Introduction
Nearly three-quarters of all deaths in low-and-middle-income countries are attributed to non-communicable diseases (NCDs), and India experiences the highest number of NCD-related mortality in the world.(1) As estimated by the World Health Organization (WHO), NCDs accounted for a staggering 66% of total deaths in India in 2022, equivalent to approximately 6,047,000 deaths.(1) Additionally, NCDs could detrimentally impact India’s economy, potentially reducing its gross domestic product by an estimated 0.8%.(Reference Okunogbe, Nugent, Spencer, Ralston and Wilding2) The aetiology of NCDs is multifaceted, encompassing behavioural factors (e.g. poor nutrition, physical inactivity, and substance use), environmental factors (e.g. unhealthy food environments, and limited healthcare access), and infections.(Reference Ezzati, Pearson-Stuttard, Bennett and Mathers3) However, the influence of these factors is significantly modulated by the socio-economic characteristics of the population. To tackle the increasing burden of NCDs in India,(4) the Ministry of Health and Family Welfare, Government of India, developed a national NCD prevention strategy and renders technical and financial support to States and union territories under the National Programme for Prevention and Control of Cancer, Diabetes, Cardiovascular Diseases and Stroke (NPCDCS).(5) Launched in 2010 within the National Health Mission, the NPCDCS implements community-level NCD prevention initiatives across all districts of India.
Modifiable lifestyle factors, such as diet, physical activity, and tobacco and alcohol use, are strongly implicated as risk factors for NCDs,(1,Reference Mohan, Khunti and Prabhakaran6) partly due to their potential link with overweight and obesity, a key metabolic risk factor that can accelerate the development of NCDs. Mitigating the rising burden of overweight and obesity in India could significantly reduce the overall burden of NCDs.(Reference Chaudhary and Sharma7,Reference Rai, Kumar, Singh, Singh, Acharya and Singh8) According to the 2019–2021 National Family Health Survey,(9) the national prevalence of overweight and obesity is 22.9% among men and 24% among women aged 15–49 years. While the NPCDCS provides some guidelines for preventing overweight and obesity, its effectiveness remains to be established. Furthermore, the global literature on interventions targeting overweight and obesity predominantly focuses on children, particularly those of school age.(Reference Adab, Pallan and Lancashire10–Reference Hodder, O’Brien and Lorien12) Consequently, our understanding of potential strategies for preventing overweight and obesity and its determinants in adults is comparatively limited.
Although the burden of NCDs in India has grown substantially since 1990, the country lacks sufficiently detailed data on local NCD epidemiology to effectively inform research and policy formulation,(Reference Arokiasamy13) including high-quality data on overweight and obesity and their determinants. Research on overweight and obesity in India has predominantly relied on cross-sectional datasets with a focus on socio-economic risk factors,(Reference Rai, Kumar, Singh, Singh, Acharya and Singh8) resulting in a scarcity of information on the role of lifestyle factors.(Reference Alvarez-Saavedra, Levasseur and Seetahul14–Reference Gupta, Tamanna, Siddika, Haider, Apu and Haider16) Additionally, operational definitions of lifestyle indicators, such as poor diet quality, inadequate physical activity, and excessive alcohol use, are often inconsistent. These challenges, along with concerns about the external validity of community-based studies in India, further limit our understanding of local overweight and obesity epidemiology.
