1. Introduction
Childhood obesity is one of the most serious public health challenges of the 21st century worldwide. The prevalence of obesity and overweight among children, which is not a unique issue in Western developed countries, has risen dramatically in some developing countries, especially China [
1]. According to the China Chronic Disease and Nutrition Surveillance 2015–19 survey, the rates of obesity and overweight among children under 6 years old are 3.6% and 6.8%, respectively [
2]. As the most populous country in the world, the total number of children younger than 6 years old is enormous in China. Many studies have shown that obesity or overweight in childhood is likely to persist into adulthood [
3,
4,
5,
6,
7]. Childhood obesity can lead to various health problems, even premature death [
8,
9,
10]. Obesity and overweight have imposed a heavy economic burden on China and induces considerable medical expenditures each year [
11,
12,
13].
Many studies have indicated that the influx of Western fast food in recent decade conveyed health consequences, especially obesity and overweight [
14,
15,
16,
17,
18,
19]. Since the reform and opening-up, China’s economic growth has achieved miraculous levels. With the vigorous development of the market economy, the catering industry has developed rapidly. In addition, income levels and requirements for quality of life have gradually improved. Food-away-from-home has become an intensely popular trend [
20,
21,
22]. Influenced by socioeconomic development and changes in people’s lifestyle, the Chinese food environment has undergone tremendous changes in recent decades. The food environment has a marked influence on individuals’ dietary habits and food consumption, and it is considered to be an important contributor to obesity [
23,
24,
25,
26,
27,
28,
29,
30,
31]. Among all types of restaurants, fast-food restaurants are favored by most people because of their delicious food, convenience, and relatively low prices. On the other hand, fast-food restaurants have received the most criticism. Many children start eating fast food during preschool in China. With exposure to fast-food restaurants, preschool children and their parents are more likely to increase the frequency of eating fast food due to immediate gratification and convenience, leading to obesity and other health problems [
32,
33]. The plasticity of childhood is great, so early intervention by family and government can prevent the occurrence of obesity and other chronic diseases with time. Therefore, it is crucial to determine the associations between fast-food restaurants and childhood obesity.
Although there are a large number of studies concerning the associations between fast-food restaurants and obesity, they have not all reached the same conclusions [
34,
35,
36,
37]. However, many studies have reached the consensus that by using the number and the distance as proxies of the proximity of fast-food restaurants, it may be difficult to avoid the challenge of endogeneity. They typically utilize a feature of the environment, specifically proximity to highways or highway exits, as an instrument. The number and distance are generally measurements of the availability of highways or highway exits. Highway exits are a determinant of the locations of fast-food restaurants; thus, the number of highway exits per county has been used as an instrument [
38,
39]. Nevertheless, there are some limitations to this method; for instance, some counties do not have any highway exits. Thus, the locations of fast-food restaurants are not determined by highway exits in these counties, and this instrument did not provide any information. Additional studies have used the straight-line distance to the nearest highway as an instrument [
4,
40,
41,
42].
An instrument that is valid in one environment may not be valid in others [
39]. Employing the proximity of highways as an instrumental variable is reasonable for American studies because the placements of fast-food restaurants are usually around highways. However, the Chinese fast-food industry is entirely different, and the placement of fast-food restaurants is not based on accessibility to the highway. Fast-food restaurants tend to be located in densely populated areas with convenient public transportation. In addition, China has a completely developed public transport system that is convenient for both urban and rural residents. Many people choose to commute by bus because of traditional habits and the low cost of the bus system. Therefore, it is more appropriate to consider the availability of the bus when choosing the instrument in our study. Specifically, we chose the straight-line distance from a kindergarten to the nearest bus stop as an instrumental variable. This instrumental variable is in line with the socioeconomic situation in China. More importantly, it is highly associated with the availability of fast-food restaurants and does not affect other factors that might contribute to obesity. At the same time, we expect that the distance from a kindergarten to the nearest bus stop negatively correlates with the number of fast-food restaurants within a specific radius and positively correlates with the distance to the nearest fast-food restaurant.
There has been some research regarding the associations between the food environment and obesity in different countries, especially focusing on preschool children. Koleilat et al. found that both the number of convenience stores and the number of supermarkets had significant and positive associations with obesity in children 3–4 years of age in Los Angeles County [
43]. Salois concluded that the environmental infrastructure was significantly associated with obesity in low-income children 2–4 years of age in the United States (U.S.), and there was heterogeneity between urban and rural areas [
44]. Lakes and Burkart demonstrated a positive association between the density of fast-food restaurants and obesity using survey data from children 5–6 years of age in Berlin [
45]. There were many samples in additional studies as well, including preschool children [
46,
47,
48,
49]. Nevertheless, most studies from China involve children greater than 7 years old, adolescents or adults. Xu et al. found that community exposure to Western fast-food restaurants was positively associated with adults’ body weight status in the rural population, and this association was temporally dynamic [
50]. Using cross-sectional data in 2019 and adjusted multiple linear regression, Zhang et al. identified some significant associations between the neighborhood food environment and obesity in older Chinese adults 65 to 80 years of age [
51]. In summary, we found no evidence in China on the associations between fast-food restaurants and obesity specifically in preschool children.
