ABSTRACT
Background and Aims
High fasting blood glucose (FBG) ranks fifth among the global risk factors for disease burden, influenced by both individual and environmental factors. However, the roles of social and neighborhood environmental factors remain underexplored in China. This study aimed to estimate the relationship between neighborhood environment, social factors, and FBG among adult residents, and explore potential interaction.
Methods
This cross‐sectional study included 440 participants from two public hospitals in China. We collected data on social factors, FBG, and participant characteristics using electronic questionnaires, anthropometric measurements, and physical examination. The neighborhood environment features were measured using the geographic information systems (GIS) methods. Linear regression analysis was used to examine associations between neighborhood environment features and social factors and FBG.
Results
This study found that marital status (married VS unmarried; β = 0.21, p < 0.001) and the highest quartile of residential housing density (Q4 vs Q1, β = 0.16, p = 0.03) had significant associations with higher FBG levels. The quartiles of the normalized difference vegetation index (NDVI) (Q2, β = −0.14, p = 0.04; Q3, β = −0.19, p = 0.004; Q4, β = −0.21, p = 0.003) and sports facility density (Q2, β = −0.18, p = 0.002; Q3, β = −0.20, p = 0.008; Q4, β = −0.30, p < 0.001) were significantly associated with lower FBG. Furthermore, the second, third, and fourth quartiles of sports facility density and the fourth quartile of residential housing density showed stronger associations with FBG by the interaction with educational level (high vs low). There were also significant interaction effects between residential housing density and subjective economic status, and between NDVI and marital status on FBG.
Conclusion
This study suggests that marital status, residential greenness, and densities of sports facility and residential housing are associated with FBG, and demonstrates potential interaction. Future research is necessary to further clarify these relationships and explore the underlying mechanisms.
Keywords: blood glucose, interaction, neighborhood environment, social factors
Key points
Existing research has identified several key factors (e.g., genetic predisposition, demographic characteristics, lifestyle factors) influencing FBG. However, evidence linking health‐related social factors and the neighborhood environment to FBG remains limited and inconclusive.
This study identified key factors associated with FBG, including marital status, residential greenery, and the density of sports facilities and residential housing.
The findings highlight the need to consider the social and environmental determinants of health in addition to individual lifestyle interventions to promote the metabolic health of the residents more effectively.
1. Introduction
Blood glucose is a critical physiological parameter that reflects the metabolic homeostasis of the body. From 1990 to 2021, the global exposure level and burden of risk factors for high fasting blood glucose (FBG) significantly increased, and it plays an increasingly crucial role in non‐communicable diseases, such as diabetes, cardiovascular disease (CVD), and others [1]. Given its pivotal role in the non‐communicable diseases of human beings, a comprehensive understanding of factors that may influence blood glucose regulation is necessary for developing better intervention strategies.
Existing research has identified several key factors influencing an individual's blood glucose levels. These factors include but are not limited to genetic predisposition, demographic characteristics (e.g., age, sex, and ethnicity), lifestyle factors (e.g., dietary pattern, physical activity, and smoking), and obesity [2]. The Social Determinants of Health theory, however, provides a broader conceptual framework by emphasizing that health outcomes are jointly shaped by biological, behavioral, social, and environmental conditions [3]. From this perspective, blood glucose levels, as an important metabolic health indicator, should be viewed not merely as the result of individual biological and behavioral risks but also as an outcome influenced by broader social factors (e.g., income, education, and living status) and neighborhood environments (e.g., residential greenness and sports facility density) [4]. Incorporating these upstream determinants into FBG research may provide a more comprehensive understanding of glycemic health and inform multilevel interventions targeting both individuals and their living environments.
Although studies have identified associations of marital status and employment status with FBG levels [5], there is still a lack of research that comprehensively explores other potential social factors influencing blood glucose. For instance, personal monthly income has been shown to correlate with physical health status [6], and living alone or having a lower educational level has been associated with an increased risk of CVD [7]. However, the direct associations between these health‐related social factors and blood glucose remain underexplored, thus requiring further research.
With the development of urbanization, a growing body of literature has reported the associations between neighborhood environments, including neighborhood greenery and built environments, and various health outcomes [8, 9]. The neighborhood built‐environment encompasses man‐made structures and infrastructure for human activities, such as residential areas, food facilities, recreational facilities, etc. Existing research has shown that fast‐food restaurants and park access are associated with obesity [10], and that green space is linked with diabetes and major CVD outcomes [8, 11]. A population‐based health study enrolling 39,019 participants demonstrated that exposure to more abundant residential green space was associated with decreased FBG levels [11]. Increased neighborhood greenness may promote healthier movement behaviors (e.g., increased physical activity and reduced sedentary behavior) and reduce psychological stress, which may enhance insulin sensitivity and improve glucose metabolism [12]. In addition, residential greenness may mitigate exposure to environmental stressors, such as air pollution, which has been suggested to be associated with impaired glucose regulation [13]. These behavioral, psychological, and environmental pathways provide a plausible explanation for the potential protective effects of residential greenness on blood glucose regulation. Nonetheless, the evidence regarding the associations between neighborhood environmental factors and FBG levels remains limited and inconclusive.
Furthermore, social and neighborhood environment factors may interact to impact individuals' glucose regulation. Several existing studies have explored the relationships between social and neighborhood environment factors and health behaviors and outcomes, such as diet, physical activity, obesity, and high blood pressure [14, 15, 16]. For example, the health benefits of residential greenness on high blood pressure may be attenuated in individuals with lower socioeconomic status, as they may have differential access to and utilization of neighborhood resources [16]. To the best of our knowledge, there is limited literature on the associations and potential interaction of social and neighborhood environments with FBG levels. Filling this gap is essential for a more thorough understanding of the factors influencing glucose metabolism.
