Translate this page into:
FADIS: Fast-food attributed diabetes index study: An ecological State-level exploration of nutritional transitions and diabetes burden in India
Present address: †Department of Endocrinology, and #Department of Hematology & Oncology, University of Alabama, Birmingham, USA
For correspondence: Dr Jeevan Yadav Nammi, Department of Endocrinology, University of Alabama, Birmingham, United States e-mail: jeevannammi1010@gmail.com
-
Received: ,
Accepted: ,
How to cite this article: Nammi JY, Pasala R. FADIS: Fast-food attributed diabetes index study: An ecological State-level exploration of nutritional transitions and diabetes burden in India. Indian J Med Res. 2026;163:237-42. doi: 10.25259/IJMR_1802_2025.
Abstract
Background and objectives
India’s rapid urbanisation, economic liberalisation, and rising fast food consumption are driving a nutritional transition that parallels the growing epidemic of non communicable diseases, particularly type 2 diabetes mellitus. This study examines the association between fast food expenditure and diabetes prevalence across 18 Indian States, stratified by gender and urban-rural residence.
Methods
The fast-food attributed diabetes index study (FADIS) utilised publicly available State-level data to evaluate correlations between fast food spending and diabetes prevalence. A novel framework, DIAGRAM (diabetes and intake gradient analysis model), was applied to parse these associations across urban and rural men and women. The WEIGHT (women’s elevated intake and glucose-health evaluation tracker) model assessed the prevalence of overweight as a predictor of diabetes in women.
Results
We found a strong positive correlation between fast food expenditure and diabetes prevalence among urban populations, specifically urban men (r=0.653, P=0.003) and urban women (r=0.619, P=0.0062). Rural patterns were less consistent and weaker. The WEIGHT model identified the prevalence of overweight as a significant contributor to diabetes risk among women.
Interpretation and conclusions
In this ecological analysis, we found that fast-food expenditure had a strong correlation with diabetes prevalence, particularly among urban men and women. In women, overweight prevalence rather than fast-food expenditure emerged as the significant predictor in multivariable analyses. These findings highlight the need for gender-sensitive and region-specific public health strategies to address India’s evolving nutritional transitions and diabetes burden.
Keywords
Diabetes mellitus
Fast food expenditures
Feeding behaviour
India
Public health
Statistics and numerical data
India is navigating a complex nutritional transition, shifting from traditional, home-prepared diets to industrially prepared and energy-dense foods.1 As this transition accelerates, the prevalence of noncommunicable diseases2 such as diabetes mellitus, has sharply risen.3-5 In urban environments, especially, accessibility to fast foods, defined as industrially prepared, energy-dense, ready-to-eat or ready-to-heat meals, snacks, and beverages typically high in refined carbohydrates, added sugars, unhealthy fats, and sodium, has increased dramatically over the past two decades.1,6,7 Evidence from extensive cohort studies shows that frequent fast-food intake is strongly linked with diabetes incidence.8 U.S. cohorts found a twofold higher risk of type 2 diabetes among individuals consuming fast food ≥2 times per week. In comparison, the Singapore Chinese Health Study reported a 27% higher diabetes risk with similar intake.9 A meta-analysis of global studies further estimated a 19% increase in diabetes risk for each additional fast-food serving per week.10 In India, the ICMR–INDIAB study reported that 45% of urban diabetics frequently consumed fried snacks compared to 25% of non-diabetics.11
In India, the rise in fast-food consumption is closely linked to increased household income and expenditure on processed and convenience foods, especially among urban middle-class populations.12 This trend reflects broader socio-economic and lifestyle transitions, including changes in work culture, such as longer working hours, dual-income households, and reduced time for home cooking.2,13
The FADIS (fast food attributed diabetes index study) project was designed to explore these dynamics quantitatively by analysing State-level data on household fast-food expenditure and diabetes prevalence. The study aims to examine how fast-food spending correlates with diabetes prevalence across 18 Indian States, stratified by gender and urban–rural residence.
Methods
This cross sectional study was undertaken by the department of Medicine, Mayo Clinic, Rochester, USA and the waiver for ethical clearance was obtained from the Mayo Clinic Institutional Review Board (IRB) for this study.
Study design and framework
This is a cross-sectional ecological study analysing State-level data from 18 Indian States to examine the relationship between fast-food expenditure (as a proportion of total monthly per capita food spending) and diabetes prevalence by gender. The study was done in adherence to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines to ensure transparent and standardised reporting14.
