Measuring Obesity in Brazil
Obesity is defined as an excess of body fat. That is an abstract idea, but what is its operational definition? You can start off by asking people to define themselves as being obese or not, but this invokes subjective cultural norms that may shift over time. In the late twentieth century, skinny models such as Twiggy and Kate Moss are considered to be chic in the western world. A century or two ago, the ideal west woman is much more rotund. In Chinese history, there are two famous beauties, one was rather corpulent and the other renowned for her tiny waist. And then an anorexic will probably consider anything to be overweight.
It is preferable, therefore, to have an objective measure of obesity. If being obese means being overweight, then surely weight has to figure in this measure somewhere. But consider a person who weighs 70 kilograms. Is that overweight or not? Well, that depends on the height of the person. For a very short person, that amount of weight requires a wide girth to carry it. But a very tall person may look quite gaunt at that weight. Therefore, height should also figure somewhere in the measure (but be careful about specifying that the height should be measured 'without shoes on'!).
The most commonly accepted guideline based on weight and height to determine underweight and overweight is the Body Mass Index, defined to be:
Body Mass Index (BMI) = (weight in kilogram) / (square of height in meters)
After calculating the BMI, the following classifications are often used:
We will now look at some anthropometrical measures from the TGI Brasil study. This is a survey of 10,624 persons between the ages of 12 to 64 years old conducted during year 2002. During the survey, the information about height and weight is taken, from which the BMI is calculated. According to the TGI Brasil study, the mean BMI is 24.1, which puts the average Brazilian at the upper end of the 'normal' range.
The first question that we ask is this, "Are rich people obese?" According to the TGI Brasil, the mean BMI among the socio-economic classes of AB, C and DE are respectively 24.1, 24.1 and 24.2 respectively. Therefore, we see no significant differences in mean BMI by socio-economic class. But if we look at the percentages of people who have BMI over 30, then the number is 11.0% in the total sample, and 9.8%, 11.1% and 12.2% in the AB, C and DE classes respectively. So the answer to the original question is "No" and in fact rich people are less likely to be obese. Today, there is no mass starvation in Brazil, but the nutrition and diets are still far from ideal for the poor, and the excess of starch and fat plus the lack of exercise among the poor cause more incidences of obesity.
The next question relates to the incidences of obesity by age/sex groups. This is where BMI breaks down. For children who are still growing, the bones grow faster than the body fat could keep up, so that children have a lower BMI which picks up as they grow older. In the next chart, we show the relationship between BMI by age/sex groups. Therefore, given this type of systematic pattern, it is obviously important to go beyond the single classification scheme and establish age-specific Body Mass Index classifications.
(data source: TGI Brasil)
The moral of the less here is that the haste to find a quick-and-easy quantitative measure should not blind people to the presence of systematic deviations that invalidate the simple measure. Can you think of any other exceptions? Well, how about pregnant women?
(posted by Roland Soong, 5/18/2003)
(Return to Zona Latina's Home Page)