Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Key influences on dysglycemia across Fujian’s urban-rural divide

  • LiHan Lin ,

    Contributed equally to this work with: LiHan Lin, XiangJu Hu

    Roles Data curation, Formal analysis, Methodology, Project administration, Resources, Software, Writing – original draft, Writing – review & editing

    Affiliation College of Physical Education, Huaqiao University, Quanzhou, China

    ⨯
  • XiangJu Hu ,

    Contributed equally to this work with: LiHan Lin, XiangJu Hu

    Roles Conceptualization, Data curation, Funding acquisition, Supervision, Visualization, Writing – review & editing

    Affiliations School of Public Health, Fujian Medical University, Fuzhou, China, Department for Chronic and Noncommunicable Disease Control and Prevention, Fujian Provincial Center for Disease Control and Prevention, Fuzhou, Fujian, China

    ⨯
  • XiaoYang Liu,

    Roles Data curation, Formal analysis, Software, Writing – review & editing

    Affiliation College of Physical Education, Huaqiao University, Quanzhou, China

    ⨯
  • GuoPeng Hu

    Roles Conceptualization, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Validation, Writing – review & editing

    hugp@hqu.edu.cn

    Affiliation College of Physical Education, Huaqiao University, Quanzhou, China

    ⨯

Abstract

Background

Screening and treatment of dysglycemia (prediabetes and diabetes) represent significant challenges in advancing the Healthy China initiative. Identifying the crucial factors contributing to dysglycemia in urban-rural areas is essential for the implementation of targeted, precise interventions.

Methods

Data for 26,157 adults in Fujian Province, China, were collected using the Social Factors Special Survey Form through a multi-stage random sampling method, wherein 18 variables contributing to dysglycemia were analyzed with logistic regression and the random forest model.

Objective

Investigating urban-rural differences and critical factors in dysglycemia prevalence in Fujian, China, with the simultaneous development of separate predictive models for urban and rural areas.

Result

The detection rate of dysglycemia among adults was 35.26%, with rates of 34.1% in urban areas and 35.8% in rural areas. Common factors influencing dysglycemia included education, age, BMI, hypertension, and dyslipidemia. For rural residents, higher income (OR = 0.80, 95% CI [0.74, 0.87]), average sleep quality (OR = 0.89, 95% CI [0.80, 0.99]), good sleep quality (OR = 0.89, 95% CI [0.80, 1.00]), and high physical activity (PA) (OR = 0.87, 95% CI [0.79, 0.96]) emerged as protective factors. Conversely, a daily sleep duration over 8 hours (OR = 1.46, 95% CI [1.03, 1.28]) and middle income (OR = 1.12, 95% CI [1.03, 1.22]) were specific risk factors. In urban areas, being male (OR = 1.14, 95% CI [1.02, 1.26]), cohabitation (OR = 1.18, 95% CI [1.02, 1.37]), and central obesity (OR = 1.35, 95% CI [1.19, 1.53]) were identified as unique risk factors. Using logistic regression outcomes, a random forest model was developed to predict dysglycemia, achieving accuracies of 75.35% (rural) and 76.95% (urban) with ROC areas of 0.77 (rural) and 0.75 (urban).

Conclusion

This study identifies key factors affecting dysglycemia in urban and rural Fujian residents, including common factors such as education, age, BMI, hypertension, and dyslipidemia. Notably, rural-specific protective factors are higher income and good sleep quality, while urban-specific risk factors include being male and central obesity. These findings support the development of targeted prevention and intervention strategies for dysglycemia, tailored to the unique characteristics of urban and rural populations.

Citation: Lin L, Hu X, Liu X, Hu G (2024) Key influences on dysglycemia across Fujian’s urban-rural divide. PLoS ONE 19(7): e0308073. https://doi.org/10.1371/journal.pone.0308073

Editor: Palash Chandra Banik, Bangladesh University of Health Sciences, BANGLADESH

Received: March 29, 2024; Accepted: July 17, 2024; Published: July 31, 2024

Copyright: © 2024 Lin et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All relevant data are within the manuscript and its Supporting Information files.

Funding: This study was sponsored by the Fujian provincial natural science foundation, grant number 2020J01087(Guopeng Hu) and Fuian provincial health technology project, grant number 2020GGA023(Xiangju Hu). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist.

Introduction

Dysglycemia, which includes prediabetes and diabetes, represents a growing global health challenge with significant public health implications [1–3]. Prediabetes is an early stage characterized by higher-than-normal blood glucose levels, which increases the risk of developing type 2 diabetes (T2DM) [4]. Without timely intervention, prediabetes can progress to diabetes, a condition characterized by elevated blood glucose due to insufficient insulin production or ineffective insulin use, leading to complications such as cardiovascular disease, kidney failure, and neuropathy [5–7]. Key factors influencing dysglycemia include genetic predisposition, individual health, lifestyle habits, environmental factors, and social security [8–13].

In China, the largest developing country, the dual urban-rural economic structure has led to significant disparities in healthcare services, social security, and resource allocation between urban and rural areas, affecting income levels, educational attainment, and lifestyles [14–16]. These disparities contribute to different risks of chronic diseases, including T2DM. For instance, higher prevalence rates of T2DM in rural areas compared to urban areas in the United States are attributed to the greater commonality of poverty, obesity, and smoking in these rural settings. [17].

Recent advancements in medical science and artificial intelligence have made machine learning a vital tool for analyzing multidimensional medical data. Machine learning techniques, such as Logistic Regression (LR) and the Random Forest (RF) algorithm, are valuable for predicting and explaining health outcomes and assisting in developing prevention strategies and clinical treatment plans [18,19]. LR provides quantitative interpretations of significant factors through odds ratios (ORs), while the RF algorithm, known for its resilience to noise and reduced risk of overfitting, effectively handles complex data environments [20,21]. These methods are widely recognized for their applications in disease prediction and risk factor assessment [22,23].