Based on the above analysis, there are limitations from three aspects in the existing literature. First, possibly due to a lack of data, there has been little evidence about the associations between fast-food restaurants and obesity in preschool children in China. Second, the challenge of endogeneity is often ignored when the multivariate linear model is exploited in many studies. Third, the samples from China used in most studies come from a given region and are not representative of China as a whole. We filled these gaps and contributed to the existing literature in three aspects. First, we combined three unique cross-sectional datasets from 2018, physical fitness data of 75,730 preschool children, the availability of fast-food restaurants from the Gaode map application programming interface (API), and the financial data of each kindergarten from the Center for Education Economics and Statistics of China to estimate the associations between fast-food restaurant proximity surrounding kindergartens and obesity. Second, combining with the socioeconomic situation in China, we chose the distance from a kindergarten to the nearest bus stop as an instrumental variable to validly address the possible endogeneity problem. Finally, our sample at the individual level comes from 785 kindergartens in 82 cities and 23 provinces in China. These kindergartens are located in the eastern, central, and western regions of China, including areas at all levels of economic development. Therefore, it reflects the general situation in China.
In this study, we estimated the associations between the availability of fast-food restaurants and preschool childhood obesity using two-stage least squares (2SLS). We also analyzed the heterogeneity between urban and rural kindergartens and between females and males. Before this empirical analysis, three hypotheses are proposed:
Hypothesis 1 (H1). The number of fast-food restaurants surrounding kindergartens has a positive correlation with childhood obesity.
Hypothesis 2 (H2). The straight-line distance from a kindergarten to the nearest fast-food restaurant has a negative correlation with childhood obesity.
Hypothesis 3 (H3). The correlations between fast-food restaurants differ between urban and rural kindergartens and between females and males.
4. Discussion
Our study contributes to the existing literature with firm evidence that fast-food restaurants have a positive correlation with obesity in China. Robust findings appeared after the potential endogeneity problem was addressed. We also identified heterogeneity between urban and rural kindergartens.
First, we combined three unique data sources, which are all for preschool. In general, most Chinese studies about health issues use data from the China Health and Nutrition Survey (CHNS) [
18,
50,
58]. However, due to confidentiality, it is impossible to obtain specific information at the individual level, such as the residential address or name of the school. Some studies have also used data at the individual level by questionnaires or surveys. However, the data are generally from one region, and the results are not generalizable to all of China [
13,
51]. Our data are from 85 cities and 23 provinces, which reflects the universal situation in China. Our study specifically focused on preschool children, which filled the gaps about the associations between fast-food restaurants and obesity in preschool children. In addition, all our datasets are official and objective, providing specific and solid evidence.
Second, we used two proxies of the proximity of fast-food restaurants: the number of fast-food restaurants within a specific radius and the straight-line distance to the nearest fast-food restaurant. These are direct and valid for measuring the availability of fast-food restaurants surrounding each kindergarten. More importantly, changing the proxy variable did not affect the positive associations between of fast-food restaurants and obesity, which was in line with our expectations.
Third, we exploited the 2SLS method to address the potential endogeneity problem by using an instrument. This instrument, the distance to the nearest bus stop, is a proxy of the proximity of public transportation and passes all instrumental variable tests. Different from previous American studies using the proximity of highways as an instrument [
4,
38,
39,
40,
41,
42], our choice is based on the fact that China has a completely developed public transport system, and fast-food restaurants tend to be located in places that are easily accessible. Therefore, instead of using the proximity of highways as an instrument such as in previous American studies, using bus stops proximity is more appropriate in China.
Fourth, the comparison between the results of OLS and 2SLS indicates some fascinating findings. The results of OLS show that fast-food restaurants have no statistical correlations with preschool childhood obesity. In contrast, the result entirely changed once the potential endogeneity was addressed, and these results were always robust. Fast-food restaurants have a positive and significant association with obesity, which is consistent with Lakes and Burkart [
45]. Compared with previous studies, our findings are consistent with Asirvatham et al. [
42], indicating that this association decreases as the distance increases. Using two definitions of fast-food restaurants, we found that Western fast-food restaurants exerted a greater influence on obesity than the broader definition.
Moreover, many studies have considered heterogeneity and reported different results in subsamples [
39,
44]. We considered the heterogeneity between different types of kindergarten and gender. The results of the urban subsamples were similar to those of the whole sample, but the correlations between fast-food restaurants and obesity were not statistically significant in the rural subsample. This difference between urban and rural kindergartens indicates that preschool children in urban areas are more sensitive to fast-food restaurants. The correlations between fast-food restaurants and childhood obesity are all significant in both the female and male subsamples, and the values of the estimated coefficients were slightly different between females and males. Therefore, the correlation between fast-food restaurants and obesity is different only between urban and rural areas for preschool children.
Regrettably, there are several limitations in our study. First, we did not have access to comprehensive, time-varying data. Thus, we were unable to capture the dynamic trend of how the associations between fast-food restaurants and childhood obesity change over time. These problems are addressed later when panel data are available. Second, due to the lack of data on eating patterns and consumption behaviors, we could not analyze the deeper cause and mechanism of the effect of fast-food restaurants on obesity. Third, we did not obtain information on the family status of each child to control for the factors of the parents. Using the financial data that were available at the kindergarten level, we addressed this problem to some extent.
Finally, there have been no laws or regulations restricting the distribution of fast-food restaurants surrounding kindergartens in China yet. However, our findings confirm that a positive association between fast-food restaurants, especially Western fast-food restaurants, and preschool childhood obesity does exist. This provides policymakers the inspiration that appropriate regulations are possibly necessary and might need to limit the proximity of fast-food restaurants surrounding kindergartens. Furthermore, infrastructure construction can be strengthened, such as constructing more parks or playgrounds around kindergartens. Health behavioral guidance should be provided to parents to prevent the prevalence of chronic diseases as early as possible.