Is the rural population of Puducherry district healthy in terms of the burden of non-communicable diseases? Findings from a cross-sectional analytical survey.

INTRODUCTION: Epidemiological transition remains a key contributor to the rising prevalence of non-communicable diseases (NCDs) across developing nations. Population-specific NCD risk factors estimates derived using World Health Organization (WHO) 'STEP-wise approach' are crucial for devising evidence-based public health interventions to combat NCDs.
OBJECTIVE: To estimate the prevalence of behavioral and biological risk factors for NCDs among the rural adult population of Puducherry district in India.
METHODOLOGY: STEPS survey was conducted by following all three steps (behavioral, physical measurements and biochemical risk factors) of NCD risk factor assessment. A total of 790 participants were selected from 50 villages through multistage cluster sampling method. STEPS instrument was used to assess behavioral risk factors, physical measurements and biochemical (fasting blood glucose and total cholesterol) risk factors.
RESULTS: Tobacco use and alcohol consumption were present among 11.3% (95% Confidence Interval (CI): 9-13.6%) and 19.2% (95% CI: 16.5-22.4%) of the population, respectively. Low physical activity, inadequate intake of fruits and vegetables, overweight and obesity were observed among 29.3% (95% CI: 26.2-32.7%), 89.8% (95% CI: 87.6-92%), 15.6% (95% CI: 13.1-18.3%) and 38.9% (95% CI: 35.4-42.2%), respectively. About 28.2% (95% CI: 25.2-31.6%) had hypertension and 24.4% (95% CI: 20-29%) had diabetes mellitus. Abdominal obesity was twice highly prevalent among women. Tobacco and alcohol use were more common among men, whereas low physical activity, obesity and hypercholesterolemia were higher among women.
CONCLUSION: Public health interventions to promote healthy lifestyle need to be initiated especially to increase physical activity, intake for fruits and vegetables, and quitting of tobacco and alcohol consumption in the rural population of Puducherry.

Introduction

NCDs are increasing worldwide, especially among the low-and middle-income countries where more than three-fourths of the world NCD deaths occur.[1] In India, NCDs account for 61% of all deaths.[2] Between the years 1990 and 2016, NCDs such as diabetes and cardiovascular diseases shared the major portion of NCD deaths in the country. During the same period, the Union Territory (UT) of Puducherry had recorded a higher middle rate (0.4) of epidemiological transition indicating substantial increase in the prevalence of various behavioral and biological risk factors of NCDs in comparison to other Indian states.[3] These behavioral risk factors when not recognized and intervened timely by a health system, lead to development of metabolic risk factors and consequent rise in NCDs in the population. Therefore, considering the rapid epidemiological transition occurring in the rural parts of Puducherry district, and lack of estimates for NCD risk factors in the population, we conducted a survey to estimate the prevalence of behavioral and biological risk factors of NCDs using World Health Organization (WHO) prescribed 'STEP-wise approach to NCD risk-factor surveillance' method.[4]

Methods

Study setting

Puducherry is one of the four districts of Puducherry UT located in the southeastern coast of India. The total population of the district is 9.5 lakh, of which 31.77% resides in the rural areas (Census, 2011). Puducherry is ranked seventh among all Indian states in Human Development Index.[5] Literacy rate in Puducherry is 85.44% and it has a sex ratio of 1029 females per 1000 males.[6] This survey was carried out in rural areas (villages) of Puducherry district.

Study design and sample size

We conducted a cross-sectional analytical study using WHO prescribed STEPS approach for a period of 7 months (February to August) in 2019. Sample size was calculated using OpenEpi (version 3.01) keeping the prevalence of risk factors as 50%, confidence interval (CI) as 95%, and design effect as 1.75. The sample size of 672 derived using these values was raised to 790 considering an anticipated 15% (672/0.85 = 790) non-response rate. This study was part of a larger survey[7] whose findings from rural component are presented in this paper.