Given the distinct social structures and historical contexts across nations, understanding dysglycemia susceptibility among urban and rural populations is essential, particularly in China’s pronounced urban-rural divide. This study uses face-to-face survey data from Fujian, China, and applies univariate analysis, logistic regression (LR) analysis, and random forest (RF) models to identify critical determinants of dysglycemia prevalence. Initially, 18 variables were considered: demographic, social security, lifestyle, and physiological health. Through rigorous univariate and LR analysis, key determinants were identified: common factors such as education, age, BMI, hypertension, and dyslipidemia; rural-specific protective factors like higher income, good sleep quality, and high physical activity; and urban-specific risk factors including being male, cohabitation, and central obesity. By developing a random forest model based on these determinants, the study aims to improve screening efficiency for at-risk individuals. By developing a random forest model based on these determinants, the study aims to improve screening efficiency for at-risk individuals. This research highlights the differences in dysglycemia risk factors between urban and rural areas and proposes targeted, evidence-based strategies for prevention and intervention.

Methodology

Study design and setting

This study was based on the Chinese Adults Noncommunicable Disease and Nutrition Surveillance (Fujian segment), a cross-sectional study investigating health-related behaviors and dysglycemia among adults in Fujian Province, China. The baseline dataset used in this study was collected from June 2019 to March 2021.

Considering the geographical, economic, and population characteristics of Fujian Province, a multi-stage cluster random sampling method was employed. In the first phase, 16 districts and counties were selected from among the 86 in Fujian Province using a probability proportionate to size sampling method. In the second stage, within each selected district or county, 6 townships (or streets) were randomly selected using the same method. In the third stage, within each selected township (or street), 3 village committees (or communities) were chosen using simple random sampling, with each having at least 100 households. In the fourth and final stage, within each selected household, 1 individual was surveyed, with the sample size calculated using the Kish Leslie formula [24]. The survey targeted permanent residents of Fujian Province who had lived in the survey areas for 6 months or more, were aged 18 years or older, and excluded pregnant women and individuals with cognitive or language impairments.

A standardized Social Factors Special Survey Form (SFSSF-2019) was utilized, combining face-to-face questionnaire interviews with medical examinations. Uniformly calibrated instruments were employed for physical and laboratory examinations to measure blood pressure, blood lipids, height, and weight, as referenced in studies such as [25–28]. Such as, blood pressure was measured using the Omron HBP-1300 electronic blood pressure monitor on the right upper arm [29]. Measurements were taken three times in a resting state, with intervals of more than five minutes between each measurement. Blood glucose levels were determined using fasting plasma glucose (FPG) [30] and a 2-hour post-75 g oral glucose tolerance test (OGTT) [31] venous blood samples (participants with a history of diabetes did not undergo the glucose challenge).

Dependent variable

The diagnosis of dysglycemia was defined according to the American Diabetes Association (ADA) [4], diabetes was defined as FPG > = 7.0 mmol/L and/or OGTT > = 11.1 mmol/L. Prediabetes was defined as FPG > = 5.6 mmol/L and < = 6.9 mmol/L, OGTT > = 7.8 mmol/L and < 11.0 mmol/L.

Independent variable

Combining literature analysis and data, the study classified the 18 variables influencing dysglycemia into 4 categories: demographic variables, social security variables, lifestyle variables, and physiological health variables.

Demographic variables encompassed gender, age, place of residence, level of education, Body Mass Index (BMI), annual income, and marital status, while social security variables included health insurance coverage. The rural and urban residence indicates household living regions and is defined by the National Bureau of Statistics of the People’s Republic of China [32]. Urban areas include cities and towns, that have higher population density, diverse economic activities, and better access to services, while rural areas refer to regions outside of the designated urban boundaries, and are characterized by lower population density, primary reliance on agriculture, and limited access to healthcare. Regarding the socioeconomic background, compared to other countries, China has a large urban-rural disparity in terms of economic income [33]. The categorization of annual income levels was based on the distribution of per capita disposable income for 2023 as reported by the National Bureau of Statistics of China, with classifications into three tiers: Lower income (Under 20,442 yuan, about 2,815 USD), Middle income (Between 20,442 yuan to 50220 yuan, 2,815–6,916 USD) and Higher income = 2(over 50,220 yuan, about 6,916 USD). Physiological health variables consisted of self-rated health status, dyslipidemia, hypertension, central obesity, and chronic diseases. Central obesity was defined according to the "Guidelines for the Prevention and Control of Overweight and Obesity in Chinese Adults," with a waist circumference ≥ 90 cm for men or ≥ 85 cm for women indicating central obesity [34]. The definition of chronic disease incidence included the presence of coronary heart disease, malignant tumors, chronic digestive system diseases, neck and lumbar diseases, chronic obstructive pulmonary disease (COPD), osteoarthritis, cerebrovascular disease, and chronic urinary system diseases, among eight types of chronic conditions. Lifestyle variables were measured by daily sleep duration, sleep quality, self-rated sleep quality, sedentary behavior, PA levels and frequency of alcohol consumption. Sleep duration was categorized into three types according to the recommendations of the American Academy of Sleep Medicine: insufficient sleep (t < 6 hours/day), appropriate sleep (6 hours/day ≤ t < 8 hours/day), and excessive sleep (t ≥ 8 hours/day) [35]. We assessed participants’ sleep quality over the past month using the Pittsburgh Sleep Quality Index (PSQI) [36]. The PSQI scores range from 0 to 21, with higher scores indicating poorer sleep quality. For analysis, we categorized the PSQI scores into three groups: Bad for PSQI scores greater than 10, Average for PSQI scores between 5 and 10, and Good for PSQI scores of 5 or less [37,38]. PA levels were classified into low, moderate, and high based on the scoring rules of the International Physical Activity Questionnaire (IPAQ) Short Form [39]. Sedentary behavior (SB) is defined as an average of more than 6 hours per day spent sitting or lying down, excluding time spent sleeping at night [40]. The remaining independent variables are shown in Table 1 below.

