Translate this page into:
Social and rural determinants of glycemic control in type 2 diabetes mellitus
For correspondence: Dr Ramya Raghavan, Department of Life Sciences, Sri Sathya Sai University for Human Excellence, Kalaburagi 585 313, Karnataka, India e-mail: ramya.r@sssuhe.ac.in
-
Received: ,
Accepted: ,
How to cite this article: Ananthakumar B, Raghavan R. Social and rural determinants of glycemic control in type 2 diabetes mellitus. Indian J Med Res. 2026;163:231-6. doi: 10.25259/IJMR_1884_2025
Abstract
Background and objectives
The global epidemic of type 2 diabetes mellitus (T2DM) is disproportionately severe in rural and agricultural populations, driven by the complex interplay of socioeconomic, behavioural, and genetic risk factors. While these determinants are acknowledged individually, few studies have validated a parsimonious, integrated measurement model to effectively structure and assess their collective influence on glycaemic control within a rural primary care context. We aimed to address this critical gap.
Methods
We conducted a cross-sectional, observational study utilising a purposive sample of 1,011 patients with uncontrolled T2DM recruited from a rural diabetic healthcare setting. To empirically test the structural relationships among key risk factors, we employed a confirmatory factor analysis (CFA) to validate a four-factor measurement model. This model integrated 14 observed indicators across the latent constructs of alcohol addiction, economic conditions, dietary habits, and family history of T2DM.
Results
The proposed four-factor model demonstrated an excellent practical fit to the observed data, confirming its structural validity. All factor loadings were statistically significant (P<0.001). The magnitude of practical misfit was medium (Cohen’s w=0.37), validating the model’s overall explanatory power. Factor intercorrelations revealed a significant positive correlation between alcohol addiction and economic condition (r=0.109, P=0.006), and a significant negative relationship between diet and family history of T2DM (r=-0.154, P=0.014).
Interpretation and conclusions
Our findings establish alcohol addiction, economic conditions, dietary habits, and family history as robust, interrelated determinants of glycaemic control in rural primary care. The validated structural model provides an evidence-based tool for risk stratification and personalised intervention.
Keywords
Addiction
Exposomics
Health equity
Primary health care
Rural health
In rural areas, geographical isolation, limited resources, and local culture can make social determinants of health more complex.1,2 While research has begun to explore how rural social factors affect type 2 diabetes mellitus (T2DM), multifactorial assessment-based risk models are lacking to show links between education, tobacco use, work status, socioeconomic position, and diabetes.3 Environmental aspects such as food options, walkability, and green spaces may also play a role, suggesting rural communities face unique risks.4 Challenges like poverty, limited healthcare, and rural living may add to diabetes risk.5 Our study questions traditional views on diabetes and highlights the roles of addiction and lifestyle.
The term ‘rural exposomics’ refers to the comprehensive study of environmental exposures in rural settings and how they relate to health and disease. Past research has looked at socioeconomic status and lifestyle but often shows family history as the independent predictor of T2DM.6 Lower socioeconomic position places individuals in a less healthy residential environment and increases the odds of exposure to financial strain and food insecurity. In rural areas, addictions can affect both physical and mental health, raising diabetes risk.7 Economic hardship can delay medical care in rural settings, creating a risk profile distinct from that in urban areas.8
We hypothesise that in close rural communities, genetic, environmental, and lifestyle risks are interconnected. The conceptual framework is that T2DM risk arises from the synergistic interaction between non-modifiable biological factors (family history) and modifiable behaviours and environments (economic conditions, dietary patterns, and addiction).9 No previous large-scale analysis has examined T2DM risk stratification in rural populations.10,11
We assert that the intergenerational transmission of genetic predispositions, environmental factors, and lifestyle risks significantly compounds the risk of T2DM in rural families. Therefore, the objective of this study was to identify and validate the underlying factors influencing glycaemic control in patients with type 2 diabetes mellitus using a confirmatory factor analysis. Here, we elucidate the cumulative effects of combining lifestyle factors with genetic susceptibility, economic hardship, poor diet, and limited healthcare access.
