Wednesday, February 8, 2012

Is it Your Neighbor(hood)'s Fault That You are an Obese Couch Potato? Plus: Higher Incomes Increase Obesity Risk in Men, Better Education Decreases Risk in Men & Women

Image 1: If this photo looks as if it was taken in your neighborhood, statistics say that you will have a harder time than others warding off obesity.
If you have not already been aware that the major weightloss obstacles are not so much of physio- than of psychological, or I should say behavioral natural, Monday's blogpost on the inability or unwillingness of the majority of the study participants in the Krebs study should have reminded you that there is more to losing weight than having an "optimal" diet plan. In this context, the results of a recently published paper from the University of Ottawa, Canada (Prince. 2012), comes to mind, in which Stepanie A. Prince and her colleagues report the results of a large-scale cross-sectional multi-level analysis of the association between neighborhoods, physical activity and obesity in Ottawa.

Mislead and misfed, but by no means unable to afford leading a healthier life-style

Although obviously of epidemiological nature, the study is remarkable in that it relies on relatively recent data (2006-2008) from 494,000 residents from 86 neighborhoods in Ottawa, who were part of the Ottawa Neighborhood Study (ONS). My usual advice not to confuse correlation (no matter how "significant") and causation, does yet still apply. A statement like "for each additional km² of park area per 1,000 inhabitants, the odds of being physically active increased by 17% in the female inhabitants of the respective neighborhood" (cf. figure 1, left; respective value: 1.17) does thusly signify that readily available parks areas seem to encourage women to be more physically active in their leisure time. It does not mean that moving to an apartment from which you can see the joggers doing their rounds in Central Park will turn a couch potato into a sporting ace. After all, it is at least as likely that people who like to jog try to make sure that they move to an area, where they can easily pursue their hobby.
Figure 1: Model predictions for the influence of environmental, social and contextual parameters on physical activity (left) and overweight/obesity rate (right) in men and women; * p < 0.05 (data based on Prince. 2012)
Against that background, I decided to include only factors which showed a statistically significant association with either obesity or physical activity in at least one of the two sexes in the data in figure 1. Now, you rarely have a rule without exception, and in this case, I also included the associative strength with the so-called census-based socio-economic status (SES) index, because the non-existent influence of social status on obesity rates goes just about as nicely against the notion that you cannot eat healthy if you are on a budget, as the +39% increased obesity risk of men in households with incomes >30,000$ (cf. figure 2)
Figure 2: Model predictions for the influence of the individual parameters household income and education overweight/obesity rate in men and women; * p < 0.05 (data based on Prince. 2012)
Moreover, the strong negative association -44% for men and -45% for women between having at least a college degree and being overweight/obese, would would support the argument that it is less about not being able to afford a "healthy lifestyle" than about not having learned / not being able to teach oneself to do so... or, if you will, being more susceptible to the misleading information from the food industry and less aware of the pitfalls of shopping in convenience stores (+17% increased risk of obesity in women per additional convenience store for 1,000 inhabitants) and eating at fast food outlets.
Note: Neither the age nor parameters, such as the number of indoor or outdoor recreational facilities, grocery stores, specialty stores and (normal) restaurants which, at least taken in isolation, had no statistically significant influence on either physical activity or obesity rates.
With respect to the +22% and +39% increased obesity risk per additional fast-food outlet (per 1,000 inhabitants), it should also be mentioned that this is only one out of several of the environmental factors, which had statistically significant influence only on obesity rates in women from the respective neighborhoods - a phenomenon, which could yet be a consequence of the fact that men are less likely to be in their respective neighborhood during working hours than women, so that the susceptibility to fast food stores is probably not gender-specific ;-)

The bitter or delighting truth about your neighbor(-hood ;-)

It is nevertheless quite remarkable that the associative strength of individual criteria such as age, income, education etc. showed an overall much more pronounced variation (0.98, p<0.05), than one of the area-specific variables, the variance of which did not reach statistical significance for either of the two study outcomes, i.e. physical activity level or overweight/obesity (cf. figure 1). This would suggest that environmental influences in the area we live in is in fact a more reliable indicator of our likelihood of ending up as an overweight couch potato or lean physical culturist than our incomes, education or say our age. A finding, which I would say is actually quite remarkable, don't you think so?