Obesity error rates high due to measurement method, say researchers

The rates of overweight and obesity may be seriously underestimated due to errors in the most commonly used measuring methods. Error rates can be as high as 45 to 50 per cent, according to research from NUI Maynooth.

"We may be underestimating how serious the problem really is," said Prof Donal O'Neill of the university's department of economics. He did a biostatistical analysis of data from the US on 33,994 people and found high error rates depending on the measuring methods used.

The work also suggests that the growth in obesity began much earlier than assumed, thus slowing a public health response to the problem.

“Our findings have important policy implications,” he said. “If we are going to understand the causes of obesity we have to be able to measure it properly.”

Gold standard

The “gold standard” for assessing whether a person is overweight or obese involves immersing them in water, but this is not a practical approach, he said. As a substitute body mass index (BMI) is often used or bioelectrical impedance analysis (BIA).

BMI is a calculation based on weight in kilogrammes divided by height in metres squared. A person is considered overweight if their BMI is above 25 and obese if it is above 30.

BIA involves the person standing on a metal device and passing a weak electrical current through their feet. The greater the electrical resistance, the more fat a person carries.

These measures have known limitations, he said. “The question is how do you take several imperfect measures and bang out an accurate measure.”

Prof O’Neill decided to compare BMI and BIA along with a simple assessment of weight status, measuring waist circumference.

A waist measure above 102cm for a man or 88cm for a woman indicates a high risk of obesity, he said.

Errors

His analysis showed all three methods produced errors, but the waist measure was consistently lowest at just 3 per cent, Prof O’Neill said. “If you are not obese there is not much error with BMI. With BMI you have about a 45 per cent chance that if you are obese it will fail to class you as obese and that is the problem.”

BMI ranks athletes as obese because of muscle mass, but older people can be read incorrectly. “You lose muscle mass as you get older and your weight and BMI are going down but it it is muscle you are losing not fat,” he said.

The electrical impedance approach proved to have the highest error rates. It over estimated obesity, giving “a very high probability of a false positive”, Prof O’Neill said.

The simple waist measure “exhibits high degrees of accuracy suggesting this measure may provide a cheap and effective measure of latent obesity,” he said.