Methods
This cross-sectional study was conducted at the department of Endocrinology outpatient, Sri Sathya Sai University for Human Excellence, Karnataka, India after obtaining the ethical clearance from the Institute Ethics Committee.
Study design and setting
The study design was a self-reported survey conducted from July 2024 to October 2024. A cross-sectional, observational study using a purposive sample of 1,011 patients with uncontrolled T2DM was conducted in a rural primary care setting. Data were collected via a self-administered survey instrument. A confirmatory factor analysis (CFA) assessed the construct validity and dimensionality of the survey instrument. Eligible participants who visited the rural hospital’s outpatient department were provided with detailed information about the research and subsequently invited to provide written informed consent. The survey was administered individually in a private setting to ensure confidentiality.
Sampling and study population
To investigate rural perspectives of T2DM patients, the study population was selected by purposive sampling from those attending the diabetic clinic for routine medical care. The inclusion criteria were rural residents over the age of 18, with a confirmed diagnosis of type 2 diabetes mellitus. Survey participants needed to be capable of providing informed consent and able to complete the 14-item survey instrument.
Exclusion criteria
Patients with other types of diabetes (type 1 diabetes, gestational diabetes), with significant comorbidities (end-stage renal disease, severe psychiatric disorders), and unwilling or unable to provide informed consent were excluded.
Sample size
To ensure a minimum sample-to-indicator variables ratio of 10:1, we employed a large sample size (n=1011). Sample size was chosen to achieve adequate CFA model fit index performance and to yield smaller standard errors for factor loadings, error variances, and factor covariances. Hoelter’s critical N measures the smallest sample size required to accept the model (alpha=.05: 665), (alpha=.01: 737) at a specific alpha level. The actual sample size was kept higher than the minimum required to achieve reliable factor loadings and adequate statistical power for the hypothesised model.
Questionnaire development and validation
The questionnaire was prepared through a systematic multistep process to ensure content validity, culturally appropriate. A panel of Sri Madhusudan Sai Institute of Medical Sciences and Research public health professionals and endocrinologists reviewed the questionnaire. The survey instrument was translated from English to the local language and back translated by volunteers to ensure uniform content and vocabulary. Feedback from the panel guided revisions in wording and contextual relevance for rural populations. A pilot test was conducted with a small group of T2DM patients to assess the clarity of items and the appropriateness of responses. The rigorous process ensured the questionnaire was suitable for administration among adults with T2DM.
Survey instrument
The main instrument, the survey questionnaire, was administered to those previously diagnosed with type 2 diabetes. It included statements on various indicators, each evaluated on a 5-point Likert scale. Data were collected through face-to-face interviews using structured questionnaires administered by trained nursing assistants ( Fig. 1).
- The flow diagram illustrates methodology of the study in sequential process of identification, screening, eligibility assessment, and data analysis. Details of study design and methodology of sample selection with Inclusion and exclusion criteria, sampling frame, and recruitment process are shown.
Data analysis
Data analysis included descriptive statistics, chi-square tests, and confirmatory factor analysis, performed using R programming.12 CFA was conducted to test the proposed four-factor model. Data suitability was assessed using the Kaiser-Meyer-Olkin (KMO) measure and Bartlett’s test of sphericity. Model fit was evaluated using standard indices, including the comparative fit index (CFI), Tucker-Lewis index (TLI), root mean square error of approximation (RMSEA), and standardised root mean square residual (SRMR). Standardised factor loadings, R-squared values, and latent factor correlations were examined to assess practical clinical significance and discriminant validity. The statistical CFA analyses were conducted using the open-source JASP software, based on R.12 CFA model was graphically generated using R script manually using Diagramme R package with actual data.
