Understandably, there is a very strong focus in obesity research on the diets of schoolchildren with many schools now attempting to implement healthy eating policies. Equally, there has been considerable concern about the existence of food retail outlets nearby to schools to which the schoolchildren have access. A group at the University of Oxford has recently published a meta-analysis of all relevant studies, which set out to examine the relationship between obesity outcomes and the proximity of food retail outlets to schools.
The authors completed a search of 10 on-line library databases and identified several thousand studies but, as ever, in meta-analyses, many of the initial studies were rejected for a variety of reasons leaving the authors with 30 full studies which met all of the a priori inclusion criteria. Each study had to have defined exactly what was meant by the retail food environment and to have measure quantitatively the relationship between food purchase patterns and obesity-related outcomes. Most of the papers were published between 2011 and 2013 and most were cross-sectional with children ranging in age from five to seventeen years. More than three quarters had sample sizes of over 1,000.
Of the 30 studies, the majority used a defined “buffer zone” around the school but some used route maps between the pupil’s home and school. GIS (Geographic Information Systems) software was the main source of information on retail outlets either within the designated buffer zone or school route. In general the buffer zone applied a distance of between 0.1 to 3.0 miles while the route approach generally used distances of 50 to 100 meters from the road travelled to and from school. The main outcome studied was the child’s BMI (kg/m2). The second most frequent measure of outcome was food intake but this appeared generally to be related to a narrow range of foods: fruit and vegetables, soda drinks or fast food. Some of course used several measures and just three used the overall diet quality index of the schoolchildren which would have included all sources of foods at all times of the day.
One study focused on fast food purchases and found a statistically significant positive association between fast food purchasing and the density of fast food outlets. Ten studies examined the relationship between food outlets in general and the consumption of sugar sweetened beverages and of fast food, including crisps, sweets, biscuits, fried food, sugar sweetened beverages and fast foods. Within these 10 papers, a total of 54 associations were examined and only two of these showed a statistically significant association. Four papers examined the association between fruit and vegetable consumption and food retail outlets and within these a total of 32 associations were examined. Only three showed statistically significant associations. Within the 30 studies, only three had data on the overall quality of the pupil’s diet and food retail outlet density. Two of these showed a significant association between diet quality of food outlets. In one case, the data showed a significantly higher diet quality index among pupils attending a school where the nearest retail outlet was greater than 1 km away as compared to those where the distance was less than 1km away. The second study found that the greater the distance to the nearest grocers the better was the overall diet of the pupils.
This is an important paper for several reasons. Firstly, it is a very well conducted study published in a high impact journal. Secondly, it highlights how the existence of evidence is happily ignored by those policy makers who want to place restrictions on the availability of food outlets within the vicinity of schools. Thirdly, it shows that the outcome variables which are easy to measure such as fruit and vegetable intake, soda intake or BMI yield fairly useless conclusions since they do not relate the one aspect of the determinants of food choice (school associated food outlets) to the totality of the effects of all food choice in terms of overall daily nutritional quality. Once again, we see a majority of studies in what is a very important area of public health nutrition, bedevilled by bad design. In the three studies, which did look at the overall quality of the pupil’s diet and density of food retail outlets two showed some significant associations. Now do two swallows make a summer?. No, but they point the way forward for the conduct of scientifically rigorous studies in this very important area of public health nutrition. To discover that the proximity of food outlets influenced specific food intake is of zero importance in public health nutrition. We need to know the full accurate daily nutrient intakes and only then can we judge whether any aspect of the obesigenic environment id truly influencing overall nutritional quality.