A systematic review of dietary patterns in India(Reference Green, Milner, Joy, Agrawal and Dangour17) concluded that the consumption of high-fat diets was associated with higher body mass index. Furthermore, a systematic review and dose-response meta-analysis(Reference Golzarand, Salari-Moghaddam and Mirmiran18) estimated that alcohol use, particularly heavy drinking, increased the odds of overweight and obesity. While the effectiveness of physical activity interventions on weight loss and maintenance is unclear,(Reference Robinson and Stensel19) systematic reviews(Reference Bellicha, van Baak and Battista20) convincingly demonstrated that increasing physical activity leads to weight loss. However, some argue(Reference Allison, Bier and Locher21) that such interventions may be less effective at the population level and when not combined with dietary modification. Prior research in India has documented an inverse association between all forms of tobacco use and body mass index.(Reference Pednekar, Gupta, Shukla and Hebert22) However, global literature suggests a more complex relationship, with smoking potentially decreasing body weight in the short term but increasing it in the long term.(Reference Chiolero, Faeh, Paccaud and Cornuz23)
The empirical evidence demonstrating the role of lifestyle risk factors in the development and mitigation of overweight and obesity in India is mixed and inconclusive. This is particularly true for rural populations, where research on the complex interplay of dietary habits, physical activity levels, tobacco and alcohol consumption patterns, and their impact on weight status remains limited. Further investigation using robust study designs and comprehensive data collection is needed to distinguish the specific influences of these lifestyle factors within the diverse socio-economic and cultural contexts of rural India. To address these evidence gaps, we analysed data from a prospective cohort study conducted as part of a large rural health and demographic surveillance system.(Reference Ghosh, Barik and Majumder24,Reference Rai, Barik, Bromage, Dhali and Chowdhury25) Using data collected between 2017–2018 and 2022–2023, we estimated the prevalence, incidence and remission of overweight and obesity. We also examined relationships between key lifestyle risk factors (measured at baseline) and both the incidence and remission of overweight and obesity.
Methods
Data and study population
Data used for this study were collected as part of the Birbhum Population Project (BIRPOP), a Health and Demographic Surveillance System located in the Birbhum district of the state of West Bengal, India.(Reference Ghosh, Barik and Majumder24,Reference Rai, Barik, Bromage, Dhali and Chowdhury25) Established by the Department of Health and Family Welfare, Government of West Bengal in 2008, BIRPOP functions under the ambit of the Society for Health and Demographic Surveillance (SHDS). Largely rural, the district of Birbhum is classified as one of the “backward” districts by the Government of India, and BIRPOP-SHDS covers four administrative blocks (Suri 1, Sainthia, Rajnagar, and Mohammad Bazar) which represent nearly 16% of the district’s population. At its inception, a cohort of 13,053 self-weighted households (59,395 individuals, 51.1% male and 48.9% female) was randomly sampled using the 2001 census sampling frame and accounting for a 10% expected non-response rate.(Reference Ghosh, Barik and Majumder24,Reference Rai, Barik, Bromage, Dhali and Chowdhury25) BIRPOP-SHDS conducts routine surveys (collecting data on vital events; antenatal, natal and postnatal tracking; and verbal autopsy), as well as focused surveys and interventions related to socio-economics, nutrition, NCDs and other areas. Details on sampling procedures and surveys implemented in BIRPOP-SHDS can be obtained from its published reports.(Reference Rai, Kumar, Singh, Singh, Acharya and Singh8,Reference Ghosh, Barik and Majumder24–Reference Rai, Barik and Chowdhury27)
During 2017–2018, BIRPOP-SHDS undertook a survey to understand the role of modifiable determinants of body mass index(28) among rural Indians aged ≥18 years (sample: 11,399). Information on demographic and socio-economic characteristics, dietary practices, physical activity, and alcohol and tobacco use were collected and participant’s height and weight were measured. A follow-up survey, conducted in 2022–2023, measured height and weight among 8,974 (78.7%) of those in the initial survey. Attrition of 2,425 individuals was attributed to permanent/ temporary migration (38.1%), absenteeism (34.6%), pregnancy (17.3%) and death (10%).