Results
Demographics
The demographics of the 1,011 survey participants are summarised in Table. The participants’ ages ranged from 29 to 89 yr. Out of which all of them are natives of nearby villages or those who have lived for more than 30-40 yr. Family history such as cerebrovascular accident/stroke (CVA), dyslipidaemia (cholesterol), ischaemic heart disease (IHD)/CAD, kidney disease, and peripheral vascular disease (PVD), were reported in smaller percentages when combined with diabetes and hypertension. Family history of T2DM with/without complications is shown in Supplementary Table I.
| Variable | Count (%) |
|---|---|
| Age group (yr) | |
| 29 to 39 | 41 (4.1) |
| 40 to 49 | 177 (17.5) |
| 50 to 59 | 305 (30.2) |
| 60 to 69 | 313 (31.0) |
| 70 to 79 | 154 (15.2) |
| 80 to 89 | 21 (2.1) |
| Medical history (associated disease complications) | |
| Diabetes | 323 (32.0) |
| Diabetes, hypertension | 579 (57.3) |
| Diabetes, hypertension, cerebrovascular accident/stroke | 17 (1.7) |
| Diabetes, hypertension, dyslipidaemia | 23 (2.3) |
| Diabetes, hypertension, ischemic heart disease coronary artery disease | 47 (4.6) |
| Diabetes, hypertension, kidney | 16 (1.6) |
| Diabetes, hypertension, peripheral vascular disease | 6 (0) |
| Gender | |
| Female | 496 (49.1) |
| Male | 510 (50) |
| Prefer not to say | 5 (0.5) |
| Diet | |
| Mixed diet | 844 (83.5) |
| Vegetarian | 146 (14.4) |
| Vegetarian with egg | 21 (2.1) |
Descriptive statistics of the Likert scale responses for all survey items are presented in the Supplementary Table II.
Assumption checks for factor analysis
To help determine if the factor analysis is appropriate, the Kaiser-Meyer-Olkin (KMO) yielded values above 0.5, with most indicators demonstrating adequate to good sampling adequacy. The Bartlett’s test of sphericity (χ2=2662, df=91, P<0.001) validated that factor analysis is appropriate.13 The Cronbach’s alpha for survey items was above 0.7, indicating good internal consistency. Using principal component analysis, four latent factors were identified, the factor loadings for items less than <0.50 being suppressed. Principal component analysis determined the number of factors; a scree plot of the eigen values was computed. The number of factor latent variables was set at four based on the Scree plot. The new indicator variables based on the loading were then tested by confirmatory factor analysis.
Model specification and assumptions
The confirmatory factor analysis (CFA) validated a four-factor measurement model examining relationships between alcohol addiction, economic conditions, dietary habits, and family history. The analysis included 14 observed indicators distributed across the four latent constructs, with 1011 participants providing complete data.
CFA model fit assessment
The model fit indices assessed by Chi-square were significant (P<0.001), and Cohen’s w value for the factor model was 0.372, which is classified as a medium effect size.
The comparative fit index (CFI=0.973), Goodness of fit index (GFI=0.981), and Tucker-Lewis Index (TLI=0.966) were noted. In addition, fit measures RMSEA and SRMR were less than 0.08. The CFI and GFI indices, with values close to or exceeding 0.95, suggest a good fit as prescribed.14 RMSEA values of less than 0.05 indicate a good fit, but values as high as 0.08 indicate a realistic fit. These indices collectively suggest that the proposed four-factor model provides a strong representation of the relationships within the data. CFA model fit assessment was within satisfactory ranges, and the four latent factors obtained are validated.
Factor loadings and intercorrelations
All factor loadings were statistically significant (P<0.001), indicating that each indicator makes a significant contribution to its respective latent factor.