In this study, we aimed to investigate the associations and explore the potential interaction between social factors, neighborhood environment features, and FBG levels among residents from two public hospitals in China.
2. Materials and Methods
2.1. Study Design and Participants
This study was a secondary analysis of data from a cross‐sectional study, which aimed to investigate the relationship between neighborhood environment, lifestyle, and cardiovascular health risk among adult residents. The original study used a convenience sampling method to recruit participants from the physical examination centers of two public hospitals in Hubei Province, China, from September 2023 to January 2024. This study was approved by the Medical Ethics Committee of Wuhan University (Ref. No. WHU‐LFMD‐IRB2023038) and Xiangyang Central Hospital (Ref. No. 2023‐135). All participants provided written consent after being informed of the purpose, content, and confidentiality of the study information. Data were collected by trained research assistants who invited eligible participants to complete the standardized electronic questionnaires, undergo anthropometric measurements, and participate in a physical examination in person.
The participants' inclusion criteria were: (1) age≥ 18 years; (2) living in their current dwelling for at least 1 year; and (3) being able to complete the questionnaires. Participants were excluded if they suffered from any mental illness, cognitive impairments, serious physical illness, or had limited physical activity due to other reasons. In addition, participants with a self‐reported diagnosis of type 1 or type 2 diabetes were excluded from the present analysis to minimize the potential effects of antidiabetic medication use and disease‐related metabolic alterations on FBG levels. We assumed a medium effect size f 2 = 0.15 and set α = 0.05 and the power (1‐β) = 0.9, to estimate sample size using G*Power 3.1.9.2. The ultimately estimated sample size was 375, with the total number of predictors set at 90 (including 6 social factors, 12 dummy variables representing neighborhood environmental factors, and 72 potential interaction terms between social and neighborhood environmental factors). Of the 630 participants recruited in the original study, we excluded 190 due to missing data on FBG (n = 111) and social and neighborhood environment factors (n = 79), leaving a final analytical sample of 440 participants, which provided sufficient statistical power for the analyses.
2.2. Data Collection and Measurements
2.2.1. Social Factors
This study assessed six social factors using self‐designed questions, including marital status, educational background, employment status, personal monthly income, subjective economic condition, and living status. Marital status was categorized as unmarried (including single, divorced, and widowed) and married. Educational background was divided into low educational level (below college) and high educational level (college or above). Employment status was defined as employed if participants were currently in normal working conditions, and unemployed otherwise. Personal monthly income (Chinese yuan) was classified into ≤ 5000, 5000–10000, and > 10000. Subjective economic status was measured by the question, “How do you feel about your family's financial situation compared to others?”, with five response options: very poor, poor, general, good, and very good. Living status was categorized as living alone or not living alone according to the number of household members living together.
2.2.2. Neighborhood Environments
This study assessed four neighborhood environment features using the Geographic Information Systems (GIS) method, including residential greenery and densities of sports facility, residential housing, and food retail facility. Participants' residential addresses were geocoded using Amap, and a 1500 m radius buffer zone around each address was used for measurement. Residential greenery was evaluated by calculating the Normalized Difference Vegetation Index (NDVI), and the remaining three environmental variables were measured by density. NDVI is a remote sensing indicator that quantifies vegetation by calculating the normalized ratio between near‐infrared and red light, with values ranging from −1 to 1 [17]. Data on NDVI were obtained from Earth Data (https://www.earthdata.nasa.gov/), and ArcGIS 10.8 software was used to calculate the average NDVI within each buffer. The calculated final NDVI values of 440 participants ranged from 0 to 1, which could better represent the proportion of greenness [17]. The densities of the sports facility, residential housing, and food retail facility were calculated by dividing their numbers obtained by employing the POI search function by the area of the buffer.
2.2.3. Fasting Blood Glucose
The results of FBG tests were extracted from participants' medical records. A venous blood sample was collected from the median cubital vein before breakfast, after a minimum 8‐hour fasting period for all participants. The blood samples were sent to the specialized blood testing department in the hospital to measure FBG.
2.2.4. Covariates
Demographic characteristics, lifestyles, and body mass index (BMI) were considered as potential covariates in this study. Demographic characteristics included age and sex, with age being converted to a dichotomous variable using 60 years as the cut‐off point. Physical activity level, smoking status (currently/previously smoking or never smoking), and alcohol consumption (yes or no) were measured as lifestyle indicators. We used the Chinese version of the International Physical Activity Questionnaire Short Form (IPAQ‐SF) to assess participants' physical activity levels [18], which were then divided into three groups (low, moderate, and high) according to the recommendation of the IPAQ working group.
2.3. Data Analysis
To evaluate potential selection bias due to missing data, baseline demographic characteristics (age and sex) and lifestyle factors (smoking status, alcohol consumption and physical activity level) were compared between participants included in the analysis (n = 440) and those excluded due to missing FBG or social and neighborhood environmental data (n = 190). Chi‐square tests were applied for categorical variables, and independent‐samples t‐tests were used for continuous variables. No statistically significant differences were found between the two groups (all p > 0.05), indicating that the exclusion of missing data did not introduce substantial selection bias.
In the final analytical sample, continuous variables were described by mean values and standard deviations (SD) or median and inter‐quartile range (IQR), and categorical data were expressed as frequencies and percentages. Neighborhood environment variables (including NDVI, and densities of sports facilities, residential housing, and food retail facilities) were originally continuous but, due to their skewed distributions, were converted into quartiles (Q1–Q4) for analysis. These quartile‐transformed variables were entered as categorical predictors using dummy coding, with the lowest quartile (Q1) as the reference group. With age and gender as auxiliary variables, the 64 missing BMI values were imputed using multiple imputation [19].