The diabetes and intake gradient analysis model (DIAGRAM) was developed to measure the gradient of disease prevalence in relation to fast-food spending levels across different States. The model segments data by gender and residence type (urban versus rural), enabling refined comparisons that reflect India’s heterogeneous food environment.
Data selection and variables
The study was conducted in May 2025, and data were drawn from two nationally representative sources. Diabetes prevalence and related health indicators were obtained from the fifth round of the National Family Health Survey (NFHS-5, 2019–2021)15, which provided State-wise, gender- and region-specific data on self-reported diabetes, mean body mass index (BMI), overweight and obesity prevalence, frequency of fried food consumption, and median years of education. Complementary expenditure data were extracted from the National Sample Survey (NSS, 2022–2023) by the Ministry of Statistics and Programme Implementation (MoSPI)16, specifically from household-level information on monthly spending (original survey questionnaire files provided as Supplementary Files). For this analysis, we calculated State-wise urban and rural monthly per-capita expenditure on fast foods, defined in accordance with the ICMR-NIN 2024 classification of ultra-processed foods (UPFs), which include items such as commercially produced breads, packaged snacks, sugar sweetened beverages, breakfast cereals, ready-to-eat meals, bakery products, sauces, and packaged spreads; but exclude minimally processed staples like rice, plain dairy ingredients, fiber supplements, and 100% fruit juices. A detailed list of included food items and beverage categories is provided in Supplementary Table. Given that our study includes individuals with known diabetes, who may have modified their dietary habits (e.g., avoiding added sugars or sweet snacks), this table enhances transparency regarding the expenditure items captured.
States were included in the final analysis only if complete data were available for all key variables across gender and region categories. Variables were cleaned and harmonised across datasets, and the final merged data were analysed in R.
The outcome variable, diabetes prevalence, was based on self-reported physician diagnosis as recorded in the NSS 2022–2023 survey. While our study hypothesis focuses on dietary risk factors, we recognise that prevalent cases may have altered their consumption patterns post-diagnosis, which can introduce reverse causality. The primary independent variable was the proportion of monthly per capita food expenditure spent on fast foods. Fast-food expenditure was calculated as the percentage of total monthly per capita food expenditure spent on fast-food items. Additional covariates included the percentage of individuals reporting fried food consumption, mean BMI, prevalence of overweight (BMI ≥ 25), obesity among the overweight (BMI ≥ 30), and median years of education.
Statistical analysis
Descriptive statistics were calculated for the primary outcome variable. Pearson correlation coefficients were computed to examine the relationship between diabetes prevalence and fast-food expenditure, separately for men and women in urban and rural settings. For secondary analyses, data from NSS 2022–2023 and NFHS were merged at the State and gender level. A multiple linear regression model with additive effects was used to examine the relation between dietary predictors and diabetes prevalence. State-level clustering was considered in sensitivity analyses using mixed-effects models.
Results
Out of the combined 36 States and union territories initially considered, data from 18 States met the inclusion criteria and were included in the analysis. States were included if complete State-level data were available for diabetes prevalence (NFHS-5, 2019–2021), monthly per capita fast-food expenditure (NSSO, 2022–2023), overweight and obesity prevalence, fried food consumption frequency, and a median year of education. States with incomplete or inconsistent data in the NSS survey across any of these variables were excluded. Data on monthly food expenditure were unavailable for approximately half of the States. States with missing expenditure data were excluded from analyses.
Kerala showed the highest diabetes prevalence across genders and regions, with rates of ⁓4.0 to 4.1% among both men and women in urban and rural areas. Telangana exhibited a relatively high diabetes prevalence of 3.8% among urban men. In Tamil Nadu, Karnataka, Telangana, Kerala, and Andhra Pradesh, fast-food expenditure accounted for 28–34% of the total monthly per capita food expenditure, equivalent to approximately INR 1,000–INR 1,450 per person per month. The overall monthly per capita food expenditure in urban areas averaged INR 4,120 and in rural regions INR 2,900 (NSSO 2022 23), indicating that in high-consuming States, nearly one-third of food spending is directed toward fast foods.