Sampling method

Multistage, stratified and geographically clustered sampling method was employed using Census 2011 data as the baseline. Two-staged sampling method was utilized. In the first stage, of the 62 revenue villages in Puducherry district, fifty villages were selected randomly. In the second stage, 16 households were selected from each village using systematic random sampling method to achieve the sample size of 790. On reaching the household, one adult (18–69 years old) member who has resided in the household for at least the past 6 months was chosen using Kish method.[8]

Data collection instrument

STEPS instrument (version 3.2) was used for collecting behavioral and biological risk factors.[9] The questionnaire was culturally adapted, Tamil translated and pre-tested before the survey. Translation of the questionnaire was done by two authors of this manuscript independently, whose mother tongue was Tamil, to better reflect the nuances of the questionnaire in the target language. Discrepancies identified in the translations were discussed and resolved with the involvement of other authors of the study. The translated version was then presented to an expert committee who were familiar with the constructs of the study questionnaire. The expert members reviewed all the items of the questionnaire for appropriateness to produce the pre-final version. This version was piloted on a small sample (about 30 participants) to resolve any further discrepancies in the translated questionnaire before arriving at the final questionnaire.

In the study, risk factors were assessed in three steps. Socio-demographic and behavioral risk factors (tobacco use, alcohol consumption, inadequate intake of fruits and vegetables, physical inactivity) were assessed in the first step. Physical measurements such as blood pressure (BP), height, weight and waist circumference (WC) were recorded in Step 2. Fasting blood glucose (FBG) and total cholesterol were assessed in fasting blood samples collected from alternate participants in Step 3.

Data collection procedure

We initiated data collected after obtaining approval from JIPMER Institute's Ethics Committee for Human Studies on 03/10/2018 (JIP/IEC/2018/0246). A trained investigator obtained informed written consent from the study participants after describing the study purpose, procedure and the potential risks and benefits of their participation. At each selected village, a household was randomly selected using village map and household addresses; thereafter, systematic sampling method was used to choose the required number of households in the village. When members of a household were unavailable, two more visits were made at a time convenient to them. All three steps of the survey were carried out in the households of the participants.

Behavioral risk factors (Step 1)

Demographic information included education, occupation, and income including basic details such as gender, age, residence and contact details. History of tobacco use (smoke and smokeless forms), alcohol consumption (quantity), fruit and vegetable intake (quantity and frequency) was also obtained. Physical activity was asked at work, travel and leisure activities using Global Physical Activity Questionnaire (GPAQ).[10] Show cards were used to describe type and intensity of physical activities and serving of fruits and vegetables.

Physical measurements (Step 2)

Standard procedure was followed to measure the height, weight and WC.[4] Height and weight were measured in standing position with barefoot and light clothing using portable stadiometer and electronic weighing scale, respectively. Height and weight were recorded to the closest 0.1 cm and 100 g, respectively. WC was recorded using a Seca constant tension tape to the nearest 0.1 cm. BP was measured in seated position. Three BP readings were recorded in sitting position at 3 min intervals using electronic device (OMRON, HEM 7120, Omron Corporation, Kyoto, JAPAN), and the average of the last two readings was used to determine the BP level. All equipments used were regularly calibrated before and during the data collection.

Biochemical measurements (Step 3)

A written instruction on the pre-requisites for fasting blood sample collection was handed over to the alternate participants (50% of the study sample) on the previous day of sample collection. Blood samples were collected after 10-12 h of overnight fasting. Five milliliter of blood sample was collected in seated position and estimation was done at the Central Laboratory of Institute's Biochemistry Department.

FBG and total cholesterol were estimated enzymatically using the glucose oxidase peroxide and cholesterol oxidase peroxide method, respectively and both measured at 520 nm. They were determined utilizing commercially available kits adapted to clinical chemistry autoanalyzer based on spectrophotometry (Beckman Coulter Inc, Brea, CA, USA).

Information obtained from all three steps of the survey were recorded in the electronic version of the questionnaire loaded in the data collection software ODK collect (version 1.25.1).