thumbnail
Download:
Table 1. Variable assignment.

https://doi.org/10.1371/journal.pone.0308073.t001

Data processing

Logistic regression analysis was used to identify factors influencing dysglycemia, and a predictive model was constructed using the random forest algorithm. Data preprocessing and the construction of the RF model were completed in Python 3.9, using Sklearn, NumPy, Matplotlib, and Pandas. Univariate analysis was conducted using the Chi-square test, while multivariate analysis employed LR statistics, with the significance level set at <0.05. This part of the work was performed in SPSS 26. For a detailed research flowchart, refer to Fig 1 below.

thumbnail
Download:
Fig 1. The flow chart of this study.

https://doi.org/10.1371/journal.pone.0308073.g001

Ethics

This research is a branch of the Chinese Adults Noncommunicable Disease and Nutrition Surveillance project, conducted within Fujian Province, China, and led by the National Center for Chronic and Noncommunicable Disease Control and Prevention, Chinese Center for Disease Control and Prevention. The project received approval from its ethics committee (#201819). Adhering strictly to ethical standards, we ensured that all participants were fully informed of the study’s purpose, procedures, potential risks, and benefits prior to their participation, and had signed a written informed consent form.

Results

Urban-rural differences in variables

Upon exclusion of outliers and missing data, information from 26,157 participants (mean age 53.77, SD ± 14.41) was analyzed. Among these respondents, a significant proportion of rural participants had not pursued education beyond primary school (63.0%). The prevalence of overweight and obesity among urban respondents was 46.1%, higher than the 40.5% observed in rural respondents. Furthermore, a significant fraction of urban respondents, representing 66.8%, belonged to the low-income tier. Dysglycemia was diagnosed in 9,224 participants, yielding a prevalence rate of 35.56%, with 6,567 cases in rural areas and 2,657 in urban areas. These findings are elaborated in Table 2.

thumbnail
Download:
Table 2. Descriptive statistics.

https://doi.org/10.1371/journal.pone.0308073.t002

Univariate analysis of factors influencing dysglycemia in urban-rural adults

Among 26,157 urban and rural residents in Fujian Province, the prevalence rates of dysglycemia were 34.1% and 35.8%, respectively, with rural residents exhibiting a higher rate than their urban counterparts (χ2 = 6.905, P<0.01). Significant differences were observed in urban residents across 13 factors (P<0.05), including age, BMI, gender, education, marital status, daily sleep duration, SB, PA, chronic disease, self-rated health, hypertension, dyslipidemia, and central obesity. Similarly, rural elders showed significant variances in depressive symptoms across 13 factors (P<0.05), specifically age, BMI, education, annual income, marital status, daily sleep duration, sleep quality, PA, chronic disease, self-rated health, hypertension, dyslipidemia, and central obesity, as detailed in Table 3.

thumbnail
Download:
Table 3. Univariate analysis of variables influencing dysglycemia in urban-rural adults.

https://doi.org/10.1371/journal.pone.0308073.t003

Logistic regression analysis of key factors of dysglycemia

The results indicate that higher education background is significantly associated with lower odds of dysglycemia in both urban and rural areas. Specifically, in rural areas, higher education is associated with reduced odds of dysglycemia, with odds ratios (OR) as follows: primary school (OR = 0.782, 95% CI = 0.714–0.855), junior high school (OR = 0.868, 95% CI = 0.796–0.946), and junior college and above (OR = 0.765, 95% CI = 0.680–0.848). In urban areas, the OR for individuals with junior college education and above is 0.815 (95% CI = 0.708–0.939).

Additionally, common factors associated with higher odds of dysglycemia in urban and rural residents include BMI (overweight: OR = 1.486, 95% CI = 1.127–1.957; obese: OR = 1.800, 95% CI = 1.322–2.453), age (65 and older: OR = 1.372, 95% CI = 1.211–1.555), hypertension (OR = 1.515, 95% CI = 1.353–1.697), and dyslipidemia (OR = 1.570, 95% CI = 1.419–1.736).

For rural residents, higher income, sleep quality, and PA are associated with lower odds of dysglycemia. Specifically, an annual income exceeding 50,220 yuan (OR = 0.799, 95% CI = 0.735–0.870), average sleep quality (OR = 0.892, 95% CI = 0.802–0.993), good sleep quality (OR = 0.894, 95% CI = 0.801–0.997), and high PA (OR = 0.871, 95% CI = 0.788–0.963) are associated with reduced odds of dysglycemia. Conversely, a daily sleep duration over 8 hours (OR = 1.146, 95% CI = 1.025–1.280), middle income (OR = 1.122, 95% CI = 1.033–1.219), and moderate PA (OR = 1.095, 95% CI = 1.013–1.183) are associated with higher odds of dysglycemia in this demographic.

In urban populations, being male (OR = 1.135, 95% CI = 1.024–1.258), cohabitation marital status (OR = 1.132, 95% CI = 1.022–1.366), and central obesity (OR = 1.351, 95% CI = 1.192–1.530) are associated with higher odds of dysglycemia.

Urban-rural dysglycemia prediction model

Random forest model.

The RF model is an ensemble learning algorithm comprising multiple decision trees. It utilizes the bagging algorithm for random sampling of the dataset to generate multiple training sets. Each training set is analyzed using a decision tree as the base classifier, with the final prediction being the outcome of a majority vote across multiple trees [21].

Model construction.

Employing the Random Forest model, an important analysis was conducted on factors affecting dysglycemia in rural and urban residents, followed by the construction of predictive models. The samples were divided into training and testing sets in a 7:3 ratio, with Rural (N = 18359) and Urban (N = 7798) cohorts. Risk factors identified through logistic regression analysis served as training variables, including Age, BMI, Annual income, Daily sleep duration, PA, Hypertension, and Dyslipidemia for the Rural area, and Age, Gender, BMI, Marital status, Hypertension, Dyslipidemia, and Central obesity for the Urban area.

