Quote of note:

But when Ezzati and his colleagues looked at individual states, they found huge disparities. In Texas, for example, estimates corrected for inaccurate self-reporting suggest that 30% of men and 37% of women were obese in 2000. By comparison, self-reported figures from the BRFSS of the same year suggest that between 18% and 24 % of men and women were obese. A similar discrepancy was found in the 1990 data.

Ezzati stresses that correcting the data for self-reporting shows that the problem of obesity is in fact much worse – particularly in the southern states – than previously thought. Socioeconomic changes may have caused this region of the country to have a greater obesity problem, he speculates.

US states grossly underestimate levels of obesity
22:00 26 April 2006
NewScientist.com news service
Roxanne Khamsi

Widespread wishful thinking has left some US states with gross underestimates of their obesity rates, a new study suggests.

People generally misreport their height and weight to paint a slimmer picture of themselves when answering health-survey questions by phone, say researchers. And as a consequence, obesity levels in some southern states have been underestimated by as much as 50%.

“This study can make a serious contribution to help us get to grips with the obesity epidemic,” comments Morgan Downey, director of the American Obesity Association in Washington, DC. He adds that the findings have “enormous implications on the policy level”.

Widespread wishful thinking has left some US states with gross underestimates of their obesity rates, a new study suggests.

People generally misreport their height and weight to paint a slimmer picture of themselves when answering health-survey questions by phone, say researchers. And as a consequence, obesity levels in some southern states have been underestimated by as much as 50%.

“This study can make a serious contribution to help us get to grips with the obesity epidemic,” comments Morgan Downey, director of the American Obesity Association in Washington, DC. He adds that the findings have “enormous implications on the policy level”.

Face-to-face

About a third of the US population is obese, according to the US Centers for Disease Control and Prevention (CDC) – more than twice the proportion of three decades ago.

To gauge the nation’s health, researchers rely on information gathered by the agency’s National Health and Nutritional Examination Survey (NHANES). The survey involves face-to-face interviews at participants’ homes along with a clinical evaluation at a nearby mobile examination centre. However, this examines only a handful of sites throughout the US.

The CDC also supports the much larger Behavior and Risk Factor Surveillance Survey (BRFSS), which takes a state-by-state view of obesity and gathers information through telephone surveys.

Inconsistencies between the two surveys have appeared in the past, which epidemiologists chalk up to the different methods of collecting information.

Fact and fiction

The new analysis demonstrates how the actual measured weight of participants is higher than that which they self-report in the NHANES. Moreover, the self-reported weights appeared higher in the NHANES – where people were measured straight after - than in the BRFSS. The analysis then uses this data to adjust the BRFSS data, giving a more realistic view of obesity rates.

This makes sense given the context of the surveys, explains Majid Ezzati of the Harvard School of Public Health in Boston, Massachusetts, US, who led the analysis. “If you know you’re going to be measured in person you’re less likely to exaggerate your height or understate your weight,” he says.

Ezzati’s team compared the actual, measured NHANES obesity numbers with the BRFSS figures collected by phone interviews. The researchers found that women under 60 years of age who answer questions over the phone probably understate their weight by about 4 kilograms on average.

Men, on the other hand, were more likely overstate their height by about 2.5 centimetres, according to the study.