The food environment, which could also be called a nutrition environment, refers to the physical presence of food that affects a person’s diet, a person’s proximity to food store locations, the distribution of food stores, food service, any physical entity by which food may be obtained, or a connected system that allows access to food [1
]. The food environment plays an important role in food choice, eating patterns, and energy intake [2
]. It is widely reported that the food environment is associated with the increasing epidemic of childhood and adult obesity [3
]. Recent efforts on obesity prevention have partly focused on studying the role of environmental factors, and national level policies on the food environment have been implemented in many countries [7
It was reported that the food environment within stores, as it relates to healthy food availability, quality, and price, may contribute to the dietary intake disparities in many countries [10
]. Specifically, higher quality food stores, such as specialized fruit and vegetable (F&V) markets, were associated with greater F&V intake after controlling for individual-level characteristics, according to a study conducted in a Brazilian city [10
]. The proportion of in-store shelf space for skim, 1% fat, and 2% fat milk was reported by a study conducted in the United States to be associated with low-fat milk consumption [11
]. Besides price and convenience, purity, freshness, association with specific places, and ‘Pakistani-ness’ were reported as the bases for making decisions about ‘good food’ in Pakistan [12
Over the past several decades, China has been undergoing economic development and urbanization at an accelerated rate [13
]. In 1950, 13% of people in China lived in cities. By 2010, the urban share of the population had grown to 45%; this is projected to reach 60% by 2030 [14
]. Meanwhile, the total energy intake of Chinese people increased by 80%, from 1635 kcal/capita/d in 1961 to 2943 kcal/capita/d in 2003. By 2006, less than 1 percent of people were getting 10 percent of their energy from fat, 44 percent were getting 30 percent of their energy from fat, and nearly two-thirds were getting more than 10 percent of their energy from animal products [15
]. The prevalence of overweight and obese children increased from less than 2% in 1985 to 15% in 2010, and the overall rate of overweight and obese people in major cities is over 20% [16
]. However, the food environment after such huge societal change in China, and its contribution to the obesity prevalence in China, has yet to be described. Answers these questions would help provide evidence for food environment improvement and potentially limit the obesity epidemic in China.
Lack of valid, reliable measures of food environments hinders food environment studies in China and other Asian countries. Many food environment measurement tools exist for other countries [19
]. Among these, the Nutritional Environment Measures Survey in Stores (NEMS-S) is an observational measure to assess the food environment within retail food stores [21
]. The NEMS-S measures the availability of healthier options, price, and quality for 10 categories of food products. The NEMS-S is characterized by its relative ease of use and ability to adapt to different settings and populations [23
]. The 10 categories of food products included in the NEMS-S were based on the types of food products that contributed the most fat and calories, which are different from those in China, to the American diet. So, before using the NEMS-S in China, food products were culturally adapted in this study to fit the Chinese food culture. This study aimed to adapt and validate a Chinese version of the Nutritional Environment Measurement Survey for Stores (C-NEMS-S) in order to provide a tool for measuring the food environment in China for future studies.
The present study adapted and tested the first tool to measure the retail food environment in China. The original NEMS-S was adapted to the Chinese culture. The C-NEMS-S adapted in this study had high inter-rater reliability and was able to display the differences in healthy food availability between large-sized supermarkets and convenience stores, as well as the price differences between healthier options and regular options.
The NEMS-S is regarded as an observational measure or environmental auditing tool. The auditing ability of raters is critical to obtaining high reliability observations. For the original NEMS-S, the inter-rater kappa coefficient ranged from 0.83 to 1.00 [26
]. For the Brazilian version of the NEMS-S, the inter-rater kappa coefficient ranged from 0.69 to 1.00, and the ICC raged from 0.75 to 1.00 [21
]. The inter-rater reliability of the C-NEMS-S in the present study was only slightly lower than that of the original and the Brazilian version. Nonetheless, both the ICC and kappa coefficient were acceptable, ranging from moderate to high (0.41 to 1.00 for the ICC, 0.52 to 1.00 for the kappa coefficient) [34
Compared with other food categories, milk, bread, instant noodles, and beverages had a lower ICC or kappa. This may be because packaged food categories occupied larger proportions of shelf space. In order to determine healthier options for these food categories, careful reading of food labels and packaging information is needed. Some raters may, therefore, fail to find healthier options for these food categories during this process. Though the ICC and kappa for these food categories were acceptable, future training programs, to teach raters how to measure in-store food environments, should pay more attention to these packaged food categories.
According to the results, large-sized supermarkets had a significantly greater amount of healthier options available for all food measures (all items were statistically significant (p
< 0.05), except for sugar-free beverages) compared with convenience stores. The result was consistent with previous studies on differences between supermarkets and convenience stores. According to previous studies, supermarkets could provide healthier foods at reasonable prices [36
]. For the original NEMS-S in the United States and the adapted NEMS-S in Brazil [21
], supermarkets obtained higher scores on healthier food availability. Hence, supermarkets were hypothesized to score high in the C-NEMS-S, as well. According to the results in the present study, the C-NEMS-S obtained higher availability for all food measures in large-sized supermarkets. These findings in the present study support the ability of the C-NEMS-S to differentiate between supermarkets and convenience stores.
Except for meat and poultry, all healthier options cost more than regular options. This result indicates the ability of the measurement scoring to identify the differences between healthier options and regular options. In addition, this result also suggests that access to various food outlets might not be a problem for the food environment in China, but more attention should be paid to the consumer environment. The price discrepancy between healthier options and regular options may become a target for future policies aimed at improving the food environment in China.
There were several limitations in the present study. Firstly, dietary habit is different between northern China and southern China. The current validation survey of the C-NEMS-S was only conducted in Shenyang, a city in northern China. In addition, the field work happened in winter, so the seasonality of the food offerings could have played a role in influencing reliability testing and findings. However, the C-NEMS-S was developed based on guidelines suitable for the whole of China by referencing the Chinese Dietary Guidelines (2016) [29
] and Chinese food categories in the China Food Composition handbook [28
]. Food categories and species were selected based on a nation-wide perspective and were not limited to northern Chinese cuisine or limited by seasonality. With further validation surveys in cities in southern China and during different times of the year, the C-NEMS-S may be verified to be suitable for the whole of China. Secondly, a small sample size was used to conduct the environment audit in stores and to test the reliability of the C-NEMS-S. To cover more areas of China, a larger sample size is warranted in future validation studies. Thirdly, snacks were not included in the final version of the C-NEMS-S as a food measure. During evaluations, it was too difficult to determine low fat or low sugar healthier snacks from the large variety of snacks. This is a problem to be amended in updated versions of the C-NEMS-S in our future studies.