Among individuals with self-reported diabetes, 63% were either overweight or obese. The proportion of obesity was higher among women (44%) than among men (34%). In Punjab, both urban and rural women demonstrated an obesity prevalence of 14.2%, contrasted with a diabetes prevalence of 2.2%, resulting in a difference of 12 percentage points. Similar disparities were observed in Tamil Nadu and Andhra Pradesh, where obesity rates among women considerably exceeded diabetes prevalence. These States also exhibited high expenditures on fast food and a higher consumption of fried foods, alongside higher mean BMI and overweight percentages.
The mean diabetes prevalence was similar across genders and regions, with men and women showing comparable rates in both rural (men: 2.01%, women: 1.99%) and urban areas (men: 2.01%, women: 1.99%). Average expenditure on fast foods (% per month per capita) was higher in urban populations (men: 25.9, women: 25.9) than in rural populations (men: 21.2, women: 21.2). Findings are mentioned in Table I.
| Gender | Region | Diabetes prevalence (%) mean (SD) |
Expenditure on fast food (% per capita per month) Mean (SD) |
r | P value |
|---|---|---|---|---|---|
| Men | Rural | 2.01 (0.96) | 21.16 (3.07) | 0.389 | 0.110 |
| Urban | 2.01 (0.96) | 25.89 (4.15) | 0.653 | 0.003 | |
| Women | Rural | 1.99 (0.90) | 21.16 (3.07) | 0.394 | 0.105 |
| Urban | 1.99 (0.90) | 25.89 (4.15) | 0.619 | 0.006 |
r, correlation coefficient; SD, standard deviation
The DIAGRAM model revealed statistically significant associations between fast food consumption and diabetes prevalence among urban populations. Among both urban men and women, there was a strong and statistically significant correlation between fast food spending and diabetes prevalence. Specifically, the correlation was higher in urban men (r=0.653, P=0.003) and slightly lower, but still significant, in urban women (r=0.619, P=0.006). In urban areas, people who spent more of their food budget on fast food reported higher rates of diabetes. In rural areas, this relationship was not statistically significant ( Table I). Multiple linear regression models stratified by gender and region were used to examine associations between diabetes prevalence and potential risk factors, including overweight prevalence, obesity prevalence, mean BMI, fried food consumption, and median years of education ( Table II).
| Variable | Beta estimate | Std. Error | P value |
|---|---|---|---|
| Men-Urban | |||
| Fried food (%) | 0.028 | 0.018 | 0.147 |
| Mean BMI | -1.207 | 1.131 | 0.307 |
| Overweight (%) | 0.266 | 0.144 | 0.089 |
| Obese (%) | 0.043 | 0.221 | 0.849 |
| Median education (yr) | -0.002 | 0.268 | 0.993 |
| Men-Rural | |||
| Fried food (%) | 0.028 | 0.018 | 0.147 |
| Mean BMI | -1.207 | 1.131 | 0.307 |
| Overweight (%) | 0.266 | 0.144 | 0.089 |
| Obese (%) | 0.043 | 0.221 | 0.849 |
| Median education (yr) | -0.002 | 0.268 | 0.993 |
| Women-Urban | |||
| Fried food (%) | 0.024 | 0.012 | 0.073 |
| Mean BMI | 0.260 | 0.352 | 0.475 |
| Overweight (%) | 0.163 | 0.071 | 0.041 |
| Obese (%) | -0.114 | 0.124 | 0.375 |
| Median education (yr) | -0.076 | 0.100 | 0.462 |
| Women-Rural | |||
| Fried food (%) | 0.024 | 0.012 | 0.073 |
| Mean BMI | 0.260 | 0.352 | 0.475 |
| Overweight (%) | 0.163 | 0.071 | 0.041 |
| Obese (%) | -0.114 | 0.124 | 0.375 |
| Median education (yr) | -0.076 | 0.100 | 0.462 |
BMI, body mass index
The WEIGHT model (women’s elevated intake and Glucose-Health Evaluation Tracker) found overweight prevalence as a significant predictor of diabetes in both urban and rural women (β=0.163, P=0.04). Fried food consumption demonstrated a borderline significant positive association with diabetes prevalence in women (urban: β=0.024, P=0.07; rural: β=0.024, P=0.07), indicating a possible dietary influence.
Conversely, among men, none of the examined variables reached statistical significance, though overweight prevalence showed a marginal trend towards association with diabetes prevalence (urban: P=0.08; rural: P=0.08). Mean BMI, obesity prevalence, fried food consumption, and education level did not significantly predict diabetes prevalence in men.