Linear regression models were exploratory and were used to analyze the association between social and neighborhood environmental factors (independent variables) and FBG level (dependent variables). Firstly, we analyzed the relationship between each social and neighborhood environmental factor and FBG separately, using an unadjusted univariate linear regression model and a series of adjusted multiple regression models by controlling for potential covariates. Independent variables that showed a significant association with FBG were then included in multiple linear regression analyses. We constructed three models for all linear regression analyses: (1) the first model was unadjusted; (2) the second model adjusted for participants' demographic characteristics; (3) the third model adjusted for participants' demographic characteristics, lifestyle factors, and BMI. To examine whether the association between neighborhood environment and FBG levels varied by social factors, interaction terms were further included in the final multiple linear regression analysis [20], and visualization was performed using the “ggplot2” package. All the above linear regression analyses were exploratory.
All analyses were performed using R (Version 4.4.2) with a two‐sided test, and a p < 0.05 was considered statistically significant.
3. Results
3.1. Characteristics of Participants
Participant characteristics are summarized in Table 1. A total of 440 participants were included in the current study. The mean age of participants was 36.90 ± 11.39 years, and 58.2% were female. Among the participants, 15.2% were current or previous smokers, 58.0% drank alcohol, and nearly a quarter engaged in a high level of physical activity (24.3%). The mean value of FBG was 5.11 ± 0.50 mmol/L.
Table 1.
Characteristics of participants (n = 440).
| Variables | Categories | N (%)/Mean ± SD/median (IQR) |
|---|---|---|
| Demographic characteristics and lifestyles | ||
| Age, years, mean ± SD | 36.90 ± 11.39 | |
| Age groups | ＜60 | 423 (96.1%) |
| ≥ 60 | 17 (3.9%) | |
| Sex | Female | 256 (58.2%) |
| Male | 184 (41.8%) | |
| Smoking status | Current or previously smoking | 67 (15.2%) |
| Never smoking | 373 (84.8%) | |
| Alcohol consumption | Yes | 255 (58.0%) |
| No | 185 (42.0%) | |
| Activity level | Low | 165 (37.5%) |
| Moderate | 168 (38.2%) | |
| High | 107 (24.3%) | |
| BMI, kg/m2, mean ± SD | 22.95 ± 3.17 | |
| FBG, mmol/L, mean ± SD | 5.11 ± 0.50 | |
| Social factors | ||
| Educational background | Low educational level | 38 (8.6%) |
| High educational level | 402 (91.4%) | |
| Marital status | Unmarried | 144 (32.7%) |
| Married | 296 (67.3%) | |
| Living alone | Yes | 31 (7.0%) |
| No | 409 (93.0%) | |
| Employment status | Employed | 393 (89.3%) |
| Unemployed | 47 (10.7%) | |
| Personal monthly income, Chinese Yuan | ≤ 5000 | 103 (23.4%) |
| 5000 ~ 10000 | 237 (53.9%) | |
| ＞10000 | 100 (22.7%) | |
| Subjective economic condition | Very poor | 3 (0.7%) |
| Poor | 15 (3.4%) | |
| General | 311 (70.7%) | |
| Good | 99 (22.5%) | |
| Very good | 12 (2.7%) | |
| Neighborhood environments | ||
| NDVI, median (IQR) | 0.34 (0.189) | |
| sports facility density, median (IQR) | 26.67 (29.7) | |
| Residential housing density, median (IQR) | 39.56 (43.7) | |
| Food retail facility density, median (IQR) | 76.00 (58.0) |
Abbreviations: BMI, body mass index; FBG, fasting blood glucose; IQR, inter‐quartile range; NDVI, Normalized Difference Vegetation Index.
Most participants had higher education (91.4%) and were employed (89.3%). Approximately two‐thirds of participants were married (67.3%), and 7.0% lived alone. More than half of the participants had a personal monthly income of 5,000–10,000 (53.9%), and 70.7% perceived their subjective economic condition as general.
3.2. Associations Between Social and Neighborhood Environmental Factors and FBG
The results of the bivariable linear regression analyses for each social and neighborhood environmental factor are presented in Supplementary Table 1 and Supplementary Table 2. The results showed that better educational background (β = −0.12, p = 0.01) had a statistically significant correlation with lower FBG levels, whereas being married (vs. unmarried; β = 0.25, p < 0.001) and reporting better subjective economic status (β = 0.10, p = 0.03) were significantly associated with higher FBG levels. Regarding neighborhood environmental factors, compared with the lowest quartile, the second, third, and highest quartiles of NDVI (Q2, β = −0.17, p = 0.004; Q3, β = −0.22, p < 0.001; Q4, β = −0.24, p < 0.001) were significantly associated with lower FBG levels. The highest quartile of sports facility density (Q4 vs Q1, β = −0.16, p = 0.007) was also linked with lower FBG levels, and the highest quartile of residential housing density (Q4 vs Q1, β = 0.13, p = 0.02) was correlated with higher FBG levels.
The results of multiple linear regression are summarized in Table 2. Among social factors, marital status (married vs unmarried, β = 0.21, p < 0.001) was significantly associated with higher FBG levels. Compared with the lowest quartile, the second, third, and highest quartiles of NDVI (Q2, β = −0.14, p = 0.04; Q3, β = −0.19, p = 0.004; Q4, β = −0.21, p = 0.003) and sports facility density (Q2, β = −0.18, p = 0.002; Q3, β = −0.20, p = 0.008; Q4, β = −0.30, p < 0.001) had significant associations with lower FBG levels. The highest quartile of residential housing density (Q4 vs Q1, β = 0.16, p = 0.03) was significantly correlated with higher FBG levels. All associations remained significant after controlling for covariates.
Table 2.
