMI is partially responsible for observed racial/ethnic disparities in CRC survival. This paper presents new methods to quantify the impact of a hypothetical intervention on BMI on racial/ethnic disparities in survival. We do this by fitting a Bayesian density regression model for BMI, coupled with an accelerated failure time model for survival. We perform a simulation that shows our proposed approach performs as well as or better than current methodology allowing for a shift in means only, and that standard practice of categorizing BMI leads to large biases when interactions are present and the distribution of BMI is not normal.