Body mass index, smoking, age and cancer mortality among women: a classification tree analysis

Although obesity is an established risk factor for cancer mortality among women, little is known about how body mass index (BMI) is interacting with certain lifestyle behaviors and sociodemographic characteristics to increase the risk of cancer mortality. The purpose of this study was to use classification trees to examine possible interactions between BMI, smoking, age, poverty level and marital status in identifying high-risk subgroups for all-cause cancer mortality in women.

Data from the US National Health Interview Survey linked with the National Death Index from 1990-2004 were used in this study. The data were restricted to women with complete information on all risk factors considered in the analyses. The data were analyzed using descriptive statistics and classification trees.

Findings revealed that younger women (41-50 years old) who smoked had higher cancer mortality if they were underweight. Even among 51-60-year-old never smokers, we observed an inverse relationship between BMI and mortality. High BMI, however, was not a protective factor in 51-60-year-old women who were not married and did not smoke.

Classification tree analysis confirms and extends current knowledge about the role of BMI and smoking on cancer mortality by providing preliminary risk profiles. The findings indicate that it is imperative for cancer-related studies to examine BMI and smoking within age-specific groups and in the context of several sociodemographic factors that are known to independently affect cancer mortality.