Multiple associations between social and environmental factors and fasting blood glucose.
| Model 1 | Model 2 | Model 3 | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| β‐Coefficient (95% CI) | p | SE | t | β‐Coefficient (95% CI) | p | SE | t | β‐Coefficient (95% CI) | p | SE | t | |
| Educational background (high vs low) | −0.05 (−0.24, 0.08) | 0.30 | 0.08 | −1.04 | −0.04 (−0.23, 0.10) | 0.43 | 0.08 | −0.80 | −0.04 (−.23, 0.09) | 0.40 | 0.08 | −0.85 |
| Marital status (married vs unmarried) | 0.21 (0.12, 0.31) | < 0.001 | 0.05 | 4.40 | 0.21 (0.12, 0.32) | < 0.001 | 0.05 | 4.46 | 0.20 (0.11, 0.31) | < 0.001 | 0.05 | 4.16 |
| Subjective economic condition | 0.08 (−0.01, 0.14) | 0.09 | 0.04 | 1.69 | 0.06 (−0.03, 0.13) | 0.21 | 0.04 | 1.27 | 0.06 (−0.03, 0.13) | 0.21 | 0.04 | 1.26 |
| NDVI (Ref: Q1, the lowest) | — | — | — | — | — | — | ||||||
| Q2 | −0.14 (−0.30, −0.01) | 0.04 | 0.07 | −2.11 | −0.14 (−0.30, −0.02) | 0.03 | 0.07 | −2.24 | −0.14 (−0.30, −0.02) | 0.03 | 0.07 | −2.21 |
| Q3 | −0.19 (−0.36, −0.07) | 0.004 | 0.07 | −2.91 | −0.18 (−0.36, −0.07) | 0.004 | 0.07 | −2.90 | −0.18 (−0.36, −0.06) | .005 | 0.07 | −2.82 |
| Q4 (the highest) | −0.21 (−0.40, −0.09) | 0.002 | 0.08 | −3.05 | −0.20 (−0.38, −0.07) | 0.004 | 0.08 | −2.90 | −0.20 (−0.39, −0.08) | 0.004 | 0.08 | −2.93 |
| Sports facility density (Ref: Q1, the lowest) | — | — | — | — | — | — | ||||||
| Q2 | −0.18 (−0.33, −0.07) | 0.003 | 0.07 | −3.02 | −0.16 (−0.31, −0.05) | 0.007 | 0.07 | −2.73 | −0.16 (−0.32, −0.05) | 0.006 | 0.07 | −2.77 |
| Q3 | −0.19 (−0.38, −0.06) | 0.008 | 0.08 | −2.67 | −0.17 (−0.36, −0.03) | 0.02 | 0.08 | −2.38 | −0.17 (−0.36, −0.03) | 0.02 | 0.08 | −2.37 |
| Q4 (the highest) | −0.30 (−0.52, −0.18) | < 0.001 | 0.08 | −4.16 | −0.28 (−0.49, −0.16) | < 0.001 | 0.08 | −3.82 | −0.28 (−0.50, −0.16) | < 0.001 | 0.09 | −3.85 |
| Residential housing density (Ref: Q1, the lowest) | — | — | — | — | — | — | ||||||
| Q2 | 0.08 (−0.04, 0.23) | 0.16 | 0.07 | 1.40 | 0.09 (−0.03, 0.24) | 0.12 | 0.07 | 1.55 | 0.08 (−0.04, 0.23) | 0.16 | 0.07 | 1.40 |
| Q3 | 0.10 (−0.06, 0.27) | 0.21 | 0.09 | 1.26 | 0.10 (−0.06, 0.28) | 0.20 | 0.08 | 1.30 | 0.09 (−0.07, 0.27) | 0.24 | 0.09 | 1.18 |
| Q4 (the highest) | 0.16 (0.02, 0.38) | 0.03 | 0.09 | 2.18 | 0.16 (0.02, 0.39) | 0.03 | 0.09 | 2.21 | 0.16 (0.02, 0.38) | 0.03 | 0.09 | 2.13 |
Note: Model 1: unadjusted multiple linear regression model (F = 6.290, R2 = 0.150, Adjusted R2 = 0.126). Model 2: adjusted for demographic characteristics (F = 6.158; R2 = 0.169, Adjusted R2 = 0.141). Model 3: adjusted for demographic characteristics, lifestyle factors, and body mass index (F = 4.905; R2 = 0.173, Adjusted R2 = 0.138). The inter‐quartile range (IQR) is 0.189 for NDVI; 29.7 for sports facility density; and 43.7 for residential housing density. Q1 represents the lowest density/normalized Index. Q4 represents the highest density/normalized Index.
Abbreviation: NDVI, normalized difference vegetation index.
3.3. Interactions Between Social and Neighborhood Environmental Factors on FBG
Figure 1, Figure 2, and Figure 3 illustrate the interactions between social and neighborhood environmental factors on FBG levels. Table 3 presents the detailed statistical results of the interaction effects. The second, third, and fourth quartiles of sports facility density (Q2: β = −0.65, p = 0.01; Q3: β = −0.69, p = 0.03; Q4: β = −0.89, p = 0.007) and the fourth quartile of residential housing density (Q4 vs Q1: β = 0.75, p = 0.02) showed stronger associations with FBG through interaction with educational level (high vs low). Subjective economic status had a significant interaction with residentiazl housing density at the highest quartile (Q4 vs Q1, β = 0.41, p = 0.02), indicating that the association between the highest quartile of residential housing density and FBG levels was stronger among individuals with lower subjective economic status. Additionally, the association between marital status (married vs unmarried) and FBG was strengthened through interaction with the third quartile of NDVI (Q3 vs Q1, β = 0.41, p = 0.02).
Figure 1.
FBG levels compared by neighborhood environment across education background. Environmental variables were categorized into quartile‐based dummy variables, with Q1 (lowest quartile) as the reference group (Level 0), and Q2–Q4 as comparison groups (Level 1). The inter‐quartile range (IQR) is 0.189 for NDVI; 29.7 for sports facility density; and 43.7 for residential housing density. Q1 represents the lowest density/normalized index. Q4 represents the highest density/normalized index. Panels A–C show the interaction between education and environmental quartile dummies (Q2–Q4 vs Q1) in relation to FBG levels, respectively. Abbreviation: NDVI, Normalized Difference Vegetation Index.