Incidence and remission of overweight including obesity
To calculate body mass index (BMI, an individual’s weight in kilograms divided by their height in metres squared), a standardised height and weight measurement protocol and inclusion and exclusion criteria were followed to obtain height and weight data in the study population.(Reference Rai, Jaacks, Bromage, Barik, Fawzi and Chowdhury29) Height was measured using a stadiometer (Bioplus Stature Meter, model IND/09/2005/815) and weight was measured using a certified electronic scale (Omron HN-283). These measurements were taken by trained surveyors of BIRPOP-SHDS. “Overweight/obesity” was defined as a BMI of ≥23 kg/m2 according to WHO guidelines developed for Asian Indians.(Reference Misra, Chowbey and Makkar30) Following definitions used in an earlier study with the BIRPOP-SHDS population,(Reference Rai, Jaacks, Bromage, Barik, Fawzi and Chowdhury29) incidence of overweight/obesity was defined as a transition from normal BMI (18.5–22.9 kg/m2) in 2017–2018 to overweight/obesity in 2022–2023, while remission was defined as a transition from overweight/obesity in 2017–2018 to normal BMI in 2022–2023. Smoothed plots (graphically presented using local polynomial smoothing) of mean BMI during 2017–2018 and 2022–2023 and change from 2017–2018 to 2022–2023 by gender an age are presented in Figure 1. At baseline (2017–2018), a mean BMI of approximately 20.7 kg/m2 was observed among men from their mid-twenties to late fifties. In contrast, mean BMI among women peaked at 40 years (nearly 21.3 kg/m2) and decreased thereafter. The change in BMI from 2017–2018 to 2022–2023 was highest among men and women aged 20 years at baseline (0.5 kg/m2) and lowest around 80 years old (nearly –0.5 kg/m2).
Figure 1. Local polynomial-smoothed plots of mean (with 95% confidence interval or CI) body mass index (BMI) during 2017–18 (top left), 2022–2023 (top right), and change in BMI from 2017–2018 to 2022–2023 (bottom left) by age and gender.
Lifestyle risk factors
Lifestyle risk factors for overweight/obesity were represented by indicators on dietary habits, physical activity, alcohol consumption, and tobacco use. Dietary habits were assessed using food group-specific frequency options,capturing consumption of meat, milk, lentils, fruits, vegetables, and added salt. Consumption of four categories of processed or prepared foods (potato chips/wafers, other snacks, paratha/luchi/puri, and biriyani/fried rice) and food consumed outside the home were also assessed. Physical activity was measured as the number of days per week that participants engaged in at least 10 minutes of both vigorous and moderate physical activity during both leisure and non-leisure time, walking (assessed among both employed and unemployed participants), and use of different transportation modalities (unemployed participants only). Participants were asked about their alcohol consumption - those who reported ever consuming alcohol were then asked about current consumption (defined as any consumption in the past 30 days) and their average daily volume of alcohol consumed. Frequency of smoking and smokeless tobacco use were also assessed, and participants were categorised as ‘often/daily smokers/users’ and ‘non-smokers/users’.
Statistical approach
The prevalence of overweight/obesity, with associated 95% confidence intervals (CIs), was calculated for both men and women in 2017–2018 and 2022–2023. Prevalence was examined within strata defined by key background characteristics and lifestyle risk factors. Information on background characteristics included each participant’s age, years of schooling, occupation, religion, social group, wealth quintile, mass media exposure, distance from their house to the nearest healthcare facility, and health insurance coverage. Social groups, namely Scheduled Castes (SCs), Scheduled Tribes (STs), Other Backward Classes (OBCs), and Others, were defined according to the Constitution of India. The Constitution recognises SCs, STs, and OBCs as economically disadvantaged compared to the ‘Others’ group.(Reference Rai, Jaacks, Bromage, Barik, Fawzi and Chowdhury29) Information on household goods and durables was used to derive a wealth index using principal component analysis.(Reference Vyas and Kumaranayake31) To measure mass media exposure, respondents were asked how often they read newspapers, listen to radio, or watch television (almost every day or high exposure, once or more per month medium exposure, or once or less per three months or low exposure).(Reference Singh, Rai and Singh32) Modified Poisson regression models with robust error variance(Reference Zou33) were used to estimate relative risks (RRs) with 95% CIs of overweight/obesity incidence and remission associated with each lifestyle risk factor, adjusting for potential confounders. The statistical software Stata version 18(34) was used for all analyses and statistical significance was defined based on two-tailed P-values <0.05.