Discussion
The findings from our study corroborate global patterns showing strong links between ultra-processed food intake and diabetes.17-19 However, the State-level granularity of the FADIS framework presents new insight into how these links manifest regionally within India. The absence of statistically significant associations between fried food consumption, mean BMI, and diabetes prevalence among men may be explained by reverse causation. Individuals with known diabetes, especially men, are more likely to modify their diet after diagnosis, potentially lowering their intake of fast foods, fried snacks, and sugary beverages. Since the analysis is based on self-reported diabetes, this behavioural modification may attenuate observed associations in men.
The use of the DIAGRAM model allowed us to map these correlations with greater nuance rather than implying causality. These results point toward a possible urban/rural divide: in urban settings, increased access to convenience food outlets, sedentary lifestyles, and marketing exposure20-22 may be playing a more active role in diabetes risk. Under reporting of diabetes is more likely in rural areas due to limited access to diagnostic services, low health literacy, and delayed healthcare-seeking behaviors.9,10 In contrast, over reporting may occur in urban regions where better awareness and increased screening lead to higher self-reported diagnoses. This disparity could partly explain why urban areas showed stronger correlations between fast-food expenditure and diabetes prevalence, whereas rural trends appeared weaker.
The WEIGHT model’s identification of overweight as a significant diabetes predictor among women points toward sociobiological vulnerabilities and systemic issues, including food marketing, caregiving roles, and lower physical activity levels in urban Indian women.6,23
While earlier studies have assessed food processing and health outcomes globally18,24,25, few have applied such granular region-specific analysis within India. In our study, significant regional heterogeneity was observed in fast-food expenditure. Southern States such as Tamil Nadu (33.7%), Karnataka (32.9%), Telangana (32.4%), Kerala (28.4%), and Andhra Pradesh (28.0%) reported the highest per capita fast-food spending, whereas several northern and northeastern States reported rates below 15%. These findings suggest that State-specific regulatory strategies and regional nutrition policies may be more effective than nationwide interventions. The FADIS framework fills that gap and offers a reproducible method for future studies in other low- and middle-income countries undergoing similar transitions.
As this is an ecological study, observed correlations between fast-food expenditure and diabetes prevalence cannot be interpreted as causal, and individual-level risk associations were not assessed. Diabetes prevalence was self-reported, which may lead to underestimation, especially in rural settings. We were unable to restrict the analysis to newly diagnosed diabetes cases. Observed associations may reflect reverse causation, where individuals with diabetes reduce consumption of fast/fried foods following diagnosis. The fast-food variable reflects household expenditure patterns rather than individual consumption, and does not capture specific food types or caloric intake. Important behavioral and clinical variables, such as physical activity, smoking, alcohol use, family history, and healthcare access, were not captured in our datasets. These may act as potential confounding factors influencing both dietary patterns and diabetes prevalence, and their absence may partly limit the precision of our findings. As this is an ecological analysis, our findings represent population-level correlations and should not be interpreted as individual-level causal relationships. The NFHS dataset does not classify overweight and obesity using the Indian-specific BMI cut-offs (>22.9 kg/m2 for overweight and ≥25 kg/m2 for obesity). In our analysis, a BMI ≥25 kg/m2 was used to define overweight/obesity, which may underestimate the actual prevalence when compared to Indian cut-offs. Additionally, the exclusion of some States due to incomplete data may limit the generalisability of the findings.
In conclusion, the FADIS study provides a robust, data-driven analysis of the intersection between dietary change and diabetes prevalence in India. With its DIAGRAM and WEIGHT models, it provides a scalable methodology for mapping population-level nutritional patterns and their correlations with diabetes prevalence. These findings should not be interpreted as risk estimates at the individual level. Key findings show that while urban fast-food expenditure correlates with diabetes prevalence, particularly among men and women, overweight prevalence remains the strongest predictor among women. These results emphasise the importance of gender-focused and regionally tailored interventions but should not be interpreted as evidence of causality. To address these challenges, India must prioritise nutrition education campaigns, urban food policy reform, and community-level interventions targeted at high-risk populations. The tools and insights from FADIS can be extended to other countries experiencing similar nutrition transitions, making it a valuable model for global public health planning.
Author contributions
JYN: Conceptualization, methodology, resources, manuscript writing; RP: Methodology, data curation, formal analysis, manuscript writing. All authors have read and approve the final printed version of the manuscript.