Figure 2.
FBG levels compared by neighborhood environment across marital status. Environmental variables were categorized into quartile‐based dummy variables, with Q1 (lowest quartile) as the reference group (Level 0), and Q2–Q4 as comparison groups (Level 1). The inter‐quartile range (IQR) is 0.189 for NDVI; 29.7 for sports facility density; and 43.7 for residential housing density. Q1 represents the lowest density/normalized index. Q4 represents the highest density/normalized index. Panels A–C show the interaction between education and environmental quartile dummies (Q2–Q4 vs Q1) in relation to FBG levels, respectively. Abbreviation: NDVI, Normalized Difference Vegetation Index.
Figure 3.
FBG levels compared by neighborhood environment across subjective economic condition. Note: Environmental variables were categorized into quartile‐based dummy variables, with Q1 (lowest quartile) as the reference group (Level 0), and Q2–Q4 as comparison groups (Level 1). The inter‐quartile range (IQR) is 0.189 for NDVI; 29.7 for sports facility density; and 43.7 for residential housing density. Q1 represents the lowest density/normalized index. Q4 represents the highest density/normalized index. Panels A–C show the interaction between education and environmental quartile dummies (Q2–Q4 vs Q1) in relation to FBG levels, respectively. Abbreviation: NDVI, Normalized Difference Vegetation Index.
Table 3.
Interaction effects between social and environmental factors and fasting blood glucose.
| Educational background (high vs low) β (95%CI) | Marital status (married vs unmarried) β (95%CI) | Subjective economic condition β (95%CI) | ||
|---|---|---|---|---|
| NDVI (Ref: Q1, the lowest) | Q2 | −0.01 (−0.514, 0.493) | 0.20 (−0.116, 0.519) | −0.02 (−0.259, 0.297) |
| Q3 | −0.14 (−0.777, 0.501) | 0.41 (0.082, 0.746) | 0.15 (−0.110, 0.407) | |
| Q4 | −0.28 (−1.036, 0.479) | 0.15 (−0.188, 0.495) | 0.06 (−0.226, 0.344) | |
| Sports facility density (Ref: Q1, the lowest) | Q2 | −0.65 (−1.163, −0.137) | 0.15 (−0.138, 0.437) | 0.05 (−0.201, 0.294) |
| Q3 | −0.69 (−1.230, −0.075) | 0.06 (−0.332, 0.446) | −0.06 (−0.367, 0.238) | |
| Q4 | −0.89 (−1.536, −0.247) | 0.02 (−0.362, 0.394) | −0.04 (−0.364, 0.281) | |
| Residential housing density (Ref: Q1, the lowest) | Q2 | 0.54 (−0.294, 1.366) | 0.04 (−0.253, 0.338) | 0.11 (−0.135, 0.351) |
| Q3 | 0.60 (−0.010, 1.208) | 0.14 (−0.244, 0.517) | 0.11 (−0.209, 0.425) | |
| Q4 | 0.75 (0.113, 1.402) | 0.09 (−0.476, 0.300) | 0.41 (0.072, 0.744) | |
Note: Interaction effects were the β value of the product interaction term and its 95% CI. Adjusted for demographic characteristics, lifestyle factors, and body mass index. The inter−quartile range (IQR) is 0.189 for NDVI; 29.7 for sports facility density; and 43.7 for residential housing density. Q1 represents the lowest density/normalized index. Q4 represents the highest density/normalized ndex. Bold indicates p < 0.05.
Abbreviation: NDVI, normalized difference vegetation index.
4. Discussion
In the present study, we comprehensively analyzed the associations between a range of social and environmental factors with FBG levels among residents in China. The results revealed that marital status, residential greenery, and densities of residential housing and sports facility were independently associated with FBG levels, even after adjusting for demographic characteristics, lifestyle factors, and BMI. This study also identified potential interactions between social and neighborhood environmental factors on FBG levels.
In terms of social factors, married participants showed higher FBG levels compared to unmarried individuals, including those who were single, divorced, or widowed. Aligning with our findings, a previous study among healthcare workers reported that married individuals had a significantly higher risk of prediabetes compared to their unmarried counterparts, with elevated triglyceride levels acting as a partial mediator [21]. However, several studies demonstrated inconsistent findings [22, 23]. For instance, one study showed that being divorced was associated with a higher incidence of diabetes mellitus compared with those who were married and had never married [22]. Marriage has traditionally been considered to have a positive influence on health outcomes due to factors such as increased household income, reduced stress, and improved sleep quality. The inconsistency between our findings and previous studies may be explained by several factors. One possible explanation could be the role of overweight and obesity, which are more prevalent among married individuals and are well‐known to significantly influence FBG levels [24]. The theory of social facilitation of eating suggests that people tend to eat more when dining with others, such as family or friends, compared to when eating alone [25]. Several mechanisms, such as prolonged meal times, mood influences, appetite processes, and social norms, could contribute to this phenomenon in married individuals, which potentially results in higher caloric intake and weight gain. Another possible explanation for the discrepancy in findings may be attributable to the methodological differences. In this study, we grouped men and women for analysis and combined single, widowed, and divorced participants into a category due to the small sample of each group, and this may have obscured potential differences in FBG levels among these distinct groups. Our findings suggest that the relationship between marital status and health, particularly regarding glycemic control, is complex. Further research is needed to explore in depth their relationship and underlying mechanisms from multiple perspectives, including diet, weight, and psychological factors (e.g., marital quality and happiness).
