Prevalence and determinants of anaemia among men in rural India: Evidence from a nationally representative survey

Anaemia among men is a significant health issue which has not been given due importance. Only a handful of studies have captured the prevalence of anaemia among men. There is dearth of evidence base on anaemia among men in India. Therefore, this study attempts to fill this research gap by examining the socioeconomic, geographic, health-related, and behavioural differentials of anaemia among rural men in India. We analysed a cross-sectional sample of 61,481 men aged between 15–54 and living in rural areas from the National Family Health Survey (NFHS-5), conducted in 2019–21. Bivariate statistics and multivariable logistic regression were employed to assess the factors associated with anaemia. In rural India, three out of ten men were found to be anaemic. Older men [49–54 years] (Odds Ratio: 1.10, 95% CI, 1.00–1.21), men without a formal education (OR: 1.36, 95% CI, 1.26–1.47), those from Scheduled Tribes (OR: 1.48, 95% CI, 1.39–1.58) and men who belonged to the poorest wealth quintile (OR: 1.24, 95% CI: 1.25–1.35) had a higher risk of anaemia. Men who were underweight were more likely to be anaemic (OR: 1.36, 95% CI: 1.30–1.43). When compared to the central region, men from the eastern (OR: 1.47, 95% CI: 1.39–1.55) parts of India had higher a risk of anaemia. The findings suggest the need to recognise anaemia among men as a public health issue. When developing policy, significant variation in socioeconomic, geographic, health-related, and behavioural factors must be taken into account. Men should also be screened on a regular basis in order to reduce the national burden of anaemia.


Introduction
According to the World Health Organization (WHO), anaemia is a disorder in which the number or haemoglobin concentration of red blood cells is below normal which subsequently results in the decreased oxygen-carrying capacity of blood [1] Haemoglobin is an iron-containing protein in the red blood cells (RBC) that transports oxygen from the lungs to the tissues and carries carbon dioxide from tissues back to lungs [2]. Nutritional deficiencies, particularly iron deficiency is the main reason behind this disease although deficiencies in vitamins B9, Data and methods

Data source
The data comes from the latest round of National Family Health Survey (NFHS-5) carried out by International Institute for Population Sciences during 2019-2021 under the supervision of Ministry of Health & Family Welfare, Government of India. NFHS, the Indian version of the Demographic and Health Surveys (DHS), is a nationally representative cross-sectional survey that collects data on a wide range of demographic, socioeconomic, and health-related issues. Using a two-stage stratified random sampling method, NFHS-5 interviewed 724115 women aged 15 to 49 years and 101839 men aged 15-54 years from 636699 households. Response rate was 97% and 92% for women and men respectively. Through a series of biomarker tests and measurements, the clinical, anthropometric, and biochemical (CAB) component of the NFHS-5 provided critical estimates of the prevalence of malnutrition, anaemia, hypertension, high blood glucose levels, waist and hip circumference, Vitamin D3, HbA1c, and malaria parasites. The survey covered 707 districts from 28 states and 8 UTs. A uniform sample design, which is representative at the national, state/UT, and district level, was adopted in each round of the survey [8].
The Biomarker Schedule contained measurements of height, weight, and haemoglobin levels for children; measurements of height, weight, waist and hip circumference, and haemoglobin levels for women aged 15-49 years and men aged 15-54 years; and blood pressure and random blood glucose levels for women and men aged 15 years and over. Additionally, both men and women were requested to provide a few more drops of blood from a finger prick for laboratory testing for HbA1c, malaria parasites, and Vitamin D3 [8].
We made a request to the DHS Program to provide us with the NFHS data. Once we received the permission to use the data, we downloaded the Men's data file (MR) and the Household Member data file (PR). MR and PR datasets were then merged to avail the information on anaemia among men in India.

Ethics approval and consent to participate
The present study has used secondary data, which is available in the public domain. The dataset has no identifiable information of the survey participants. Therefore, no ethical approval is required for this study.