Operational definitions

Risk factors were determined based on the cut-off recommendations of STEPS.[4] Tobacco use (in any form in the last 30 days) and alcohol consumption in the last 1 year was considered as a current tobacco and alcohol user. Alcohol consumption of ≥60 gm of pure alcohol for men and ≥40 gm for women on an average day in the last 30 days was considered as harmful use of alcohol. Fruits and vegetables intake of less than five servings was regarded as inadequate intake. Physical activity was measured at work, travel and leisure activities which were further assessed for the intensity levels: low, moderate and severe. The duration of the physical activity captured were converted into Metabolic Equivalent Time per week (MET min/week). As per STEPS recommendation, those who had less than 600 MET, between 600 and 1500 MET and more than 1500 MET min/week were classified to have low, moderate and high levels of physical activity.

Asian cut-off of body mass index (BMI) was used to categorize study participants. BMI between 23 and 24.9 kg/m2 and ≥25 kg/m2 were regarded as overweight and obese, respectively. Abdominal obesity was defined as men having WC of ≥90 cm and women ≥80 cm.[11] Hypertension was defined as those who had systolic blood pressure (SBP) of ≥140 mm/Hg[12] and/or diastolic blood pressure (DBP) of ≥90 mm/Hg or currently on BP-lowering drugs.

Diabetes mellitus was defined as FBG level of ≥126 mg/dL[13] or currently on anti-diabetic medication. Hypercholesterolemia as those having total cholesterol level of ≥200 mg/dL[14] or currently on lipid-lowering drugs.

Statistical analysis

Data cleaning and analysis were carried out using STATA version 14.1 (StataCorp LP, College Station, TX, USA). Prevalence of various risk factors was determined by different age groups and gender. Continuous variables (such as age, number of sticks of tobacco use and amount of alcohol) and categorical variables (such as tobacco use, alcohol use, physical activity, intake of fruits and vegetables) were summarized using mean and proportions, respectively, with 95% CI. Significant difference in risk factors prevalence between the levels of independent variables (age group and gender) was determined by comparing 95% CI.

To account for the clustering effect on dependent variables, forward weighted adjusted prevalence ratio (PR) was determined for each risk factor by keeping the risk factors as dependent variables and socio-demographic characteristics (gender, age group, education, marital status, occupation) as independent variables using 'multilevel mixed-effects generalized linear model' under 'Poisson's regression'. The PRs obtained from regression models indicated the risk of having a risk factor in a selected group compared to the reference group of the independent variable. In the models, P ≤ 0.05 was considered statistically significant.

Results

The response rate for the survey Steps 1-3 were 734/790 (92.9%) and 355/395 (90%), respectively.

The mean age (SD) of the participants was 43.94 (16) years and majority (54.9%) were women. The study witnessed almost equal participation in 18-44 years (49%) and 45-69 years (51%) age groups. About half of the participants had received primary or secondary level of education. Socio-demographic characteristics of the participants are given in Table 1.

Table 1

Socio-demographic profile of study participants (n=734)

Variables	Men (n=337) n (%)	Women (n=397) n (%)	Both Genders (n=734) n (%)
Age categories (years)
18-44	157 (46.6)	203 (51.1)	360 (49)
45-69	180 (53.4)	194 (48.9)	374 (51)
Educational status
No formal education	39 (11.6)	91 (22.9)	130 (17.7)
Less than primary	55 (16.3)	83 (20.9)	138 (18.8)
Primary education	82 (24.3)	99 (24.9)	181 (24.7)
Secondary/Higher Secondary	99 (29.4)	89 (22.4)	188 (25.6)
Graduation and above	60 (17.8)	34 (8.6)	94 (12.8)
Refused to answer	2 (0.6)	1 (0.3)	3 (0.4)
Religion
Hindu	329 (97.6)	390 (98.2)	719 (98)
Christian	4 (1.2)	5 (1.3)	9 (1.2)
Muslim	4 (1.2)	1 (0.3)	5 (0.7)
Others	0	1 (0.3)	1 (0.1)
Marital status
Never married	77 (22.8)	20 (5)	97 (13.2)
Currently married	247 (73.3)	295 (74.3)	541 (73.8)
Divorced/Separated	0	4 (1)	4 (0.5)
Widowed	13 (3.9)	78 (19.6)	91 (12.4)
Occupation
Government employee	11 (3.3)	2 (0.5)	13 (1.8)
Non-Government employee	100 (29.7)	43 (10.8)	143 (19.5)
Self-employed	157 (46.6)	60 (15.1)	217 (29.6)
Student	22 (6.5)	10 (2.5)	32 (4.4)
Homemaker	0	260 (65.5)	260 (35.4)
Retired	5 (1.5)	1 (0.3)	6 (0.8)
Unemployed (able to work)	22 (6.5)	18 (4.5)	40 (5.4)
Unemployed (unable to work)	20 (5.9)	3 (0.8)	23 (3.1)