Results
From 2017–2018 to 2022–2023, the prevalence of overweight/obesity increased by nearly six percentage points among men (from 15.2% (95% CI: 14.1%–16.4%) to 21.0% (95% CI: 19.7%–22.3%)) and nearly ten percentage points among women (from 24.1% (95% CI: 22.9%–25.2%) to 33.8% (95% CI: 32.5%–35.1%)) (Table 1). Among both men and women, alcohol users had a lower prevalence of overweight/obesity than non-users, while those who used tobacco (including both smoked and smokeless tobacco) had a higher prevalence of overweight/obesity than non-users. Overweight/obesity was more prevalent among both the wealthiest individuals (compared to the poorest) and those with high mass media exposure (compared to those with low exposure).
Table 1. Prevalence (%) of overweight/obesity by key lifestyle factors and background characteristics among men and women, 2017–2018 and 2022–2023
n, sample; CI, confidence interval; km, kilometre; ml, millilitre; OBCs, other backward classes; SCs, scheduled castes; STs, scheduled tribes.
– : percentage estimates were not calculated as the covariates are continuous variables.
* Recorded as number of days per week engaged in at least 10 minutes of each activity category.
Table 2 presents the incidence of overweight/obesity and its associated lifestyle factors. Among a total of 4,056 participants with normal BMI in 2017–2018, 23.0% (95% CI: 21.8%–24.3%) developed overweight/obesity by 2022–2023. Multivariable models revealed a positive association between the use of a motorised vehicle (motorcycle, scooter, auto-rickshaw, car, etc.) among unemployed study participants and the development of overweight/obesity (RR: 1.058; 95% CI: 1.023–1.095; P: 0.001). Among 1,831 participants with overweight/obesity in 2017–2018, 13.9% (95% CI: 12.4%–15.6%) experienced remission as measured during 2022–2023 (Table 3). Each 1 ml in alcohol consumption was also inversely associated with remission (RR: 0.995; 95% CI: 0.991–0.999; P: 0.022). Among unemployed adults, vigorous home-based physical activity (including gardening, yard work, and household chores) was linked to higher odds of recovering from overweight/obesity (RR: 1.065; 95% CI: 1.008–1.125; P: 0.025). Compared to those who used tobacco often/daily, non-users were less likely to experience remission from overweight/obesity (RR: 0.689; 95% CI: 0.484–0.980; P: 0.038).
Table 2. Incidence of overweight/obesity from 2017–2018 to 2022–2023 and its associations with key lifestyle risk factors
n, sample; CI, confidence interval; ml, millilitre; RR, relative risk; (ref.), referent.
* Predictors are adjusted for age, gender, years of schooling, occupation, religion, social group, wealth quintile, mass media exposure, distance to nearest health facility, and health insurance.
– : percentage estimates were not calculated as the covariates are continuous variables.
Table 3. Remission of overweight/obesity from 2017–2018 to 2022–2023, and its associations with key lifestyle risk factors
n, sample; CI, confidence interval; ml, millilitre; RR, relative risk; (ref.), referent.
* Predictors are adjusted for age, gender, years of schooling, occupation, religion, social group, wealth quintile, mass media exposure, distance to nearest health facility, and health insurance.
– : percentage estimates were not calculated as the covariates are continuous variables.
Discussion
Using prospective cohort data from a health and demographic surveillance system in West Bengal, India, this study examined associations between key lifestyle risk factors (dietary habits, physical activity, and alcohol and tobacco use) measured at baseline and subsequent incidence and remission of overweight/obesity among rural Indian men and women between 2017–2018 and 2022–2023.