Financial support & sponsorship
None.
Conflicts of Interest
None.
Use of Artificial Intelligence (AI)-Assisted Technology for manuscript preparation
The authors confirm the use of language assistance tools such as OpenAI’s ChatGPT for assisting in the writing of the manuscript, all content generated was subsequently reviewed and edited by the author(s), and no images were manipulated using AI.
References
- Processed foods purchase profiles in urban India in 2013 and 2016: A cluster and multivariate analysis. BMJ Open.. 2022;12:e062254.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- A lot of sacrifices: Work-family spillover and the food choice coping strategies of low-wage employed parents. SocSci Med.. 2006;63:2591-603.
- [Google Scholar]
- Are exposures to ready-to-eat food environments associated with type 2 diabetes? A cross-sectional study of 347 551 UK Biobank adult participants. Lancet Planetary Health.. 2018;2:e438-50.
- [Google Scholar]
- Ultra-processed food availability and noncommunicable diseases: A systematic review. Int J Environ Res Public Health.. 2021;18:7382.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Food environment and diabetes mellitus in South Asia: A geospatial analysis of health outcome data. PLoS Med.. 2022;19:e1003970.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Nutrition and health improvements after participation in an Urban home garden program. J Nutr Educ Behav.. 2019;51:1037-46.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Dietary patterns and associations with metabolic risk factors for non-communicable disease. Sci Rep.. 2023;13:21028.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Fast food, central nervous system insulin resistance, and obesity. Arterioscler Thromb Vasc Biol.. 2005;25:2451-62.
- [CrossRef] [PubMed] [Google Scholar]
- Western-style fast food intake and cardiometabolic risk in an Eastern country. Circulation.. 2012;126:182-8.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Fast food pattern and cardiometabolic disorders: A review of current studies. Health Promot Perspect.. 2016;5:231-40.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Lessons learnt from the ICMR-INDIAB study. Natl Med J India.. 2023;36:137-9.
- [CrossRef] [PubMed] [Google Scholar]
- Evolving food choices among the Urban Indian middle-class: A qualitative study. Front Nutr.. 2022;9:844413.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Health and social determinants and outcomes of home cooking: A systematic review of observational studies. Appetite.. 2017;111:116-34.
- [CrossRef] [PubMed] [Google Scholar]
- Strengthening the reporting of molecular epidemiology for infectious diseases (STROME-ID): An extension of the STROBE statement. Lancet Infect Dis.. 2014;14:341-52.
- [CrossRef] [PubMed] [Google Scholar]
- Ministry of Health and Family Welfare, Government of India. National Health Family Survey-5 (NFHS-5 2019-2020). Available from: https://www.nfhsiips.in/nfhsuser/nfhs5.php, accessed on October 17, 2025.
- Ministry of Statistics and Program Implementation. Government of India. National Statistics Office, India. Available from: https://www.mospi.gov.in/, accessed on June 2, 2025.
- Ultra-processed foods and human health: An umbrella review and updated meta-analyses of observational evidence. Clin Nutr.. 2024;43:1386-94.
- [CrossRef] [PubMed] [Google Scholar]
- Assessing the consumption of ultra-processed foods in people with diabetes: Is a specific NOVA questionnaire always necessary? Nutrients.. 2024;17:1.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Ultra-processed food consumption and risk of diabetes: Results from a population-based prospective cohort. Diabetologia.. 2024;67:2225-3.
- [CrossRef] [PubMed] [Google Scholar]
- Towards unified and impactful policies to reduce ultra-processed food consumption and promote healthier eating. Lancet Diabetes Endocrinol.. 2021;9:462-70.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- The nutrition transition to a stage of high obesity and noncommunicable disease prevalence dominated by ultra-processed foods is not inevitable. Obes Rev.. 2022;23:e13366.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Protecting children from the harmful impact of food marketing. Eur J Paediatr Dent.. 2023;24:173.
- [CrossRef] [PubMed] [Google Scholar]
- High need for recovery from work and sleep problems are associated with workers unhealthy dietary habits. Public Health Nutr.. 2021;24:2185-194.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Consumption of ultra-processed foods and all-cause and cause-specific mortality in the southern community cohort study. Clin Nutr.. 2023;42:1866-74.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Ultraprocessed food: Addictive, toxic, and ready for regulation. Nutrients.. 2020;12:3401.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]