A statistically significant positive correlation was observed between alcohol addiction and economic condition [r=0.109, SE=0.039, z=2.749, P=0.006, 95% Confidence interval (CI) 0.031, 0.186]. Data suggest addiction could be a mediator in the relationship between economic conditions and T2DM risk. The statistically significant indicator items related to lifestyle habits, defensiveness about criticism, and feelings of guilt are strong and reliable measures. The negative loading for ‘guilty about habits’ is conceptually sound, implying that as guilt increases, the underlying construct of addiction may decrease, or vice versa. A significant negative correlation emerged between diet and T2DM family history (r=-0.154, SE=0.063, z=-2.447, P=0.014, 95% CI -0.277, -0.031), indicating that poor dietary patterns are associated with higher family history of T2DM.
The statistical analysis strongly supports the model, demonstrating that the underlying latent factors are highly distinct and reliable measures. The discriminant validity is confirmed by the non-significant or very small correlations between the four latent constructs (Supplementary Table III). The 14-item survey measures four distinct rural factors, all of which are shown to increase the risk of T2DM. The economic factors showed greater influence on lifestyle choices ( Fig. 2). Furthermore, the diet and family history are more complex factors with varied effects. The statistically significant residual covariances indicates that some shared variation remains unexplained by the measured factors (Supplementary Table IV). It implies that several other unreported variables affect rural health outcomes that warrant future investigation.
- CFA model demonstrates strong empirical support with strong discriminant validity among the four latent factors. CFA figure displays the factor structure and standardised parameter estimates from the survey of the purposive sample of N=1,011 patients with uncontrolled Type 2 Diabetes Mellitus. Ovals represent the latent factors, and rectangles represent the observed variables. The blue arrows indicate the standardised factor loadings (beta), strength of the relationship between the latent factor and the observed item. The double arrows dotted green lines between latent factors indicate the factor covariances (correlations) between the constructs.
Discussion
Our study revealed that health outcomes among rural communities are created by a unique set of interacting risk factors. We report rural socioeconomic conditions, alcohol consumption, and dietary factors together to form a distinct risk profile for rural healthcare. Our research found that alcohol-related factors combined with low socioeconomic position and economic instability affect blood sugar control in patients. Addiction habits act not merely as an independent risk factor for T2DM but also complement the diabetogenic effects of socioeconomic and biological determinants in rural populations. The overall framework supports the gene–environment interaction. The social determinants of health, such as addiction, financial strain, are linked to negative health behaviours, underscoring the need for policy action to promote health equity.15
Previous studies also show that alcohol addiction often co-occurs with poor diet, obesity, and economic deprivation, and our data found in rural settings that these factors as associated with risks for T2DM.16 Limited access to healthcare, lower educational attainment, and neighbourhood deprivation further exacerbate T2DM risk.4 It found that financial hardships, particularly difficulty paying bills, were consistently linked to worsening glycaemic control. Additionally, financial strain is associated with poorer mental, physical, and functional health. The moderate loadings across family history indicators support the genetic and familial clustering of diabetes risk. The significant negative correlation between diet and T2DM family history (r=-0.154) supports that poor dietary patterns may be more prevalent in families with diabetes history. The results for neighbourhood type suggest that other unknown factors, possibly related to rural economics, may be at play.4 Earlier reports show a relationship between social determinants of health and glycaemic control in elderly adults with diabetes.5,15 Our findings align with the literature on social determinants of health and lifestyle choices within rural families in the development of diabetes complications.17 While the model shows associations between items and factors, it cannot explain how these relationships develop over time. To confirm these findings and understand the causes, more experimental and clinical studies are needed.
This study has several limitations. First, the weaknesses associated with the ‘diet’ and ‘T2DM history’ factors suggest that incorporating additional indicators could improve their reliability. Second, since the data is cross-sectional, we cannot infer causality from the findings. Testing the model across different groups would help establish the model’s validity in different cultural or demographic contexts, applicable.10,11
Our study has contributed to the direct impact of predictable factors and the mediating or moderating variables for rural-specific populations. In summary, the study’s findings support the hypothesis and suggest that rural socioeconomic conditions and lifestyle choices are linked to T2DM outcomes. T2DM prevalence and complications in rural populations are limited by socioeconomic and psychosocial factors. The public health policy of diabetes management should shift toward multifactorial interventions tailored to the rural context.18 In a clinical scenario, these findings mean that a healthcare provider treating a rural patient with poorly controlled T2DM should not only focus on prescribing medication but must also screen for and integrate referrals addressing economic instability and substance use.