Grid search was also utilized to determine the optimal number of decision trees, with Out-of-Bag (OOB) error trends visualized as the number of trees varied. The OOB error, an unbiased estimate of the model’s generalization capability, facilitated the identification of the decision tree count that minimizes OOB, thus determining the optimal parameters for the RF model. The findings are illustrated in Fig 2.

thumbnail
Download:
Fig 2. OOB error as the tree number in RF model for rural -urban adults.

https://doi.org/10.1371/journal.pone.0308073.g002

Results analysis

Stabilizing the OOB estimates for both rural and urban RF models occurs when decision trees exceed 450. This stabilization informs the construction of dysglycemia diagnostic models with accuracies of 75.35% for rural and 76.95% for urban settings. The models’ efficacy, illustrated by ROC of 0.7733 and 0.7538 respectively, is depicted in Fig 3.

thumbnail
Download:
Fig 3. ROC curve of the random forest mode.

https://doi.org/10.1371/journal.pone.0308073.g003

Discussion

This study investigates the prevalence and determinants of dysglycemia among urban and rural populations in Fujian Province, China, utilizing data from 26,157 adults collected through comprehensive surveys and medical examinations. The analysis reveals a dysglycemia prevalence of 35.26%, with a slightly higher rate in rural areas (35.8%) compared to urban areas (34.1%), shown in Table 1. Common influencing factors across both settings include age, BMI, hypertension, dyslipidemia, and education level. Rural-specific protective factors identified are higher income and good sleep quality, while middle income and daily sleep duration exceeding eight hours are associated with increased odds of dysglycemia. In urban areas, being male, cohabiting, and having central obesity emerge as unique factors associated with higher odds of dysglycemia. Random forest models used for predicting dysglycemia achieve accuracies of 75.35% for rural and 76.95% for urban populations, and ROC of 0.7733 and 0.7538 respectively, shown in Fig 3.

Over the past few decades, China’s economy has experienced rapid growth, albeit unevenly distributed, particularly affecting rural areas. This economic imbalance represents one of the social determinants contributing to significant health issues among the elderly population in rural China [41]. Consistent with previous surveys [42], Fujian Province in China exhibits low educational attainment with a pronounced urban-rural disparity: 45.7% of rural inhabitants and 35.2% of urban dwellers have not completed primary education (Table 1). Education lays a robust foundation for sustained cognitive growth and living conditions, offering pathways to improved employment prospects and higher income. Moreover, it stimulates the adoption of positive and healthy lifestyle choices. Collectively, these benefits are linked to a reduced likelihood of dysglycemia [43,44]. Therefore, with targeted educational initiatives, we could reduce the likelihood of urban-rural residents developing dysglycemia. Initiatives such as universities for the elderly and community health outreach programs serve as effective means to disseminate health consciousness and foundational medical knowledge [45].

Consistent with our findings, several studies have established a direct correlation between aging and the increased likelihood of developing dysglycemia. [46,47]. Age-related changes in insulin sensitivity, pancreatic β-cell function, and the accumulation of visceral fat are significant contributors to this likelihood [48]. Given the established link between aging and the heightened likelihood of dysglycemia, it becomes clear that intervention measures must pivot toward lifestyle modifications. Intervention measures should emphasize lifestyle modifications. Based on the protective factors identified through LR analysis, these include encouraging physical exercise, enhancing sleep quality, and maintaining a healthy BMI. Advancing targeted and efficient screening for dysglycemia can also contribute to increasing awareness and prevention of dysglycemia. In alignment with our observations, extensive research corroborates that hypertension and dyslipidemia predispose individuals to a heightened possibility of dysglycemia [49–52]. This concordance emphasizes the intricate interplay among these conditions, underscoring the critical need for integrated approaches in screening and managing these co-occurring health risks. Beyond the recommendations, fostering medical equity across urban and rural regions to ensure that low-income rural populations have easier access to medical and rehabilitation services is crucial. Concurrently, conducting regular health assessments for populations at risk of dysglycemia can facilitate early detection, preventing the irreversible progression to diabetes and averting severe health consequences [53,54].

This study’s findings indicate that in rural areas, higher income, good sleep quality, and high-intensity PA are protective against dysglycemia, whereas middle-income, moderate-intensity activities, and daily sleep durations exceeding eight hours are associated with an increased likelihood of dysglycemia. (Table 4). Additionally, the lower education levels and higher incidence of dyslipidemia in rural areas contribute to the higher prevalence of dysglycemia in these regions (Table 2). The farmwork demands and limited health infrastructure in rural areas further contribute to the neglect of health management and regular medical check-ups, exacerbating glycemic control issues. This also explains why the prevalence of dysglycemia is higher in rural areas of Fujian Province, China, compared to urban areas [55]. In rural settings, moderate-intensity physical activities, such as walking and labor primarily due to agricultural and occupational demands, often lack the systematic and sustained intensity found in organized high-intensity exercise, which is known to effectively reduce the likelihood of dysglycemia and provide significant metabolic benefits [56,57]. Furthermore, the dietary habits of rural inhabitants, characterized by high fat and carbohydrate intake will counteract the potential metabolic benefits of these activities, increasing the likelihood of dysglycemia [58,59]. Interestingly, the phenomenon where higher income offers protection against dysglycemia in rural residents, while middle income is associated with an increased likelihood of dysglycemia, could be attributed to lifestyle habits. Higher-income enables better access to healthcare, nutritious foods, and wellness resources, alongside more opportunities for PA and health screenings [60,61]. Conversely, middle-income individuals might afford diets high in meat and processed foods, increasing the likelihood of dysglycemia, unlike lower-income counterparts who may consume more plant-based, high-fiber diets due to financial constraints [62,63]. This scenario illustrates the intricate connections between socioeconomic status, dietary habits, and health, suggesting the importance of nuanced health interventions tailored to different income levels. Consistent with our findings, numerous studies have indicated that prolonged sleep duration (exceeding 8 hours) is associated with an increased likelihood of dysglycemia [64,65]. The link between extended sleep in rural areas and heightened dysglycemia likelihood can be attributed to multiple factors. Excessive sleeping may indicate underlying health issues, such as sleep apnea or depression, which are acknowledged risk factors for dysglycemia [66,67]. Furthermore, longer sleep durations could reflect a sedentary lifestyle, with diminished time dedicated to PA, thereby elevating the chances of dysglycemia. Concurrently, this study aligns with previous research, identifying good sleep quality as a protective factor against dysglycemia [68,69].