The mean consumption of tobacco (smoke form), alcohol and other behavioral risk factors is presented in Table 2. The mean BMI and WC were significantly higher among women. FBG was higher among men whereas total cholesterol was higher among women [Table 2].

Table 2

Mean (Confidence Interval) of various behavioral and biological risk factors and physical measurements by age group and gender among the study participants (n=734)

Variables	Mean (CI) of behavioral risk factors on average per day (in numbers)
	Daily tobacco use (smoke form)	Standard drinks of alcohol consumed	Servings of fruits and vegetables intake	MET of physical activity/week
Age (years)
18-44	5.4 (3.4-7.3)	5 (4.1-6)	2.9 (2.7-3.2)	2635 (2308-2963)
45-69	9.2 (6.5-11.8)	5.8 (4.8-6.9)	2.2 (2-2.4)	2469 (2179-2760)
Gender
Male	7.9 (6-9.8)	5.54 (4.8-6.3)	2.8 (2.6-3)	3096 (2731-3462)
Female	-	2.33 (0.5-5.2)	2.3 (2.2-2.5)	2088 (1838-2338)
Overall	7.9 (6-9.8)	5.46 (4.76-6.17)	2.6 (2.4-2.7)	2551 (2333-2769)
	Mean (CI) of physical measurements
	BMI	Waist circumference (cm)	Systolic blood pressure (mmHg)	Diastolic blood pressure (mmHg)
Age (years)
18-44	24.5 (24-25)	88.3 (87-89.6)	119 (118-120)	79 (78-80)
45-69	24 (23.5-24.5)	87.3 (86-88)	125 (123-126)	81 (80-82)
Gender
Male	23.3 (22.8-23.8)	86.2 (85-87.4)	124 (122-125)	81 (80-83)
Female	25 (24.5-25.6)	89 (87.8-90.3)	120 (119-122)	79 (78-80)
Overall	24.3 (23.9-24.6)	87.8 (86.9-88.6)	122 (121-123)	80 (79-81)
	Mean (CI) of biological risk factors
	Fasting blood glucose (mg/dl)		Total cholesterol (mg/dl)
Age (years)
18-44	100.1 (92.9-107.4)		171.6 (165.8-177.4)
45-69	111.1 (102.6-119.7)		179 (172.5-185.7)
Gender
Male	109.2 (99.4-119)		169.1 (162.6-175.6)
Female	103.3 (96.8-110)		181 (175-187)
Overall	106 (100.3-111.7)		175.6 (171.1-180)

In the study, 11.3% (95% CI: 9-13.6%) were current tobacco users. Current use of alcohol was present among 19.2% (95% CI: 16.5-22.4%) of the population. Tobacco and alcohol use, including harmful use of alcohol, were significantly higher among men compared to women. Inadequate fruits and vegetables intake was present in 89.8% (95% CI: 87.6-92%) of the population, with no significant difference between genders and age groups [Table 3].

Table 3

Prevalence of various risk factors of NCDs by age group and gender among the study participants

