Mari Gallagher’s response to “Studies Question Pairing of Food Deserts and Obesity”
This is in response to the lengthy but flawed New York Times article on food deserts and obesity that appeared on April 18. I write from the perspective of a researcher who has been deeply engaged for many years in these matters and whose studies have helped stimulate solutions in my hometown of Chicago as well as other parts of the nation.
The Times piece begins with a misstatement that policy makers and first lady Michelle Obama think that all poor urban areas are food deserts. There are many poor urban areas in which residents do have significant access to healthy food options. But food deserts can and do exist in urban, rural and even suburban locations. In Chicago, many food desert residents are poor. We also identified more than 12,000 food desert households that earn $100,000 or more annually.
Ms. Kolata, who wrote the Times story, states, “It is unclear how the idea took hold that poor urban neighborhoods were food deserts,” but there is really nothing unclear about it at all. The existence of food deserts in many U.S. cities is not an idea, but an established fact.
Our research firm popularized the term “food desert” in the U.S. in 2006 with the release of a report titled Examining the Impact of Food Deserts on Public Health in Chicago. Additionally, the National Center for Public Research, of which I am the founding president, launched a highly successful three-year food desert awareness campaign shortly thereafter.
Once the awareness war was won, we retired the campaign to focus our energy on a kaleidoscope of solutions that include – but certainly are not limited to – improving healthy food access. Ms. Kolata named the National Center for Public Research, but appears to have missed or overlooked the details of the campaign or that my firm has found and reported statistically significant relationships in Chicago between a lack of access to nutritious food options and two crucial indicators of negative health impacts: higher body mass index, which is a proxy for obesity, and increased incidents of premature death by diabetes.
We found similar results in many other locations all across the country, and so have many other researchers. In our latest study – which took place on the East Coast and is being finalized for release later this spring – we found a statistically significant relationship between overweight newborns and poor food access, after controlling for a number of factors (i.e. race, prenatal care, mother's education, mother's age, mother's alcohol use, mother's tobacco use, marital status and gestational age).
We have stressed throughout the course of our work that plopping down a grocery store does not mean that these problems are instantly solved. Yet Ms. Kolata’s article unfairly suggests that community leaders, policy makers, Mrs. Obama, and so many others want to “combat the obesity epidemic simply by improving access to healthy foods.” [emphasis added] To my knowledge, no one of any credibility has ever suggested that access was the entire solution or that anything involving the complicated relationship between diet and health is simple.
Healthy food access is a necessary and important foundation to build upon – we cannot choose healthy food unless we have access to it. Once we do have access, other factors that drive individuals to make unhealthy food choices come into play. Behaviors do not change overnight. We all have a lot of work to do. Thankfully, many different community, policy, government and market leaders and organizations in my hometown of Chicago and all across the country – including Mrs. Obama – are working on aspects of this complicated and urgent problem.
Ms. Kolata’s summary of two recent studies on the link between child obesity and access to healthy food was also misleading in several respects. She fails to note the large number of studies that have identified food deserts and the subsequent large number of studies that have found a link between living in underserved areas and poor health outcomes. The article fails to note the shortcomings of the two studies it touts, even though the authors of those studies themselves go to great lengths to describe those deficiencies.
For example, the California study clearly notes that its results stand in contrast to previous work that employed a substantially larger sample. The authors note that their own findings may result from a lack of “statistical power” – statisticians’ shorthand for a sample of insufficient size. The study by Helen Lee also notes that the longitudinal data she uses following children from kindergarten through fifth grade does an insufficient job following urban, minority children in single parent households with low levels of parental education. While we need to stay away from stereotypes, these are the groups most likely to find themselves in food deserts. In Chicago, there is almost a one-to-one overlap between the location of food deserts and the highest concentration of single women with children under 18 years of age.
Other shortcomings discussed in these studies, but not mentioned in the Times article, include problems with the data sources employed: InfoUSA and Dun & Bradstreet. The California study mentions one attempt to validate the data sources used in both studies described here, but substantially understates the negative conclusion of that validation exercise: “[T]he validity of common data sources used to characterize the food environment is limited. The marked undercount of food outlets and the geospatial inaccuracies observed have the potential to introduce bias into studies evaluating the impact of the built food environment.” (Liese et al., American Journal of Epidemiology, 2010). These “canned” data sources, though adequate for many purposes, are insufficient to accurately identify food retailers providing a healthy range of food options from those that are not, a labor-intensive effort that requires significant empirical research and observation.
Not all grocery stores identified by these databases are equal, and many are not even grocery stores at all, which is why we coined the terms “mainstream” and “fringe” for coding purposes. A mainstream store can be small, medium, or large. It does not need to be a chain store. It can even be a corner store – as long as it carries the types of foods that would support a healthy diet on a regular basis. Such knowledge comes only from painstaking fieldwork, verifying the presence of stores and their offerings. We purchased InfoUSA data in the past, but it was so inaccurate and incomplete that it was of no use to us.
Another shortcoming – again, discussed in these studies, but not in the Times article –
concerns the failure to account for how access to food retailers is different in suburban locations where automobile use is nearly universal, compared to urban locations where fewer residents drive and must travel on foot, by taxi, or by mass transportation to obtain the food, nutritious or not, that makes up their daily diet. Lumping suburban and urban places together in the analysis is inappropriate: A retailer that is a few miles away might be reached in a drive of a few minutes in a suburb, but this would take much longer on foot and on public transportation in a city, with an even more difficult return journey with sacks of groceries.
The article also does a disservice to all of us concerned about access to nutrition by setting up a “straw man”: an imagined world in which more grocery stores and improved access to healthy food is the only solution to problems like obesity. Anyone who claims that access is the silver bullet is indeed foolish. But our own research – and the research of many others – makes no such claims. Education, for example, can help people make better choices.
But all the knowledge in the world will not allow food desert residents to choose healthy food if they do not have access to it. Solutions to the challenge posed by obesity lie in both access and education, and in more studies that reveal what works and what doesn’t in changing peoples’ behavior. Cost, culture, and preference are also factors. The solutions do not lie in the misleading presentation of a few contrary findings, the limitations of which even their authors readily acknowledge.
Our issue is not with the two new studies; we thank the authors for their valuable contributions. Our issue is the reporter’s sloppy job of getting the facts straight. Some of this could have been settled by some simple Google searches. She muddied the water at best, misled at worst, and left the inaccurate impression that food access and the concept of food deserts does not matter.
Response by Mari Gallagher, written April 18, 2012
[New York Times, April 18, 2012, pp. A1 and A3