thumbnail
Download:
Table 4. Logistic regression results for dysglycemia (rural).

https://doi.org/10.1371/journal.pone.0308073.t004

This study’s findings indicate that in urban areas, being male, cohabitation, and central obesity were identified as unique factors increasing the likelihood of dysglycemia (Table 5). Consistent with our findings, some research indicates that in urban environments, males may have a higher predisposition to dysglycemia compared to their rural counterparts and are also more susceptible to dysglycemia than urban females [70–73]. This risk factor among urban males can be attributed to lifestyle choices and occupational stressors more prevalent in urban settings, differing significantly from those in rural areas and from the lifestyle patterns of urban females. Our study found that cohabitation in urban settings increases the likelihood of dysglycemia. Numerous studies have shown that cohabitation affects individual lifestyle habits; specifically, it may lead to weight gain [74], unhealthy dietary habits [75], and reduced willingness to exercise [76,77], all of which are established risk factors for dysglycemia. Central obesity, marked by excess fat around the stomach and abdomen, is a significant risk factor for dysglycemia in urban populations [78–80]. In the surveyed urban residents, the prevalence of central obesity is notably higher at 35.0%. Therefore, the likelihood of developing dysglycemia due to central obesity is higher in urban areas [81].

thumbnail
Download:
Table 5. Logistic regression results for dysglycemia (urban).

https://doi.org/10.1371/journal.pone.0308073.t005

This study evaluates the determinants of dysglycemia among residents of Fujian Province, China, employing LR to analyze their impact and developing a predictive model via an RF approach. It contributes threefold: firstly, it identifies critical factors for urban and rural dysglycemia, guiding researchers and informing policymakers on targeted prevention strategies; secondly, it assesses these factors’ influence from an urban-rural perspective, the laying groundwork for nuanced intervention in dysglycemia management; thirdly, the model enhances dysglycemia risk screening efficiency, addressing both urban and rural needs. Limitations include the inability of cross-sectional data to infer causality or dynamic variable interactions and the study’s context-specific findings to Fujian’s socio-cultural environment, which may limit broader applicability.

Conclusions

The study, integrating logistic regression analysis and the random forest model, identified education, BMI, age, hypertension, and dyslipidemia as common factors for dysglycemia among urban and rural residents. sleep quality, PA, daily sleep duration, and annual income emerged as key factors for rural residents, while gender, marital status, and central obesity were pinpointed as specific key factors for rural inhabitants.

Acknowledgments

We thank the National Center for Chronic and Noncommunicable Disease Control and Prevention, Chinese CDC, for their support of our Fujian branch project under the Chinese Adults Noncommunicable Disease and Nutrition Surveillance. Special gratitude to Li Xinhua, principal investigator, for his invaluable guidance. Appreciation extends to our team and participants for their crucial contributions.