Variables	Current tobacco use, % (CI)	Current use of alcohol, % (CI)	Harmful use of alcohol, % (CI)
Age (years)
18-44	9.4 (6.4-12.5)	18.6 (14.8-23)	10.6 (7.8-13.9)
45-69	13.1 (9.6-16.3)	19.8 (16-24.1)	8.8 (6.1-11.8)
Gender
Male	21.4* (17.2-25.5)	40.9* (35.9-46.5)	20.5* (16-24.6)
Female	2.8 (1.3-4.5)	0.8 (0.4-1.8)	0.5 (0-1.3)
Overall	11.3 (9-13.6)	19.2 (16.5-22.4)	9.7 (7.5-11.7)
	Low physical activity, % (CI)	Inadequate intake of fruits and vegetables, % (CI)	Raised blood pressure, % (CI)
Age (years)
18-44	29.4 (24.7-34.4)	84.4 (80.6-87.8)	18.3 (14.6-22.3)
45-69	29.1 (24.3-33.7)	94.9 (92.5-97.1)	37.7* (32.7-42.6)
Gender
Male	22.8 (18.7-27.3)	85.5 (81.6-89)	32.6 (27.5-37.3)
Female	34.8* (30.2-39.3)	93.5* (90.7-95.7)	24.4 (20.3-28.7)
Overall	29.3 (26.2-32.7)	89.8 (87.6-92)	28.2 (25.2-31.6)
	Overweight, % (CI)	Obesity, % (CI)	Abdominal obesity, % (CI)
Age (years)
18-44	14.6 (10.9-18.1)	42 (37.5-47.1)	60.2 (55.3-65.3)
45-69	16.6 (12.9-20.4)	35.8 (31-40.8)	59.4 (54.5-64.7)
Gender
Male	19.6 (15.4-23.7)	29.7 (24.9-34.7)	38.3 (32.9-43.7)
Female	12.2 (9-15.5)	46.7* (42.2-51.8)	78.2* (73.7-82.7)
Overall	15.6 (13.1-18.3)	38.9 (35.4-42.2)	59.8 (56.4-63.4)
	Diabetes mellitus, % (CI)	Hypercholesterolemia, % (CI)
Age (years)
18-44	14.4 (9.4-20.3)	21.1 (15.4-27.4)
45-69	33.2* (27-39.8)	32.8 (26.3-39.7)
Gender
Male	25.5 (19-32.4)	22.3 (16.1-28.8)
Female	23.5 (17.7-30.1)	31.6 (24.9-38.1)
Overall	24.4 (20-29)	27.3 (22.7-32.3)

*Indicates statistically significant difference between the sub-groups

In the population, 29.3% (95% CI: 26.2-32.7%) adults had low physical activity which was significantly higher among women 34.8% (95% CI: 30.2-39.3%) compared to men 22.8% (95% CI: 18.7-27.3%). More than one-third and one-tenth were obese 38.9% (95% CI: 35.4-42.2%) and overweight 15.6% (95% CI: 13.1-18.3%), respectively [Table 3].

Nearly one-fourth of the population had diabetes mellitus 24.4% (95% CI: 20-29%), raised BP 28.2% (95% CI: 25.2-31.6%) and hypercholesterolemia 27.3% (95% CI: 22.7-32.3%). Both diabetes mellitus and raised BP were significantly higher among 45-69-year-olds.

The probability of having each of the risk factors across various sub-groups of socio-demographic characteristics of the population was estimated using multilevel mixed-effects generalized linear models [Table 4]. The prevalence of tobacco (Prevalence Ratio (PR): 1.2, 95% CI: 1.16-1.3) or alcohol use (PR: 1.5, 95% CI: 1.39-1.6) was significantly higher among men. Inadequate intake of fruits and vegetables (PR: 1.09, 95% CI: 1.03-1.6), hypertension (PR: 1.15, 95% CI: 1.06-1.24), and diabetes mellitus (PR: 1.14, 95% CI: 1.04-1.24) was significantly higher among 45-69-year-olds.

Table 4

Determinants of various risk factors of NCDs among the rural population of Puducherry district (n=734)

