Advanced Search

Last Name First Name Company Name City Advanced Search

Actuary of the Future Education & Research Entrepreneurial Actuaries Financial Reporting Forecasting and Futurism Health International Investment Joint Risk Management Long-Term Care Insurance Management & Personal Marketing & Distribution Pension Product Development Reinsurance Smaller Insurance Company Social Insurance & Public Finance Taxation Technology

Research

Research Studies in Pension

Factor Affecting Retirement Mortality (FARM)

This FARM site consists of

• an Introduction
• an Abstract
• a Bibliography of research papers
• a collection of Summaries of the research papers.

Use the Table of Summaries to link to the summaries either by author or by risk factor.

Abstract | Bibliography | Introduction | Table of Summaries

Deaths Attributable to Obesity in the United States

Allison, David B, Kevin R. Fontaine, JoAnn E. Manson, June Stevens, and Theodore B. VanItallie, 1999. Journal of the American Medical Association (JAMA), October 27, 1999, 282(16). †

This study aimed to estimate the number of deaths, annually, attributable to obesity among US adults. Data was obtained from 5 prospective cohort studies and one published study. The use of several studies enabled the evaluation of the extent to which the results obtained were sensitive to a particular data set. Analyses using data from only non–smokers or never–smokers was used as a sensitivity analysis as it has been suggested that smoking status may confound the BMI–mortality �† relationship and that analyses of never–smokers yield the most valid estimates. The study measured the relative hazard ratio of death for obese or overweight persons, using a BMI between 25–30 to classify being overweight and a BMI greater than 30 as the classification of obese. The reference group for normal weight had a BMI range of 23 to 25.

The six data sets used were the Alameda Community Health Study, the Framingham Heart Study, the Tecumseh Community Health Study, the American Cancer Society Cancer Prevention Study I, the National Health and Nutrition Examination Survey I (Epidemiologic Follow–up Study) and the Nurses Health Study. These data sources were chosen as they were US specific, not derived predominately from ill, high–risk or elderly subjects and they had well–documented characteristics. Together they provided a relatively broad cross–section of studies.

The estimated number of annual deaths attributable to obesity varied with the cohort used to calculate the hazard ratios, but the findings were consistent overall. More than 80% of the estimated obesity–attributable deaths occurred among individuals with a BMI of more than 30. It was found that about 22% of adults had BMIs at this level. Hazard rates were also calculated from data for non– smokers or never smokers only, where data was available. The estimated number of annual deaths attributable to obesity was higher for non–smokers or never smokers compared with the entire population. The study found that hazard ratios generally increased with BMI, although this was not consistent until the BMI was in the upper 20s. In the higher BMI categories hazard rates tended to be higher in non–smokers or never–smokers, however, over all categories the difference was slight and inconsistent. Possible effects of differing follow–up duration were considered, although no clear association was found.

The analysis indirectly took into account differential effects of obesity by age and sex, simply by including both sexes and cross section of ages in the derivation samples. However, the study was interested in estimating the average effect of overweight or obesity across both sexes over all adult ages. Distributions of age, sex, smoking, health status, ethnicity and socio–economic status in a cohort can affect the estimated hazard ratio. A paper by Stevens et al is referred to, which found that hazard ratios for obesity decreased steadily with advancing age. It may be interesting for a future study to examine how obesity effects differ by age, sex, race and other factors. Most of the samples in this study over– represent whites and may have over–represented middle–and upper–socioeconomic status subjects. Future research may also consider cause–specific mortality.

The study only analyzed BMI at a single point in time. It may be interesting for future investigations to evaluate mortality as a function of BMI changes. The approaches used did not account for potential confounding from other sources, e.g. prevalent chronic disease, unintentional weight loss, weight fluctuation, or residual confounding from cigarette smoking. As a result, the analyses may underestimate risks of excess weight.

This study estimated the number of deaths for 1991. Since then there has been continued growth in population size and there also appears to be continued growth in the proportion of the population who are obese and severely obese. These trends are likely to increase the number of obesity– attributable deaths.

† This study is based on data from the United States.

†† BMI (Body Mass Index) is calculated as weight (in kilograms) divided by height (in metres) squared. The unit of measurement is kg/m2.