In this study, we found that residential greenery had a protective association with FBG levels, consistent with previous studies on the health benefits of greenness [26]. For example, a study reported that participants living in the greenest areas had a lower relative risk of developing diabetes mellitus and a lower relative risk of developing CVD compared with those living in the least green areas [27, 28]. Several possible mechanisms may explain the observed association. Firstly, residential greenery could influence an individual's FBG levels through its impact on promoting healthier body weight, possibly due to higher opportunities for physical activity and/or recreational activities with higher residential greenery. Secondly, residential greenery can improve air quality, as poor air quality has been associated with an increased risk of type 2 diabetes [29]. Thirdly, access to green space has been linked with improved mental well‐being, particularly through green physical activities [30], which may be another mechanism. Excessive mental stress triggers a release of cortisol, crucial for regulating metabolic functions. This study added evidence supporting the health benefits of residential greenery in glycemic control and further emphasized the importance of the construction of neighborhood green environments.
Our study also identified a positive correlation between higher residential housing density and higher FBG levels, which was consistent with previous findings on the relationship between residential housing density and health outcomes [31, 32]. A 5‐year cohort study showed that higher building coverage was associated with an increased incidence of lung cancer [31]. High‐density housing may contribute to increased pollutant emissions due to the large population density, and the compact building structures could restrict air circulation, which together exacerbate air pollution. Although high residential housing density might confer certain health benefits, these advantages are limited to low levels of air pollution [32]. Moreover, high residential housing density has been associated with increased prevalence of overweight and obesity, a well‐established risk factor for elevated FBG [33]. From a mental health perspective, residential housing density may be linked to mental health, possibly due to factors such as increased noise levels, limited personal space, and reduced access to outdoor areas, all of which can contribute to higher stress and depressive symptoms. To the best of our knowledge, this study is the first to investigate the link between residential housing density and FBG levels, which provides complementary evidence to existing literature.
Consistent with previous research [34], this study revealed that higher sports facility density was associated with lower FBG levels. This relationship may be explained by the pathway through physical activity. A higher sports facility density is likely to enhance accessibility and convenience while promoting social interaction, leading to greater participation in physical activity. Therefore, the construction and availability of sports facilities may have a protective effect on FBG levels.
The interaction analysis showed that educational background, marital status, and subjective economic status had varying interaction effects with neighborhood environment indicators. Guided by the theory of social determinants and the health ecology model, our findings provided preliminary evidence for how social and environmental factors interact to influence the metabolic health of residents. Further research is warranted to comprehensively investigate their interactions, especially in the context of rapid urbanization.
Our findings enriched our understanding of the complex relationships between multifaceted social and environmental factors and FBG levels. These insights offer new perspectives for the development of health policies and health practices to guide targeted interventions to enhance glycemic control and support diabetes prevention and management. The findings suggest that health care professionals should consider social and environmental determinants and integrate environmental modifications to maintain glycemic homeostasis and reduce the burden of metabolic disorders. Efforts to improve neighborhood greenery, increase the availability of sports facilities, and manage residential density could promote engagement in physical activities, improve air quality, and support mental well‐being, which in turn may contribute to better glycemic control and overall support for diabetes prevention and management.
Several limitations should be acknowledged in our study. We were unable to explore the mechanisms underlying the influence of social factors and neighborhood environments on FBG levels due to the study's scope. Future research should investigate these mechanisms to enhance our understanding. Additionally, the cross‐sectional design limits our ability to establish causality. Longitudinal and experimental designs are recommended to provide more robust evidence and establish causal relationships. Furthermore, the relatively young, highly educated, and employed participants may limit the external validity of our findings.
5. Conclusion
This study revealed the complexity of the association between marital status and FBG levels, and demonstrated that efforts on improving neighborhood greenery, increasing the availability of sports facilities, and managing residential density could contribute to better glycemic control. These insights offered new perspectives for health providers to develop targeted interventions to effectively enhance glycemic control and support diabetes management among residents.
Author Contributions
Xiao Ji Zhang: writing – original draft, writing – review and editing, visualization, formal analysis, methodology, investigation. Yunyan Xianyu: writing – review and editing, visualization, formal analysis, methodology. Tianyu He: writing – review and editing, formal analysis, methodology, visualization. Guiqin Zhang: writing – review and editing, investigation. Jiatong Cui: writing – review and editing, investigation. Yu Xin Cheng: writing – review and editing, investigation. Yu Jia Liu: writing – review and editing, data curation, investigation. Jie Chen: writing – review and editing, visualization. Qian Liu: writing – review and editing, visualization, data curation. Dan Luo: writing – review and editing, supervision, resources. Xiao Qin Wang: writing – review and editing, visualization, resources. Bing Xiang Yang: writing – review and editing, project administration, conceptualization. Huijing Zou: writing – review and editing, supervision, project administration, funding acquisition, methodology, conceptualization.
Disclosure
All authors have read and approved the final version of the manuscript. Huijing Zou (corresponding author) had full access to all of the data in this study and takes complete responsibility for the integrity of the data and the accuracy of the data analysis.
Ethics Statement
This study was approved by the Medical Ethics Committee of Wuhan University (Ref. No. WHU‐LFMD‐IRB2023038; approved on August 1, 2023) and Xiangyang Central Hospital (Ref. No. 2023‐135; approved on November 13, 2023). All procedures were performed in accordance with the Declaration of Helsinki, and all participants provided written consent after being informed of the purpose, content, and confidentiality of the study information.
Conflicts of Interest
The authors declare no conflicts of interest.
AI Use Declaration
The authors used DeepSeek for language polishing and enhancing the readability of the manuscript. All intellectual content, research design, data analysis and conclusions were developed independently by the authors. The authors take full responsibility for the content of this manuscript.
Transparency Statement
The lead author Huijing Zou affirms that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted.