This study observed a considerably higher prevalence and increase in the prevalence of overweight and obesity in women than in men, consistent with prior research(Reference Rai, Jaacks, Bromage, Barik, Fawzi and Chowdhury29) conducted in the BIRPOP-SHDS cohort (baseline survey: 2008; follow-up survey: 2017). These findings may be partly explained by biological factors such as age-related weight gain and hormonal and pregnancy-related effects among women.(Reference Cooper, Gupta, Moustafa and Chao35) However, sociocultural factors are also crucial to consider. Women in rural India often face constraints on their mobility and autonomy due to traditional gender roles and limited social support.(Reference Khandelwal and Reddy36) This can restrict their opportunities for engaging in physical activity. Additionally, in rural India there often exists a division of labour wherein women typically engage in less physically demanding tasks than men, further contributing to lower activity levels.(Reference Meshram, Vishnu Vardhana Rao, Sudershan Rao, Laxmaiah and Polasa37)
This study revealed a link between increased use of motorised transport and a higher likelihood of developing overweight/obesity. These findings are consistent with prior research in India, which utilised data from the National Family Health Survey (2005–2006 and 2015–2016) and identified a positive association between household motor vehicle ownership and obesity, even after accounting for socio-economic factors.(Reference Kellstedt, Washburn, Lee, Gwarzo, Ahenda and Maddock38) Similar results have been reported in studies conducted in Bangladesh.(Reference Ahmed, Yakub, Khan and Yunus39) However, this study makes a unique contribution to the existing literature in that while earlier research focused on vehicle ownership at the household level, this study examined individual vehicle use.
This study provides evidence that engaging in vigorous physical activity at home – such as gardening, yard work, and household chores - was conducive to remission from overweight/obesity. In addition to increasing energy expenditure, engaging in physical activity may improve metabolic health, further supporting weight management and overall health. These findings are encouraging, especially for women in India who may have limited opportunities to participate in structured exercise programmes. Engaging in physical activity at home, like household chores, gardening, or playing with children, could offer a practical and accessible way to increase their energy expenditure and manage weight.(Reference Tanucan, Wider and Lobo40)
Alcohol use (expressed as a continuous variable) and abstinence from tobacco use were inversely associated with remission from overweight-obesity. An earlier study reported that all forms of tobacco use are associated with low BMI in India,(Reference Pednekar, Gupta, Shukla and Hebert22) while a study conducted in the BIRPOP-SHDS cohort found that remission from overweight/obesity was less likely among alcohol users.(Reference Rai, Jaacks, Bromage, Barik, Fawzi and Chowdhury29) These findings may be attributed in part to lower levels of physical activity among alcohol and tobacco users, in addition to appetite-suppressing effects of nicotine and the obesogenic effects of alcohol, respectively.(Reference Chiolero, Faeh, Paccaud and Cornuz23,Reference Downer, Bertoia, Mukamal, Rimm and Stampfer41) It is important to note that these findings highlight complex relationships; further research is needed to fully understand these associations and their implications for public health interventions in rural India.
Furthermore, while our multivariable analysis showed no statistically significant association between dietary habits and overweight/obesity incidence or remission (Tables 2 and 3), our descriptive analysis (Table 1) showed an increasing prevalence of overweight/obesity among respondents who reported consuming more lentils, fruits, and vegetables. Diets high in plant-based foods like fruits, vegetables and lentils are generally considered to promote weight loss(Reference Mytton, Nnoaham, Eyles, Scarborough and Ni Mhurchu42–Reference Chamberlin, Wilson, Gaston, Kuo and Miles44) as these foods tend to be lower in energy density relative to their volume, promoting satiety and hence less energy intake per meal.(Reference Fung, Li and Bromage45) However, despite the high intake of lentils, fruits, and vegetables, Indians also consume a large amount of white rice and other refined grain products,(Reference Matsuzaki, Birk and Bromage46) along with processed foods which often contain high calories and low nutrients,(Reference Chaudhary and Sharma7) contributing to an increased prevalence of overweight and obesity. Furthermore, socio-economic status may confound observed prevalence estimates. Further research is needed to understand the complex interplay of factors contributing to overweight/obesity in this population.