Acknowledgment
Authors acknowledge Sadguru Sri Madhusudan Sai, Chancellor Sri Sathya Sai University For Human Excellence (SSSUHE), India, and Prashanti Balamandira Trust, India.
Author contributions
Both authors have equally made the significant contributions to the work’s conception, design, data acquisition, analysis, or interpretation. AB: Conceptualisation, methodology, data analysis, manuscript writing; RR: Data curation, drafting or critically revising data analysis, manuscript writing. All the authors have read and confirm the final printed version of the manuscript.
Financial support and sponsorship
None.
Conflicts of Interest
None.
Use of Artificial Intelligence (AI)-Assisted Technology for manuscript preparation
The authors confirm that there was no use of AI-assisted technology for assisting in the writing of the manuscript and no images were manipulated using AI.
References
- Social determinants of health and diabetes: A scientific review. Diabetes Care.. 2020;44:258-79.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Community credit scores and community socioeconomic deprivation in association with type 2 diabetes across an urban to rural spectrum in Pennsylvania: A case–control study. BMJ Public Health.. 2024;2:e000744.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Social determinants of health and type 2 diabetes in Asia. J Diabetes Investig.. 2025;16:971-83.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- The mediating role of the food environment, greenspace, and walkability in the association between socioeconomic position and type 2 diabetes — The Maastricht study. Diabetes Metab Syndr.. 2024;18:103155.
- [CrossRef] [PubMed] [Google Scholar]
- Do gene–environment interactions have implications for the precision prevention of type 2 diabetes? Diabetologia. 2022. ;65:1804-13.
- [Google Scholar]
- Integrating the exposome and one health approach to national health surveillance: An opportunity for Latin American countries in health preventive management. Front Public Health.. 2024;12:1376609.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Exploring the association of social connections and food security among adults with uncontrolled type 2 diabetes: A population-based study. J Health Popul Nutr.. 2024;43:156.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Predictors of diabetes risk in urban and rural areas in Colombia. Heliyon.. 2021;8:e08653.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Exploring influencing factors of healthy lifestyles in rural area among older adults with diabetes based on socioecological model. Sci Rep.. 2025;15:2829.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Exploration of the individual, social and environmental factors influencing dietary behaviour in shift workers with type 2 diabetes working in UK healthcare—The shift‐diabetes study: A qualitative study using the theoretical domains framework. Diabet Med.. 2024;41:e15179.
- [CrossRef] [PubMed] [Google Scholar]
- Precision treatment of beta-cell monogenic diabetes: A systematic review. Commun Med (Lond).. 2024;4:145.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- JASP Team. JASP (Version 0.19.3) [Computer software]. Available from: https://jasp-stats.org/, accessed on July 1, 2025.
- When to use and how to report the results of PLS-SEM. European Business Review.. 2019;31:2-24.
- [CrossRef] [Google Scholar]
- CB-SEM vs PLS-SEM methods for research in social sciences and technology forecasting. Technol Forecast Soc Change.. 2021;173:121092-24.
- [CrossRef] [Google Scholar]
- Financial strain measures and associations with adult health: A systematic literature review. Soc Sci Med.. 2025;364:117531.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- A causal relationship between alcohol intake and type 2 diabetes mellitus: A two-sample Mendelian randomization study. Nutr Metab Cardiovasc Dis.. 2022;32:2865-76.
- [CrossRef] [PubMed] [Google Scholar]
- A composite indicator for primary diabetes care: A cross-sectional study in Hungary. Healthcare (Basel).. 2025;13:480.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Social determinants and health disparities pertaining to diabetes in appalachia. J Prim Care Community Health.. 2023;14:21501319231192327.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]