Objective: To quantify changes in clinical and cost outcomes associated with increasing levels of body mass index (BMI) in a US setting.
Research methods and procedures: A semi-Markov model was developed to project and compare life expectancy (LE), quality-adjusted life expectancy (QALE) and direct medical costs associated with distinct levels of BMI in simulated adult cohorts over a lifetime horizon. Cohort definitions included age (20-65 years), gender, race, and BMI (24-45 kg m(-2)). Cohorts were exclusively male or female and either Caucasian or African-American. Mortality rates were adjusted according to these factors using published data. BMI progression over time was modeled. BMI-dependent US direct medical costs were derived from published sources and inflated to year 2004 values. A third party reimbursement perspective was taken. QALE and costs were discounted at 3% per annum.
Results: In young Caucasian cohorts LE decreased as BMI increased. However, in older Caucasian cohorts the BMI associated with greatest longevity was higher than 25 kg m(-2). A similar pattern was observed in young adult African-American cohorts. A survival paradox was projected in older African-American cohorts, with some BMI levels in the obese category associated with greatest longevity. QALE in all four race/gender cohorts followed similar patterns to LE. Sensitivity analyses demonstrated that simulating BMI progression over time had an important impact on results. Direct costs in all four cohorts increased with BMI, with a few exceptions.
Conclusions: Optimal BMI, in terms of longevity, varied between race/gender cohorts and within these cohorts, according to age, contributing to the debate over what BMI level or distribution should be considered ideal in terms of mortality risk. Simulating BMI progression over time had a substantial impact on health outcomes and should be modeled in future health economic analyses of overweight and obesity.