Supporting information
Supporting File
Acknowledgments
The authors would like to express their gratitude to the foundations that provided financial support for the study. Additionally, the authors recognize the collaboration of research assistants from two public hospitals, who contributed to data collection and facilitated the implementation of this study. This work was supported by the National Natural Science Foundation of China [grant number 72304213]; and the Fundamental Research Funds for the Central Universities [grant number 2042023kf0125]. The funders had no role in study design, data collection, analysis, interpretation, writing of the report, or the decision to submit the report for publication.
Zhang X. J., Xianyu Y., He T., et al., “Association and Interaction of Neighborhood Environment and Social Factors With Blood Glucose Among Adult Residents: A Cross‐Sectional Study,” Health Science Reports 9 (2026): e72967, 10.1002/hsr2.72967.
Xiao Ji Zhang, Yunyan Xianyu, and Tianyu He are co‐first authors.
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.
References
- 1. Brauer M., Roth G. A., Aravkin A. Y., et al., “Global Burden and Strength of Evidence for 88 Risk Factors in 204 Countries and 811 Subnational Locations, 1990–2021: A Systematic Analysis for the Global Burden of Disease Study 2021,” Lancet 403, no. 10440 (2024): 2162–2203, 10.1016/s0140-6736(24)00933-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Jia W., Chan J. C., Wong T. Y., and Fisher E. B., “Diabetes in China: Epidemiology, Pathophysiology and Multi‐Omics,” Nature Metabolism 7, no. 1 (2025): 16–34, 10.1038/s42255-024-01190-w. [DOI] [PubMed] [Google Scholar]
- 3. Clifford T., “Social Determinants of Health,” Journal of Perianesthesia Nursing 39, no. 2 (2024): 329–330, 10.1016/j.jopan.2024.01.008. [DOI] [PubMed] [Google Scholar]
- 4. Guo L., Zhang M., Namassevayam G., et al., “Effectiveness of Health Management among Individuals at High Risk of Stroke: An Intervention Study Based on the Health Ecology Model and Self‐Determination Theory (HEM‐SDT),” Heliyon 9, no. 11 (2023): e21301, 10.1016/j.heliyon.2023.e21301. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Ford K. J. and Robitaille A., “How Sweet Is Your Love? Disentangling the Role of Marital Status and Quality on Average Glycemic Levels among Adults 50 Years and Older in the English Longitudinal Study of Ageing,” BMJ Open Diabetes Research & Care 11, no. 1 (2023): e003080, 10.1136/bmjdrc-2022-003080. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Cho J., Son Y., Kim S., et al., “Correlation of Income Instability With 12 Non‐Communicable Disease Risks in Korea: A Prospective Cohort Study,” Journal of Korean Medical Science 40, no. 30 (2025): e180, 10.3346/jkms.2025.40.e180. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Cathro C. J., Brenn T., and Chen S. L. F., “Education Level and Self‐Reported Cardiovascular Disease in Norway‐The Tromsø Study, 1994–2016,” International Journal of Environmental Research and Public Health 20, no. 11 (2023): 5958, 10.3390/ijerph20115958. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Liu M., Meijer P., Lam T. M., et al., “The Built Environment and Cardiovascular Disease: An Umbrella Review and Meta‐Meta‐Analysis,” European Journal of Preventive Cardiology 30, no. 16 (2023): 1801–1827, 10.1093/eurjpc/zwad241. [DOI] [PubMed] [Google Scholar]
- 9. Zhang Y., Zhang Y., van Dijk T., and Yang Y., “Green Place Rather Than Green Space as a Health Determinant: A 20‐year Scoping Review,” Environmental Research 214, no. Pt 4 (2022): 113812, 10.1016/j.envres.2022.113812. [DOI] [PubMed] [Google Scholar]
- 10. Oh J. I., Lee K. J., and Hipp A., “Food Deserts Exposure, Density of Fast‐Food Restaurants, and Park Access: Exploring the Association of Food and Recreation Environments With Obesity and Diabetes Using Global and Local Regression Models,” PLoS One 19, no. 4 (2024): e0301121, 10.1371/journal.pone.0301121. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Li R., Chen G., Jiao A., et al., “Residential Green and Blue Spaces and Type 2 Diabetes Mellitus: A Population‐Based Health Study in China,” Toxics 9, no. 1 (2021): 11, 10.3390/toxics9010011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Odebeatu C. C., Darssan D., Roscoe C., Reid S., and Osborne N. J., “Residential Greenspace Indicators and Metabolic Syndrome in the UK Biobank Cohort: Mediation Through Behavioural, Environmental, Social and Biomarker Pathways,” Environmental Research 283 (2025): 122148, 10.1016/j.envres.2025.122148. [DOI] [PubMed] [Google Scholar]
- 13. Nagrani R., Wolters M., Buck C., et al., “Effect of Long‐Term Air Pollution Reduction on Insulin Resistance and Fasting Glucose in Children: A Causal Analysis,” Environmental Research 292 (2026): 122920, 10.1016/j.envres.2025.122920. [DOI] [PubMed] [Google Scholar]
