Where people shop for their food and the risk of obesity-related cancers

A team of researchers at the University of Connecticut have developed a novel tool to help understand consumer behaviour at the county level, and to study the relationship between where people shop for their food and the risk of obesity-related cancers.

Department of Allied Health Sciences researchers Assistant Professor Ran Xu and PhD student, Gaofei Zhang, Department of Geography, Sustainability, Community, and Urban Studies researchers including Associate Professor Peter Chen and PhD students Weixuan Lyu and Congcong Miao, and Qinyun Lin from the University of Gothenburg, Sweden, used nation level large-scale GPS-tracked mobility data to study where people shop for their food and location data indicating the availability of healthy food within each neighbourhood for about 359,000 retailers across the US. Then, they linked this with obesity-related cancer data to see if where they shop for food impacted health outcomes.

Ran Xu (Credit: Jason Sheldon/UConn Photo)

“Obesity is closely related to cancer and every year, 5% of the new cancer cases for men, 10% of cancer cases for women can be attributed to excessive obesity or body weight, so it is a very serious problem,” explained Xu. “From a public policy perspective, intuitively, you think that if you open healthy grocery stores, that will improve people’s diet quality and then reduce obesity.”

That is unfortunately not the case. In previous studies, Xu and colleagues found that more factors are involved than just increasing access to places to buy healthy groceries.

“We found that people go to food retailers that are mostly outside their immediate neighbourhoods. Also, in our previous research we found that by tracking where people go, we can link that with a series of cardiac metabolic diseases like obesity, hypertension, high blood pressure and diabetes” he added. “We found the strongest association with obesity. In this paper, we go deeper into that and try to see if there is a link between peoples’ food retailer visit pattern and obesity-related cancer mortality.”

The researchers used GPS data from 2018-2019 from SafeGraph paired with retailer location information from InfoGroup. The researchers designated healthy food locations as grocery stores, large warehouse clubs, and farmers’ markets. Convenience stores and fast-food establishments were designated as unhealthy food sources. This information was then compared with obesity-related cancer mortality data from the Centers for Disease Control and Prevention from 2015–2020.

They used the data to create an activity-based index of the percentage of visits to healthy food establishments made by people from each county. The location-based index was calculated based on the percentage of healthy food establishments located in each county.

“Long story short, we found that tracking people’s activity and where they go is a much stronger predictor than location for obesity-related cancer mortality events,” added Xu.

The researchers found the activity-based index predicted obesity-related cancer risk twice as effectively as the traditional location-based approach. Xu says this finding further emphasizes that more research is needed to capture the complexities of human behaviour to effectively address the obesity pandemic problem.

Chen, the geographer on the team, explained that the association also differs by many sociodemographic characteristics. For example, the association is stronger in Hispanic communities, possibly explained by the “Hispanic Paradox” where Hispanic populations tend to have better health outcomes and healthier behaviours than other groups despite the more disadvantaged risk factor profile.

“We also found that this association is strong in residents with higher socioeconomic status (SES) and we think that is probably because, for example, for people with higher SES when they go to grocery stores, they have the financial means to purchase something healthy,” added Chen.

Xu and his colleagues are leveraging digital tools like GPS data to drill down into the nuance of these problems and the seemingly conflicting results of current policy interventions.

“If you think about food deserts, which are low-income areas without grocery stores, and the government will put resources into there, for example opening new grocery stores,” said Xu. “However, based on our data, because most of the food retailer visits are outside their own neighbourhood, we have to wonder, is this really the right approach? We need to focus on the actual behaviour and activity, and that might give us better policy outcomes.”

The index the researchers developed is an important step, and Xu says their future research will begin to try to understand behaviours and their purchasing decisions. In one of their current projects, they are using GPS data to systematically evaluate the effect of opening new grocery stores in food deserts. They also plan to complement these findings with qualitative research methods, like surveys, to dig into the dynamic and multi-dimensional factors influencing why people are not shopping at healthy grocery stores that may be within their neighbourhoods. Whether due to convenience, price, brand loyalty, and a host of other reasons, further research will help illuminate some of these choices.

The findings were featured in the paper, ‘Association of activity-based food environment index with obesity-related cancer mortality in the US', published in BMC Medicine. To access this paper, please click here