The Developments Of The Perceived Food Environment Questionnaire And The Associations Between Food Environment And Obesity

BackgroundObesity continues to be a serious public health concern all around the world.According to World Health Organization(WHO)statistics,more than 1.9 billion adults aged over 18 were overweight in 2016;of those,over 650 million adults were obese.The worldwide rate of obesity showed an almost three-fold increase from 1975 to 2016.Data compiled by the World Health Organization(WHO)has shown that the prevalence of overweight and obesity in China have been consistently rising over the past few decades,accounting for 32.3%and 6.2%of the adult population in 2016,respectively.In response to this epidemic,many studies have highlighted the importance of molecular genetic or inherent factors that may cause obesity;however,there are many other factors that lack a proper explanation.Therefore,researchers and health professionals are progressively focusing on upstream factors,such as food environment and related policies.However,there is inconsistency in the association between food environment and health outcomes across race or ethnicity,sex,region,and urbanism,that inconsistency complicates explanations of finding.Few studies have examined the neighborhood food environment in developing countries,especially in China.To the best of our knowledge,at present there are no established and standardized tools to measure the perceived food environment for Chinese residents,and no systematic field surveys evaluating community food environment.These research limitations lead to difficulties in accurately measuring and evaluating the food environment in China.In this case,the aim of this study was to develop and validate a tool of perceived food environment for Chinese residents;to conduct a ground-truthing field survey in a typical area in China,and to evaluate the food environment;to conduct a preliminarily analysis between food environment and BMI by combining the food environment field survey data and population epidemiology data,so as to provide urgent suggestions to the government.Methods1.Development of food environment perception questionnaire includes five steps:(1)Collection,inventory and arrangement of survey items through systemic literature review;(2)External expert review on selected survey items and structure of the questionnaire;And revising the survey according to experts' comments and suggestions;(3)Pilot testing and cognitive interviews among a small sample of local residents;And revising the survey and conducting formal survey;(4)Item selection,e.g.test-retest reliability analysis,Cronbach's alpha analysis,correlation analysis,distinguishability analysis;(5)Assessing reliability and validity of the normal scale by test-retest reliability,homogeneity reliability,content validity,structure validity,aggregation validity and discrimination validity.2.A ground-truthing field census of food outlets was conducted in Xi Hu District,Hangzhou.A novel categorizing system was developed to better characterize food outlets in the study area.We investigated the whole District with 800-m buffer.All data were obtained through ground-truthing and direct observation.To assess food availability,we developed a spreadsheet.It contained key information about food outlets,including type of outlet and served foods,seat availability,fruit and vegetable availability,and delivery service availability.The location of each food outlet was accurately positioned using Global Positioning System(GPS)equipment.Food outlets were categorized into nine main types and 33 subtypes.We also classified food outlet types as BMI-healthy,BMI-intermediate and BMI-unhealthy on the basis of the 33 subtypes.We calculated both density per area and density per 10000 people.Geocoding and food environment spatial distribution analysis were performed using ArcGIS 10.2 software.To compare the spatial distribution of food outlets between urban and rural areas,we downloaded Landsat ETM-2017/02/28 from the United States(US)Geological Survey,the boundaries between urban and rural areas was identified by supervised classification and visual interpretation.3.The data of food environment field survey was combined with the data of WELL-China(Wellness Living Laboratory,WELL)large-scale cohort study population.We located residents' addresses on ArcGIS system.Next,800 m and 1500 m buffer area were set up with each resident as the center.Then,number of food outlets in 800 m and 1500 m buffer area were calculated tor each resident respectively.Finally,we conducted a preliminarily analysis on the association between food environment and obesity.Results1.According to literature research,expert view,pre-survey and cognitive Interview,18 items were included in formal survey,including 3 detentions and 5 sub-demensions.The test-retest reliability was 0.741 and Cronbach's alpha coefficient was 0.822.2.Three items were excluded after five statistical analyses,then the final scale contains 15 items,3 dimensions,and 5 sub-dimensions.The correlations of scales were 0.777 in test-retest reliability analyses.In internal consistency reliability analysis,the standardized Cronbach's a coefficient of scale was 0.846.The cumulative variance contribution rate of the five common factors of the formal scale was 70.05%.The average variance extraction value(AYE)of the scale was 0.512,and the correlation between the common factors is low.3.We identified 9274 food outlets in the study area.Among these,6161(66.43%)were food service places and 3,113(33.57%)were food stores.The overall density of food outlets in the study area was 29.72/km2 and 113.97/10,000 people.The density of food stores was 9.98/km2 and 38.25/10,000 people;that of food service places was 19.74/km2 and 75.72/10,000 people.4.In the study area,41.86%of all food outlets provided food-delivery services;among them,approximately 60%of food service places were food-delivery service providers.5.Total Number of BMI-unhealthy food outlets(6579)was about four times greater than that of BMI-healthy food outlets(1505).6.Density of food outlets in the urban area was 104/km2,and density of food outlets in rural area was 29/km2.7.After adjusting factors such as age,sex,marriage,education level,income,smoking,drinking and physical activity level,BMI was positively correlated(P<0.05)with the number of fast food restaurants,full-service restaurants,specialty snack shops and dessert shops in the 800m and 1500m buffer areas around residents.Conclusion1.After item selection,the perceived food environment questionnaire included 3 dimensions,5 sub-dimensions,and 15 items.2.The test-retest reliability,internal consistency reliability,construct validity,convergent validity and discriminant validity of the formal scale were good.3.In the study area,availability of food outlets was excellent,though it showed regional imbalance.4.BMI-unhealthy food outlets were far more numerous than BMI-healthy ones.5.The food-delivery service was prosperous,which shaped a more sophisticated food environment.6.The distribution of food outlets in urban and rural areas differed greatly.7.The food service places like fast food restaurants,full-service restaurants,special snack shops and dessert shops in the study area is relatively unhealthy,which contributes to increased risk of obesity.