Variables	Current Tobacco Use PR (95% CI)	Current Alcohol Use PR (95% CI)	Physical Inactivity PR (95% CI)	Insufficient Servings PR (95% CI)	Raised Blood Pressure PR (95% CI)	Raised Blood Glucose PR (95% CI)	Overweight/Obesity PR (95% CI)	Raised Cholesterol PR (95% CI)
Gender
Male	1.2* (1.16-1.3)	1.5* (1.39-1.6)	1	1	1	1	1	1
Female	1	1	1.1 (0.99-1.2)	1.07 (1-1.13)	0.92 (0.85-1.01)	0.9 (0.85-1.03)	1 (0.9-1.1)	1.1 (0.96-1.3)
Age (years)
18-44	1	1	1	1	1	1	1	1
45-69	0.99 (0.95-1.05)	0.94 (0.89-1)	1.04 (0.96-1.3)	1.09* (1.03-1.6)	1.15* (1.06-1.24)	1.14* (1.04-1.24)	0.98 (0.89-1.08)	1.1 (0.99-1.3)
Education
No formal education	1	1	1	1	1	1	1	1
Primary education	0.99 (0.93-1.05)	0.97 (0.9-1.04)	1.09 (0.98-1.2)	1.02 (0.97-1)	1.05 (0.95-1.16)	1.08 (0.95-1.23)	1.14* (1.03-1.26)	1.1 (0.94-1.25)
Secondary education	0.94 (0.86-1.01)	0.9* (0.83-0.98)	1.11 (0.98-1.2)	0.98 (0.9-1.05)	0.94 (0.83-1.07)	1.1 (0.96-1.3)	1.19* (1.06-1.35)	1.05 (0.86-1.27)
Graduation and above	0.88* (0.79-0.97)	0.88* (0.8-0.98)	1.26* (1.1-1.4)	0.94 (0.84-1.1)	0.95 (0.81-1.11)	0.92 (0.79-1.1)	1.31* (1.12-1.5)	1.06 (0.82-1.37)
Marital status
Currently married	1	1	1	1	1	1	1	1
Never married	1.1 (0.99-1.19)	1 (0.88-1.1)	0.98 (0.87-1.1)	1 (0.9-1.1)	0.9 (0.84-1.04)	0.89 (0.78-1)	0.83* (0.7-0.96)	1.09 (0.89-1.33)
Widowed/divorced	1.06 (1.98-1.14)	0.99 (0.92-1)	0.99 (0.9-1.1)	1 (0.96-1.1)	1.02 (0.92-1.13)	1.2 (0.99-1.36)	1 (0.89-1.13)	1.12 (0.97-1.3)
Occupation
Homemaker	1	1	1	1	1	1	1	1
Government employee	0.98 (0.8-1.6)	0.92 (0.72-1.2)	1.04 (0.8-1.02)	1.02 (0.84-1.22)	1.1 (0.8-1.43)	1.3 (0.85-2)	0.93 (0.72-1.19)	0.99 (0.7-1.39)
Non-government employee	0.98 (0.93-1.04)	1 (0.95-1)	0.9 (0.8-1)	0.99 (0.92-1.07)	1.1 (0.98-1.27)	0.95 (0.82-1)	0.89 (0.77-1.02)	1 (0.85-1.17)
Self-employed	1.01 (0.97-1.06)	1.05* (0.8-0.99)	0.9 (0.8-1.02)	0.96 (0.91-1.02)	1 (0.94-1.15)	0.93 (0.83-1.1)	0.87* (0.78-0.98)	1.1 (0.94-1.22)
Others	0.91* (0.86-0.96)	0.9* (0.85-0.99)	0.95 (0.83-1)	1 (0.9-1.07)	0.94 (0.84-1.04)	1.05 (0.89-1.23)	0.83* (0.72-0.94)	0.95 (0.82-1.11)

*Significant results: P≤0.05, PR: Prevalence Ratio, CI: Confidence Interval

Discussion

This was a community-based survey to estimate the prevalence of various NCD risk factors using WHO-prescribed STEPS surveillance approach in the rural population of Puducherry district. The prevalence of various risk factors and their association with key socio-demographic characteristics such as gender and age group were determined.

