Background Excess bodyweight is associated with negative effects on longevity, disability-free life-years, quality-of-life, and productivity. High prevalence of obesity and over-weight is often related to socioeconomic status, educational attainment, and ethnic group. The combination of rising obesity prevalence and increased spending on obese people has been estimated to account for 27% of the growth in US health-care expenditure between 1987 and 2001.

Methods The Behavioral Risk Factor Surveillance System (BRFSS) is the world’s largest, telephone health survey system, collecting cross-sectional data on health conditions and risk factors in all states of the US yearly since 1984. Telephone interviews provide a cheap alternative to physical examinations and are often used as a way to increase the number of participants in observational studies. Data for 3,814,344 adults from BRFSS were used to investigate trends of obesity levels in each US state. Over the period 1999–2010, the yearly trend in the percentage of the population with body mass index (BMI) in normal-weight (BMI<25), over-weight (BMI 25–29.9), and obese (BMI>30) ranges, is modelled using multinomial logistic regression, stratifying by gender and age group (20–39, 40–59,>60 years). The fitted models are used to compare trends between states and to forecast levels of obesity in the future. A previously developed micro-simulation model is used to assess the burden of 13 diseases caused by obesity and estimate the economic impact implied by the forecasted trends. Data of the US–National Health and Nutrition Examination Survey collected over the same period are used to investigate the extent of self-reporting bias in BRFSS.

Results In 2010, the proportion of obese (BMI>30) men and women in the US was 33% and 35%, respectively, whilst the proportions of over-weight (BMI 25–30) were 42% and 34%. The fitted models forecast an increase in the proportion of obese individuals to up to 70% by 2030. The results reveal increasing levels of obesity in all states, although the rate of increase varies among states. Comparing BRFSS data with data from NHANES, showed that at all ages both men and women slightly over-report their height, whilst women under-report their weight by 5kg on average. Under-reporting of weight is reduced after age 65.

Conclusion Obesity rates are rapidly and steadily increasing in the US posing a threat to population health and a substantial economic burden. As self-reporting bias may significantly underestimate BMI in women, the estimated burden of obesity may be conservative.