Interpreting the study findings should consider a few limitations. First, as assessment of lifestyle risk factors relied on participant recall, findings may be influenced by recall error and social desirability bias. Second, the lack of data on certain potential confounders (for example, sleep quality), may have introduced bias in estimation of relative risks for relationships between other lifestyle and background factors and overweight/obesity incidence and remission. Third, because dietary data were limited to broad food groups, some foods with both obesogenic and anti-obesogenic properties may have been grouped together, potentially affecting the interpretation and translatability of the results. Fourth, although our primary multivariable models include gender as a covariate to account for its potential influence on the observed associations, sample size limitations rendered it infeasible to conduct separate regression models for men and women with adequate statistical power. Finally, because lifestyle risk factors were not measured in the follow-up survey round, inference is limited to effects of factors measured at baseline only, not how changes in those factors over time might influence outcomes; by analysing repeat measurements of BMI, we could nonetheless understand how baseline risk factors are associated with BMI trajectories, which is key toward establishing relationships in observational studies of weight gain and loss.
Overall, this prospective cohort study represents a novel investigation into the long-term dynamics of weight change among rural Indian adults, shedding light on how modifiable lifestyle factors contribute to both the incidence and remission of overweight/obesity. Findings of this study may be used to guide future research on obesity epidemiology, and support design of programmes for mitigating the rapidly escalating burden of overweight/obesity and related impacts on NCD development in rural India. Future research should prioritise more detailed dietary assessment to understand windows for encouraging healthy and locally-acceptable food-based diet modifications. Furthermore, the findings underscore the importance of addressing transportation habits and promoting physical activity. To combat the growing prevalence of overweight/obesity, public health initiatives should prioritise the promotion of active lifestyles by investing in and incentivizing sustainable transportation options, such as expanding cycling infrastructure with dedicated bike lanes and bike-sharing programmes, improving public transportation networks with increased frequency and affordability, and creating pedestrian-friendly environments that prioritise walking and walkability. Additionally, promoting physical activity within the home environment, by encouraging activities such as gardening, vigorous household chores, and active play with children, can be a practical for people who may face social, cultural, or safety-related barriers to exercising outside the home, thereby empowering them to increase their physical activity levels and improve their overall health. However, this study does not suggest limiting outdoor activity. Rather, it highlights that individuals with limited opportunities to exercise outside can engage in household activities to mitigate overweight/obesity. The NPCDCS should be enhanced by incorporating community-level interventions that are both culturally acceptable and feasible within the time constraints of rural Indian men and women.(Reference Aiyar, Dhingra and Pingali47) By integrating these culturally sensitive and time-conscious interventions into the NPCDCS, the programme can effectively reach and empower rural communities to adopt healthier lifestyles and reduce their risk of overweight/obesity and associated NCDs.
Data availability statement
The data of this study are available upon reasonable request from the corresponding author and the signing of a data transfer agreement.
Acknowledgements
The authors are indebted to the household members who generously participated in surveys undertaken by the Birbhum Population Project, Society for Health and Demographic Surveillance, West Bengal, India.
Authorship
R.K.R. designed and conducted the research, analysed the data, and wrote the draft manuscript. R.K.R. and A.B. designed the survey. S.B., J.W.D.N., and A.B. provided critical feedback throughout the manuscript preparation process. With full access to the dataset, R.K.R. had primary responsibility for the final content. All authors have collectively reviewed and approved the final version of the manuscript.
Financial support
This work was supported by the Department of Health and Family Welfare, Government of West Bengal, India.
Competing interests
The authors declare none.
Ethical standard disclosure
This study was conducted in accordance with the guidelines laid down in the Declaration of Helsinki and all procedures involving research study participants were approved by the institutional ethics review board of the Birbhum Population Project, appointed by the chairperson of the Society for Health and Demographic Surveillance. Written and verbal informed consent was obtained from all subjects. Verbal consent was witnessed and formally recorded.
Open access