In the study, overall tobacco (11.3%), smoke (9%) and smokeless tobacco (3.5%) use were comparatively less as compared to Global Adult Tobacco Survey (GATS) 2 results of 19.7%, 10.8% and 11.3%, respectively.[15] The overall tobacco use in the study (11.3%) was also less, compared to the studies carried out in southern India which reported prevalence between 16 and 29.93%.[16171819] The smokeless tobacco use (3.5%) in the present study was substantially lower than those studies carried out in the northern parts of India (27.8-48.5%).[202122] Higher prevalence of smoked and smokeless tobacco in southern and northern India could be attributed to varying degrees of cultural acceptability of tobacco use.[23] The decline in the prevalence found in the study shall be viewed in the context that the tobacco use in Puducherry and at country level is witnessing a steady decline since 2010 (GATS 1).[24] This could also be attributed to significant tobacco control measures undertaken in the country since early 2000 such as implementing Cigarettes and Other Tobacco Products Act (COTPA) in 2003,[25] adopting WHO's Framework Convention on Tobacco Control in 2004,[26] National Tobacco Control Program in 2007,[27] and several key amendments from time to time including the latest on increasing pictorial warning on tobacco packs to 85%[28], which is considered as one of the 'best-buy' interventions for combatting NCDs.

The current alcohol use (19.2%) observed in the study was less when compared to the National Family Health Survey-4 (NFHS-4) (2014-2015) data for rural Puducherry (22.95%), but higher than the national prevalence (15.2%).[29] Several studies from southern states of India that followed the STEPS methodology reported increased prevalence ranging between 22.7 and 39.05%.[193031] Studies from northern India also reported higher prevalence.[2032] Lesser prevalence of alcohol use in comparison to NFHS 4 could be due to the methodological differences between NFHS 4 and the current survey.[29] The lesser prevalence observed could also be attributed to the well-established social desirability bias occurring in research surveys on health risk behaviors.

It is well-documented that adequate intake of fruits and vegetables reduces the risk for cardiovascular diseases.[33] On the contrary, it is evident from the current study that a majority (89.8%) of the rural population takes inadequate amount of fruits and vegetables. This phenomenon exists irrespective of geographic regions in the country where nearly nine out of ten people take fruits and vegetables in inadequate amounts.[34] Substantiating the evidence, a national survey revealed inadequate intake of diet by rural population who were also deficient in most of the nutrients.[35] The root cause of this finding shall be attributed to lower economic status, changing dietary patter, and misconceptions about food prevailing among rural population.[3637]

Physical inactivity is one of the key modifiable risk factors for NCDs. In comparison to the studies reported from the other parts of South India, physical inactivity is lower in the present study population.[1718303839] Studies from northern India have reported varied levels of physical inactivity ranging between 14.3 and 33.85%.[364041] Consistent with significantly higher levels of physical inactivity among women, obesity is significantly higher among women (46.7%) compared to men (29.7%). Overweight and obesity levels in the present study are also much higher compared to the other Indian studies.[313236404142] Considering these findings, the extent of pedestrian-friendly neighborhoods, access to parks, open spaces and other social amenities that are proven to promote population physical activity levels[43] need to be explored in the rural setting.

Prevalence of hypertension in rural Puducherry (28.2%) is higher compared to the other studies[16173842] from South India except a study from rural Kerala.[30] The recently concluded NFHS-4 survey reported much lesser (10.15%) prevalence of hypertension in rural Puducherry compared to the current study. One possible reason could be the participation of relatively younger population (15-49 years), and over representation of women (87.2%) in the NFHS-4 survey compared to the present study.[29] The reported prevalence (28.2%) is also higher than the pooled estimate of hypertension prevalence for South India (21.1%) generated using epidemiological evidences till 2013.[39] The higher prevalence, especially reported from a rural population shall be viewed with reference to time points when other studies compared were conducted. Because, during the last 20 years, the prevalence of hypertension in the urban areas have stabilized to about 25-30%, while it increased from 15 to 25% in rural populations of the country.[44] Higher prevalence of hypertension among men has also been reported in the other studies.[3945]

Based on the prevalence of diabetes mellitus from the study (24.4%), the rural population of Puducherry appears to have a higher prevalence compared to the studies from the neighboring states that reported between 7.8 and 21.9%.[161746] Substantiating the higher prevalence, previous evidence from rural Puducherry showed that the incidence rate of diabetes among men was twice higher compared to women.[47] Further, from the study, the prevalence of diabetes becoming comparable to hypertension (24.4% vs 28.2%) indicates the increasing levels of diabetes in the rural populations. One possible reason could be that the dietary profile of the South Indian rural populations is found to constitute majorly of carbohydrate food sources that are higher than the recommended levels per day.[48]