References

  1. 1. Lee JM, Gebremariam A, Wu E-L, LaRose J, Gurney JG. Evaluation of Nonfasting Tests to Screen for Childhood and Adolescent Dysglycemia. Diabetes Care. 2011;34: 2597–2602. pmid:21953800
  2. 2. Edwards CM, Cusi K. Prediabetes: A Worldwide Epidemic. Endocrinol Metab Clin North Am. 2016;45: 751–764. pmid:27823603
  3. 3. GBD 2019 Diseases and Injuries Collaborators. Global burden of 369 diseases and injuries in 204 countries and territories, 1990–2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet. 2020;396: 1204–1222. pmid:33069326
  4. 4. American Diabetes Association Professional Practice Committee. 2. Classification and Diagnosis of Diabetes: Standards of Medical Care in Diabetes—2022. Diabetes Care. 2022;45: S17–S38. pmid:34964875
  5. 5. Francis EC, Powe CE, Lowe WL, White SL, Scholtens DM, Yang J, et al. Refining the diagnosis of gestational diabetes mellitus: a systematic review and meta-analysis. Communications Medicine. 2023;3: 185. pmid:38110524
  6. 6. Deshpande AD, Harris-Hayes M, Schootman M. Epidemiology of Diabetes and Diabetes-Related Complications. Physical Therapy. 2008;88: 1254–1264. pmid:18801858
  7. 7. Obirikorang C, Obirikorang Y, Acheampong E, Anto EO, Toboh E, Asamoah EA, et al. Association of Wrist Circumference and Waist-to-Height Ratio with Cardiometabolic Risk Factors among Type II Diabetics in a Ghanaian Population. Journal of Diabetes Research. 2018;2018: 1–11. pmid:29670914
  8. 8. Dendup T, Feng X, Clingan S, Astell-Burt T. Environmental Risk Factors for Developing Type 2 Diabetes Mellitus: A Systematic Review. Int J Environ Res Public Health. 2018;15: 78. pmid:29304014
  9. 9. Colosia A, Khan S, Palencia R. Prevalence of hypertension and obesity in patients with type 2 diabetes mellitus in observational studies: a systematic literature review. Diabetes, Metabolic Syndrome and Obesity: Targets and Therapy. 2013; 327. pmid:24082791
  10. 10. Lyssenko V, Jonsson A, Almgren P, Pulizzi N, Isomaa B, Tuomi T, et al. Clinical Risk Factors, DNA Variants, and the Development of Type 2 Diabetes. N Engl J Med. 2008;359: 2220–2232. pmid:19020324
  11. 11. Tobias M. Global control of diabetes: information for action. Lancet. 2011;378: 3–4. pmid:21705073
  12. 12. Agliata A, Giordano D, Bardozzo F, Bottiglieri S, Facchiano A, Tagliaferri R. Machine Learning as a Support for the Diagnosis of Type 2 Diabetes. International Journal of Molecular Sciences. 2023;24: 6775. pmid:37047748
  13. 13. Banik PC, Barua L, Moniruzzaman M, Mondal R, Zaman F, Ali L. Risk of diabetic foot ulcer and its associated factors among Bangladeshi subjects: a multicentric cross-sectional study. BMJ Open. 2020;10: e034058. pmid:32114471
  14. 14. Wang Y, Liu Y, Li Y, Li T. The spatio-temporal patterns of urban–rural development transformation in China since 1990. Habitat Int. 2016;53: 178–187.
  15. 15. Treiman DJ. The “difference between heaven and earth”: Urban–rural disparities in well-being in China. Research in Social Stratification and Mobility. 2012;30: 33–47.
  16. 16. Zhu S, Yu C, He C. Export structures, income inequality and urban-rural divide in China. Appl Geogr. 2020;115: 102150.
  17. 17. O’Connor A, Wellenius G. Rural–urban disparities in the prevalence of diabetes and coronary heart disease. Public Health. 2012;126: 813–820. pmid:22922043
  18. 18. Hohmann E. Editorial Commentary: Big Data and Machine Learning in Medicine. Arthroscopy: The Journal of Arthroscopic & Related Surgery. 2022;38: 848–849. pmid:35248233
  19. 19. Wang Y, Ding Y, Liu X, Li X, Jia X, Li J, et al. Preoperative CT-based radiomics combined with tumour spread through air spaces can accurately predict early recurrence of stage I lung adenocarcinoma: a multicentre retrospective cohort study. Cancer Imaging. 2023;23: 83. pmid:37679806
  20. 20. Sperandei S. Understanding logistic regression analysis. Biochem Med. 2014; 12–18. pmid:24627710
  21. 21. Breiman L. Random Forests. Mach Learn. 2001;45: 5–32. :1010933404324.
  22. 22. Wang J, Wang Y, Chen S, Fu T, Sun G. Urban-rural differences in key factors of depressive symptoms among Chinese older adults based on random forest model. J Affect Disord. 2024;344: 292–300. pmid:37820963
  23. 23. Lin T-Y, Hsieh S-S, Chueh T-Y, Huang C-J, Hung T-M. The effects of barbell resistance exercise on information processing speed and conflict-related ERP in older adults: a crossover randomized controlled trial. Sci Rep. 2021;11: 9137. pmid:33911153
  24. 24. Kish L. Sampling Organizations and Groups of Unequal Sizes. Am Sociol Rev. 1965;30: 564. pmid:14325826
  25. 25. Yu W, Li X, Zhong W, Dong S, Feng C, Yu B, et al. Rural-urban disparities in the associations of residential greenness with diabetes and prediabetes among adults in southeastern China. Science of The Total Environment. 2023;860: 160492. pmid:36435247
  26. 26. Huang S, Lin X, Yin P, Yin Y, Zhou M, Qi J, et al. Assessment of disability weights at the provincial and city levels based on 93,254 respondents in Fujian, China: Findings from the Fujian disability weight measurement study. Chinese Medical Journal. 2024;137: 1375–1377. pmid:37612264
  27. 27. Xie X-X, Zhou W-M, Lin F, Li X-Q, Zhong W-L, Lin S-G, et al. Ischemic heart disease deaths, disability-adjusted life years and risk factors in Fujian, China during 1990–2013: Data from the Global Burden of Disease Study 2013. International Journal of Cardiology. 2016;214: 265–269. pmid:27077547
  28. 28. Hu X, Fang X, Wu M. Prevalence, awareness, treatment and control of type 2 diabetes in southeast China: A population‐based study. J of Diabetes Invest. 2024; jdi.14213. pmid:38741389
  29. 29. Meng L, Zhao D, Pan Y, Ding W, Wei Q, Li H, et al. Validation of Omron HBP-1300 professional blood pressure monitor based on auscultation in children and adults. BMC Cardiovasc Disord. 2016;16: 9. pmid:26758197
