Frankenfeld Cara L, Leslie Timothy F, Makara Matthew A
Department of Global and Community Health, George Mason University, Fairfax, VA, 22030, USA.
Department of Geography and Geoinformation Science, George Mason University, Fairfax, VA, 22030, USA.
BMC Public Health. 2015 May 14;15:491. doi: 10.1186/s12889-015-1819-x.
Social and spatial factors are an important part of individual and community health. The objectives were to identify food establishment sub-types and evaluate prevalence of diabetes, obesity, and recommended fruit and vegetable consumption in relation to these sub-types in the Washington DC metropolitan area.
A cross-sectional study design was used. A measure of retail food environment was calculated as the ratio of number of sources of unhealthier food options (fast food, convenience stores, and pharmacies) to healthier food options (grocery stores and specialty food stores). Two categories were created: ≤ 1.0 (healthier options) and > 1.0 (unhealthier options). k-means clustering was used to identify clusters based on proportions of grocery stores, restaurants, specialty food, fast food, convenience stores, and pharmacies. Prevalence data for county-level diabetes, obesity, and consumption of five or more fruits or vegetables per day (FV5) was obtained from the Behavioral Risk Factor Surveillance System. Multiple imputation was used to predict block-group level health outcomes with US Census demographic and economic variables as the inputs.
The healthier options category clustered into three sub-types: 1) specialty food, 2) grocery stores, and 3) restaurants. The unhealthier options category clustered into two sub-types: 1) convenience stores, and 2) restaurants and fast food. Within the healthier options category, diabetes prevalence in the sub-types with high restaurants (5.9 %, p = 0.002) and high specialty food (6.1 %, p = 0.002) was lower than the grocery stores sub-type (7.1 %). The high restaurants sub-type compared to the high grocery stores sub-type had significantly lower obesity prevalence (28.6 % vs. 31.2 %, p < 0.001) and higher FV5 prevalence (25.2 % vs. 23.1 %, p < 0.001). Within the larger unhealthier options category, there were no significant differences in diabetes, obesity, or higher FV5 prevalence across the two sub-types. However, restaurants (including fast food) sub-type was significantly associated with lower diabetes and obesity, and higher FV prevalence compared to grocery store sub-type.
These results suggest that there are sub-types within larger categories of food environments that are differentially associated with adverse health outcomes. These observations support the specific food establishment composition of an area may be an important component of the food establishment-health relationship.
社会和空间因素是个人及社区健康的重要组成部分。本研究旨在确定食品经营场所的子类型，并评估华盛顿特区大都市区这些子类型与糖尿病、肥胖症患病率以及推荐的水果和蔬菜摄入量之间的关系。
采用横断面研究设计。零售食品环境的衡量指标计算为不健康食品选择（快餐店、便利店和药店）来源数量与健康食品选择（杂货店和特色食品店）来源数量的比率。创建了两个类别：≤1.0（更健康的选择）和>1.0（不太健康的选择）。使用k均值聚类法根据杂货店、餐馆、特色食品店、快餐店、便利店和药店的比例来识别聚类。县级糖尿病、肥胖症患病率以及每天食用五种或更多水果或蔬菜（FV5）的数据来自行为危险因素监测系统。以美国人口普查的人口和经济变量为输入，使用多重填补法预测街区组水平的健康结果。
更健康的选择类别聚类为三个子类型：1）特色食品店，2）杂货店，3）餐馆。不太健康的选择类别聚类为两个子类型：1）便利店，2）餐馆和快餐店。在更健康的选择类别中，餐馆比例高的子类型（5.9%，p = 0.002）和特色食品店比例高的子类型（6.1%，p = 0.002）的糖尿病患病率低于杂货店子类型（7.1%）。与杂货店比例高的子类型相比，餐馆比例高的子类型肥胖症患病率显著更低（28.6%对31.2%，p < 0.001），FV5患病率更高（25.2%对23.1%，p < 0.001）。在更大的不太健康的选择类别中，两个子类型在糖尿病、肥胖症或更高的FV5患病率方面没有显著差异。然而，与杂货店子类型相比，餐馆（包括快餐店）子类型与更低的糖尿病和肥胖症以及更高的FV患病率显著相关。
这些结果表明，在更大的食品环境类别中存在与不良健康结果有不同关联的子类型。这些观察结果支持一个地区特定的食品经营场所构成可能是食品经营场所与健康关系的一个重要组成部分。