Increased prevalence of hypertension and diabetes in rural Puducherry as compared to the other regions of the country needs to be conceived in cognizance of the fact that these NCDs[3948] are on the rise over the past two decades especially in the states having high- or higher-middle epidemiological transition which includes Puducherry.[49] At this juncture, population-based interventions delivered through National Programme for Prevention and Control of Cancer, Diabetes, Cardiovascular diseases and Stroke (NPCDCS), and recent initiatives, such as population-based screening for common NCDs, and comprehensive primary healthcare and financial protection rendered through 'Ayushman Bharat' are the crucial steps towards addressing the rising burden.[505152] In addition, innovate solutions based on local needs, such as Mohalla (community or neighborhood) Clinics initiative, in the state of Delhi are crucial in making primary care for chronic conditions accessible, equitable and affordable to populations.[53]

In the study, about one-fourth of the population had hypercholesterolemia. This was lesser in comparison to other the Indian studies.[414654] The high prevalence (27.3%) of this risk factor could be substantiated by the higher levels of physical inactivity, obesity, and abdominal obesity in the study population given the established positive association between anthropometric and lipid levels in the individuals.[55]

In response to the rising burden of NCDs in the population, several key programs and policies have been introduced in the country over the last decade. About 13 national health programs that directly or indirectly contribute to health were rolled out during this period while keeping NPCDCS program as central to NCD prevention and control activities in the country.[56] In 2013, India became the first country to adopt the global monitoring framework on NCDs with the aim of reducing premature mortality from NCDs by 25% by 2025.[57] The 10 targets and 21 indicators of the national monitoring framework have provided the national and state governments the necessary direction for devising policies and programs to reduce the burden of NCDs and their risk factors in the population.[58] However, several states are facing challenges such as a lack of political commitment, inadequate manpower, financial and technical constrains, causing ineffective implementation of NCD programs.[565960]

Another major challenge faced in the NCD prevention and control efforts in the country is the non-coordination between departments that directly or indirectly affect health of the populations. Sensitizing the departments, other than health, on their roles in the NCD prevention and control, developing newer multisectoral partnerships, and undertaking joint ventures for NCD prevention and control activities could play a significant role in tacking the rising burden of NCDs.[61] For instance, the food processing sector shall be persuaded to initiate salt reduction strategies, departments for infrastructure development in the states shall be collaborated to design or redesign the available infrastructure in order to facilitate improvement in physical activity of the population.

Limitations

Our study could underestimate a few behavioral risk factors due to socially desirable responses that tend to happen in interview-based surveys. Reporting of behavioral risk factors, especially tobacco, alcohol and physical activity, relied on the participant's recall ability which has its inherent biases. However, use of standardized STEPS questionnaire by the investigator trained in the interviewing techniques including use of show cards helped minimizing these biases. In the study, although GPAQ tool was used, which is an internationally accepted tool for measuring physical activity, there are some concerns on its validity over underestimating the physical activity, specifically those of household and leisure activities in nature.[6263] This might have also contributed to the higher levels of physical inactivity among women in the study.

Conclusion

The study has provided baseline estimates of various NCD risk factors in the rural population of Puducherry. Raised BP, diabetes mellitus and hypercholesterolemia were present in one-fourth of the population. Tobacco and alcohol use were significantly higher among men, whereas insufficient fruits and vegetables intake, raised BP and diabetes mellitus were significantly higher among 45-69-year-olds. With higher levels of education, tobacco and alcohol use decreased, whereas physical inactivity and overweight/obesity increased significantly with increasing education levels. The study highlights the population groups that need to be targeted for public health interventions with population-based interventions undertaken by NPCDCS and those health promotion services supplemented through Ayushman Bharat.

Declaration of patient consent

The authors certify that appropriate patients' consents were obtained.

Financial support and sponsorship

This work was funded by the Jawaharlal Institute of Postgraduate Medical Education and Research (JIPMER) Grant No. JIP/Dean (R)/Intramural/Phs 1/2019-20.

Conflicts of interest

There are no conflicts of interest.