  30. 30. Zhou X, Pang Z, Gao W, Wang S, Zhang L, Ning F, et al. Performance of an A1C and Fasting Capillary Blood Glucose Test for Screening Newly Diagnosed Diabetes and Pre-Diabetes Defined by an Oral Glucose Tolerance Test in Qingdao, China. Diabetes Care. 2010;33: 545–550. pmid:20007941
  31. 31. Bartoli E, Fra GP, Schianca GPC. The oral glucose tolerance test (OGTT) revisited. Eur J Intern Med. 2011;22: 8–12. pmid:21238885
  32. 32. Wang Y, Liao R, Feng XL. Equity in Essential Maternal, Newborn, and Child Health Interventions in Northeastern China, 2008 to 2018. Front Public Health. 2020;8: 212. pmid:32714887
  33. 33. Xie Y, Zhou X. Income inequality in today’s China. Proc Natl Acad Sci U S A. 2014;111: 6928–6933. pmid:24778237
  34. 34. Wei Liu, Xue Wang, Shi . Comparisons of Visceral Adiposity Index, Body Shape Index, Body Mass Index and Waist Circumference and Their Associations with Diabetes Mellitus in Adults. Nutrients. 2019;11: 1580. pmid:31336951
  35. 35. Consensus Conference Panel:, Watson NF, Badr MS, Belenky G, Bliwise DL, Buxton OM, et al. Joint Consensus Statement of the American Academy of Sleep Medicine and Sleep Research Society on the Recommended Amount of Sleep for a Healthy Adult: Methodology and Discussion. J Clin Sleep Med. 2015;11: 931–952. pmid:26235159
  36. 36. Buysse DJ, Reynolds CF, Monk TH, Berman SR, Kupfer DJ. The Pittsburgh sleep quality index: A new instrument for psychiatric practice and research. Psychiatry Research. 1989;28: 193–213. pmid:2748771
  37. 37. Kim JS, Han JW, Oh DJ, Suh SW, Kwon MJ, Park J, et al. Effects of sleep quality on diurnal variation of brain volume in older adults: A retrospective cross-sectional study. NeuroImage. 2024;288: 120533. pmid:38340880
  38. 38. Babu Henry Samuel I, Pollin KU, Breneman CB. Lower cortical volume is associated with poor sleep quality after traumatic brain injury. Brain Imaging and Behavior. 2022;16: 1362–1371. pmid:35018551
  39. 39. Lee PH, Macfarlane DJ, Lam T, Stewart SM. Validity of the international physical activity questionnaire short form (IPAQ-SF): A systematic review. Int J Behav Nutr Phy. 2011;8: 115. pmid:22018588
  40. 40. Tremblay MS, Aubert S, Barnes JD, Saunders TJ, Carson V, Latimer-Cheung AE, et al. Sedentary Behavior Research Network (SBRN)–Terminology Consensus Project process and outcome. Int J Behav Nutr Phy. 2017;14: 75. pmid:28599680
  41. 41. Hu D, Zhang B, Huang M, Liu M, Xia X, Zuo Y, et al. Evaluation of a medical education policy with compulsory rural service in China. Front Public Health. 2023;11: 1042898. pmid:36817880
  42. 42. Tang S, Xu Y, Li Z, Yang T, Qian D. Does Economic Support Have an Impact on the Health Status of Elderly Patients With Chronic Diseases in China?—Based on CHARLS (2018) Data Research. Front Public Health. 2021;9: 658830. pmid:33959585
  43. 43. Meng X , D’Arcy C. Education and Dementia in the Context of the Cognitive Reserve Hypothesis: A Systematic Review with Meta-Analyses and Qualitative Analyses. Laks J, editor. PLoS ONE. 2012;7: e38268. pmid:22675535
  44. 44. Li X, Zhang Y, Zhang C, Zheng Y, Liu R, Xiao S. Education counteracts the genetic risk of Alzheimer’s disease without an interaction effect. Front Public Health. 2023;11: 1178017. pmid:37663829
  45. 45. Kuule Y, Dobson AE, Woldeyohannes D, Zolfo M, Najjemba R, Edwin BMR, et al. Community Health Volunteers in Primary Healthcare in Rural Uganda: Factors Influencing Performance. Front Public Health. 2017;5. pmid:28424765
  46. 46. Ma H, Yu G, Wang Z, Zhou P, Lv W. Association between dysglycemia and mortality by diabetes status and risk factors of dysglycemia in critically ill patients: a retrospective study. Acta Diabetol. 2022;59: 461–470. pmid:34761326
  47. 47. Gerstein HC, Santaguida P, Raina P, Morrison KM, Balion C, Hunt D, et al. Annual incidence and relative risk of diabetes in people with various categories of dysglycemia: A systematic overview and meta-analysis of prospective studies. Diabetes Res Clin Pract. 2007;78: 305–312. pmid:17601626
  48. 48. Lee S-J, Chandrasekran P, Mazucanti CH, O’Connell JF, Egan JM, Kim Y. Dietary curcumin restores insulin homeostasis in diet-induced obese aged mice. Aging (milano). 2022;14: 225–239. pmid:35017319
  49. 49. Kautzky-Willer A, Harreiter J, Pacini G. Sex and Gender Differences in Risk, Pathophysiology and Complications of Type 2 Diabetes Mellitus. Endocr Rev. 2016;37: 278–316. pmid:27159875
  50. 50. Hannon TS, Arslanian SA. The changing face of diabetes in youth: lessons learned from studies of type 2 diabetes. Ann New York Acad Sci. 2015;1353: 113–137. pmid:26448515
  51. 51. Schofield JD, Liu Y, Rao-Balakrishna P, Malik RA, Soran H. Diabetes Dyslipidemia. Diabetes Therapy. 2016;7: 203–219. pmid:27056202
  52. 52. Chehade JM, Gladysz M, Mooradian AD. Dyslipidemia in Type 2 Diabetes: Prevalence, Pathophysiology, and Management. Drugs. 2013;73: 327–339. pmid:23479408
  53. 53. Handelsman Y, Butler J, Bakris GL, DeFronzo RA, Fonarow GC, Green JB, et al. Early intervention and intensive management of patients with diabetes, cardiorenal, and metabolic diseases. Journal of Diabetes and its Complications. 2023;37: 108389. pmid:36669322
  54. 54. Skyler JS, Bakris GL, Bonifacio E, Darsow T, Eckel RH, Groop L, et al. Differentiation of Diabetes by Pathophysiology, Natural History, and Prognosis. Diabetes. 2017;66: 241–255. pmid:27980006
  55. 55. Liu Y. China’s public health-care system: facing the challenges. Bull World Health Organ. 2004;82: 532–538. pmid:15500285