- 14. C.‐P. L. van Erpecum, , S. K. R. van Zon, , Bültmann U., and Smidt N., “The Association between the Presence of Fast‐Food Outlets and BMI: The Role of Neighbourhood Socio‐Economic Status, Healthy Food Outlets, and Dietary Factors,” BMC Public Health 22, no. 1 (2022): 1432, 10.1186/s12889-022-13826-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Wang Y., Steenbergen B., E. van der Krabben, , Kooij H.‐J., Raaphorst K., and Hoekman R., “The Impact of the Built Environment and Social Environment on Physical Activity: A Scoping Review,” International Journal of Environmental Research and Public Health 20, no. 12 (2023): 6189. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Wang W., Yang C., Wang J., et al., “Lower Regional Urbanicity and Socioeconomic Status Attenuate Associations of Green Spaces With Hypertension and Diabetes Mellitus: A National Representative Cross‐Sectional Study in China,” Environmental Health and Preventive Medicine 29 (2024): 47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Kim K., Joyce B. T., Zheng Y., et al., “Associations of Urban Blue and Green Spaces With Coronary Artery Calcification in Black Individuals and Disadvantaged Neighborhoods,” Circulation 150, no. 3 (2024): 203–214, 10.1161/circulationaha.123.067992. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Craig C. L., Marshall A. L., Sj??str??m M., et al., “International Physical Activity Questionnaire: 12‐country Reliability and Validity,” Medicine & Science in Sports & Exercise 35, no. 8 (2003): 1381–1395, 10.1249/01.Mss.0000078924.61453.Fb. [DOI] [PubMed] [Google Scholar]
- 19. Woods A. D., Gerasimova D., Van Dusen B., et al., “Best Practices for Addressing Missing Data through Multiple Imputation,” Infant and Child Development 33, no. 1 (2024): e2407, 10.1002/icd.2407. [DOI] [Google Scholar]
- 20. Wei J., Liu R., Yang Z., et al., “Association of Metals and Bisphenols Exposure With Lipid Profiles and Dyslipidemia in Chinese Adults: Independent, Combined and Interactive Effects,” Science of the Total Environment 946 (2024): 174315, 10.1016/j.scitotenv.2024.174315. [DOI] [PubMed] [Google Scholar]
- 21. Chen Y. H., Lin J. J., Tang H. M., Yang C. W., Jong G. P., and Yang Y. S., “Relationship between Marriage and Prediabetes among Healthcare Workers: Mediating Effect of Triglycerides,” Medicina (Kaunas) 60, no. 9 (2024): 1418, 10.3390/medicina60091418. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. de Oliveira C. M., Viater Tureck L., Alvares D., et al., “Relationship between Marital Status and Incidence of Type 2 Diabetes Mellitus in a Brazilian Rural Population: The Baependi Heart Study,” PLoS One 15, no. 8 (2020): e0236869, 10.1371/journal.pone.0236869. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Pellón‐Elexpuru I., Van Dijk R., Van der Valk I., Martínez‐Pampliega A., Molleda A., and Cormenzana S., “Divorce and Physical Health: A Three‐Level Meta‐Analysis,” Social Science & Medicine (1982) 352 (2024): 117005, 10.1016/j.socscimed.2024.117005. [DOI] [PubMed] [Google Scholar]
- 24. Nikolic Turnic T., Jakovljevic V., Strizhkova Z., et al., “The Association between Marital Status and Obesity: A Systematic Review and Meta‐Analysis,” Diseases 12, no. 7 (2024): 146, 10.3390/diseases12070146. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Ruddock H. K., Brunstrom J. M., and Higgs S., “The Social Facilitation of Eating: Why Does the Mere Presence of Others Cause an Increase in Energy Intake?,” Physiology & Behavior 240 (2021): 113539, 10.1016/j.physbeh.2021.113539. [DOI] [PubMed] [Google Scholar]
- 26. Li F., Liu W., Hu C., et al., “Global Association of Greenness Exposure With Risk of Nervous System Disease: A Systematic Review and Meta‐Analysis,” Science of the Total Environment 877 (2023): 162773, 10.1016/j.scitotenv.2023.162773. [DOI] [PubMed] [Google Scholar]
- 27. Dalton A. M. and Jones A. P., “Residential Neighbourhood Greenspace Is Associated With Reduced Risk of Cardiovascular Disease: A Prospective Cohort Study,” PLoS One 15, no. 1 (2020): e0226524, 10.1371/journal.pone.0226524. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Yu L., Li T., Yang Z., et al., “Long‐Term Exposure to Residential Surrounding Greenness and Incidence of Diabetes: A Prospective Cohort Study,” Environmental Pollution 310 (2022): 119821, 10.1016/j.envpol.2022.119821. [DOI] [PubMed] [Google Scholar]
- 29. Yang B. Y., Fan S., Thiering E., et al., “Ambient Air Pollution and Diabetes: A Systematic Review and Meta‐Analysis,” Environmental Research 180 (2020): 108817, 10.1016/j.envres.2019.108817. [DOI] [PubMed] [Google Scholar]
- 30. Chen K., Zhang T., Liu F., Zhang Y., and Song Y., “How Does Urban Green Space Impact Residents' Mental Health: A Literature Review of Mediators,” International Journal of Environmental Research and Public Health 18, no. 22 (2021): 11746, 10.3390/ijerph182211746. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Wang L., Sun W., Moudon A. V., et al., “Deciphering the Impact of Urban Built Environment Density on Respiratory Health Using a Quasi‐Cohort Analysis of 5495 Non‐Smoking Lung Cancer Cases,” Science of the Total Environment 850 (2022): 158014, 10.1016/j.scitotenv.2022.158014. [DOI] [PubMed] [Google Scholar]
- 32. Zijlema W., Cerin E., Cirach M., et al., “Cities and Mental Health: The Role of the Built Environment, and Environmental and Lifestyle Factors in Barcelona,” Environmental Pollution 346 (2024): 123559, 10.1016/j.envpol.2024.123559. [DOI] [PubMed] [Google Scholar]
- 33. Yin C., Zhang Z., Dai S., and Chen Y., “Built Environments and Obesity: A Framework Considering Residences, Commute Routes, and Workplaces,” Cities 160 (2025): 105842, 10.1016/j.cities.2025.105842. [DOI] [Google Scholar]
- 34. Cereijo L., Gullón P., Del Cura I., et al., “Exercise Facilities and the Prevalence of Obesity and Type 2 Diabetes in the City of Madrid,” Diabetologia 65, no. 1 (2022): 150–158, 10.1007/s00125-021-05582-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supporting File
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.