Body mass index and mortality in an insured population

This study was conducted to explore the relationship between body mass index (BMI) and mortality in an insured population issued policies at standard rates or rated only for build using contemporary analytic techniques.

Many factors influence the relationship between build and mortality. Recent clinical literature on this subject often employs multivariate statistical techniques to better define this relationship and reduce the influence of confounders. BMI, a common surrogate variable for build in clinical literature, is our variable of choice in studying the relationship between build and mortality.

We studied internal data on direct and reinsurance business issued between 1975 and 1998 at standard rates or rated only for build. The policies were followed till termination (death or lapse) or to the end of 1999. The average policy duration was 4.7 years. Cox proportional hazards model runs were used to study the multivariate relationship between mortality and BMI in moderately over- and underweight insured individuals.

During follow-up, 4105 deaths were observed. Mortality was noted to very with BMI, most significantly in middle-aged male nonsmokers. Consistent with reports from the clinical literature, significant factors influencing the BMI-mortality relationship in this insured population included issue age and smoking status.

BMI is a predictor of statistically significant mortality differentials in insured populations. The strength of the BMI-mortality relationship was found to vary by age, gender, and smoking status. In our study population, the male nonsmoker subgroups tended to exhibit the strongest graded relationship between hazard of death and increasing or decreasing BMI.