  56. 56. Fekadu G, Bula K, Bayisa G, Turi E, Tolossa T, Kebebe H. Challenges And Factors Associated With Poor Glycemic Control Among Type 2 Diabetes Mellitus Patients At Nekemte Referral Hospital, Western Ethiopia. Journal of Multidisciplinary Healthcare. 2019;Volume 12: 963–974. pmid:31819470
  57. 57. Wormgoor SG, Dalleck LC, Zinn C, Harris NK. Effects of High-Intensity Interval Training on People Living with Type 2 Diabetes: A Narrative Review. Canadian Journal of Diabetes. 2017;41: 536–547. pmid:28366674
  58. 58. He D, Sun N, Xiong S, Qiao Y, Ke C, Shen Y. Association between the proportions of carbohydrate and fat intake and hypertension risk: findings from the China Health and Nutrition Survey. Journal of Hypertension. 2021;39: 1386–1392. pmid:33534340
  59. 59. Wan S, Pan D, Su M, Wang S, Wang Y, Xu D, et al. Association between socio-demographic factors, lifestyle, eating habits and hypertension risk among middle-aged and older rural Chinese adults. Nutrition, Metabolism and Cardiovascular Diseases. 2024;34: 726–737. pmid:38161126
  60. 60. Zhang Q, Huang H, Li J, Niu Y, Sun P, Cheng F. Knowledge, attitudes and practices of patients with chronic pharyngitis toward laryngopharyngeal reflux in Suzhou, China. BMC Public Health. 2023;23: 2542. pmid:38115020
  61. 61. Yao Q, Zhang X, Wu Y, Liu C. Decomposing income-related inequality in health-related quality of life in mainland China: a national cross-sectional study. BMJ Global Health. 2023;8: e013350. pmid:38035731
  62. 62. Von Salmuth V, Buijs L, Chirangi B, Vreugdenhil AC, Van Schayck OC. Health needs assessment for the double burden of malnutrition: a community-based study on nutrition facilitators and barriers in rural Tanzania. Public Health Nutr. 2023;26: 2450–2459. pmid:37581236
  63. 63. Johnston JL, Fanzo JC, Cogill B. Understanding Sustainable Diets: A Descriptive Analysis of the Determinants and Processes That Influence Diets and Their Impact on Health, Food Security, and Environmental Sustainability. Advances in Nutrition. 2014;5: 418–429. pmid:25022991
  64. 64. Yaggi HK, Araujo AB, McKinlay JB. Sleep Duration as a Risk Factor for the Development of Type 2 Diabetes. Diabetes Care. 2006;29: 657–661. pmid:16505522
  65. 65. Gottlieb DJ, Punjabi NM, Newman AB, Resnick HE, Redline S, Baldwin CM, et al. Association of Sleep Time With Diabetes Mellitus and Impaired Glucose Tolerance. Arch Intern Med. 2005;165: 863. pmid:15851636
  66. 66. Wojeck BS, Inzucchi SE, Qin L, Yaggi HK. Polysomnographic predictors of incident diabetes and pre-diabetes: an analysis of the DREAM study. J Clin Sleep Med. 2023;19: 703–710. pmid:36689314
  67. 67. Subramanian A, Adderley NJ, Tracy A, Taverner T, Hanif W, Toulis KA, et al. Risk of Incident Obstructive Sleep Apnea Among Patients With Type 2 Diabetes. Diabetes Care. 2019;42: 954–963. pmid:30862657
  68. 68. Saparwan N, Tohit NM, Salmiah MS. A cross-sectional study on the sleep quality among type 2 diabetes mellitus patients and its associated factors. Med J Malaysia. 2023;78: 627–634. pmid:37775490
  69. 69. Anothaisintawee T, Reutrakul S, Van Cauter E, Thakkinstian A. Sleep disturbances compared to traditional risk factors for diabetes development: Systematic review and meta-analysis. Sleep Med Rev. 2016;30: 11–24. pmid:26687279
  70. 70. Sabir AA, Isezuo SA, Ohwovoriole AE. Dysglycaemia and its risk factors in an urban Fulani population of northern Nigeria. West Afr J Med. 2011;30: 325–330. pmid:22752819
  71. 71. Yang W, Lu J, Weng J, Jia W, Ji L, Xiao J, et al. Prevalence of Diabetes among Men and Women in China. N Engl J Med. 2010;362: 1090–1101. pmid:20335585
  72. 72. Xu Y. Prevalence and Control of Diabetes in Chinese Adults. Jama. 2013;310: 948. pmid:24002281
  73. 73. Barua L, Faruque M, Chowdhury HA, Banik PC, Ali L. Health‐related quality of life and its predictors among the type 2 diabetes population of Bangladesh: A nation‐wide cross‐sectional study. J of Diabetes Invest. 2021;12: 277–285. pmid:32564501
  74. 74. Mata J, Richter D, Schneider T, Hertwig R. How cohabitation, marriage, separation, and divorce influence BMI: A prospective panel study. Health Psychology. 2018;37: 948–958. pmid:30234354
  75. 75. Werneck AO, Winpenny EM, Foubister C, Guagliano JM, Monnickendam AG, Van Sluijs EMF, et al. Cohabitation and marriage during the transition between adolescence and emerging adulthood: A systematic review of changes in weight-related outcomes, diet and physical activity. Preventive Medicine Reports. 2020;20: 101261. pmid:33344148
  76. 76. Rapp I, Schneider B. The impacts of marriage, cohabitation and dating relationships on weekly self-reported physical activity in Germany: A 19-year longitudinal study. Social Science & Medicine. 2013;98: 197–203. pmid:24331899
  77. 77. Nomaguchi KM, Bianchi SM. Exercise Time: Gender Differences in the Effects of Marriage, Parenthood, and Employment. J of Marriage and Family. 2004;66: 413–430.
  78. 78. Liang X, Tang X, Xi B, Qu P, Ren Y, Hao G. Abdominal obesity-related lipid metabolites may mediate the association between obesity and glucose dysregulation. Pediatr Res. 2023;93: 183–188. pmid:35437306
  79. 79. Zhong P, Tan S, Zhu Z, Zhu Z, Liang Y, Huang W, et al. Normal‐weight central obesity and risk of cardiovascular and microvascular events in adults with prediabetes or diabetes: Chinese and British cohorts. Diabetes Metab Res Rev. 2023;39: e3707. pmid:37525502
  80. 80. Wang L, Zhou B, Zhao Z, Yang L, Zhang M, Jiang Y, et al. Body-mass index and obesity in urban and rural China: findings from consecutive nationally representative surveys during 2004–18. Lancet. 2021;398: 53–63. pmid:34217401
  81. 81. Zhou L, Cao D, Si Y, Zhu X, Du L, Zhang Y, et al. Income-related inequities of adult obesity and central obesity in China: evidence from the China Health and Nutrition Survey 1997–2011. BMJ Open. 2020;10: e034288. pmid:33127627