Sample
Fig 1 depicts the process of sample selection for the present study. Out of the 111,179 eligible men aged 15-54 years selected for the state module, 101839 men who were normally inhabitants and spent the night before the survey in their homes were interviewed. 92820 men consented to have their haemoglobin levels checked. For this study, 31339 out of the 92820 men were excluded as they belonged to urban areas. Our study was limited to remaining 61481 men residing in rural areas.

Anaemia testing
The authors did not collect blood specimens for anaemia testing for this study. These were collected under NFHS-5 by health investigators from eligible men aged 15 to 54 with their consent. Blood samples were drawn from a drop of blood taken from a finger prick (or a heel prick for children age 6-11 months) [8] and collected in a microcuvette, a single-use pipette. Concentration of haemoglobin was analysed on-site with HemoCue Hb 201+ analyser. Introduced in 1990, the HemoCue Hb 201+ is a battery-operated portable device used for quantitative determination of haemoglobin level in undiluted, capillary or venous blood. It converts the haemoglobin into methemoglobin and combines it with azide to form azidemethemoglobin followed by measurement of transmittance and haemoglobin absorbance [22][23][24].

Dependent variable
The dependent variable in this study was whether or not the respondents were anaemic. Men were considered to have anaemia in any form if their haemoglobin concentration was less than 13.0 g/dL, mildly anaemic if it was 12.0-12.9 g/dL, moderately anaemic if it was 9.0-11.9 g/dL, and severely anaemic if it was less than 8.9 g/dL, according to WHO criteria [25]. The study developed a dichotomous variable for prevalence of anaemia. Men with a haemoglobin level less than 13g/dl were considered 'anaemic' and coded '1' while men having a haemoglobin level greater than 13g/dl were classified as 'not anaemic' and coded '0'. We did not take into account the three categories of anaemia: mild, moderate, and severe.

Independent variables
Definitions, categories and codes of independent variables are enlisted in Table 1. A wide range of variables were found to predict anaemia among men [26][27][28]. To illustrate, as a proxy for household income, the wealth index was chosen as a gauge of economic inequality. This is

PLOS GLOBAL PUBLIC HEALTH
Determinants of anaemia among men in rural India Table 1. Description of the variables.

Socio-economic variables
Age (in years) Age of men has five categories: [15][16][17][18][19] (0) Household wealth Household wealth is divided into five categories: poorest, poorer, middle, richer and richest. For the calculation of wealth index, households were given scores based on the number and kinds of consumer goods they own, ranging from a television to a bicycle or car, and housing characteristics such as source of drinking water, toilet facilities, and flooring materials. These scores were derived using principal component analysis. National wealth quintiles were compiled by assigning the household score to each usual (de jure) household member, ranking each person in the household population by their score, and then dividing the distribution into five equal categories, each with 20% of the population [8].
Work status Whether a man currently works or not: no (0) and yes (1) Marital status Marital status has three categories: never married (0), currently married (1), and formerly in union (2) which includes widowed, separated, divorced and deserted.

Health related variables
Body Mass Index It is defined as the ratio between the weight of a man in kilogram divided by the squared height in meter. Body mass index is divided into four categories underweight (<18.5 kg/m 2 ), normal (18.5-24.99 kg/m 2 ), overweight (25-32 kg/m 2 ) and obese (> 32 kg/m 2 ) Blood sugar level It has three categories: normal (0), random blood glucose less than 140 g/dl; high (1), random blood glucose level of 141-160 mg/dl; and very high (2), random blood glucose level of more than 160 mg/dl.

Behavioural variables
Media exposure It refers to how often men watch television, read newspapers, and listen to the radio. Exposure to mass media was defined as men using them every day or at least once a week [coded as yes (1)]. Men who used it less than once a week or never [marked as no (0)] were regarded to have no media exposure.

Consumption of nonvegetarian food
Men's consumption of egg, fish, and chicken/meat. Men who consumed the abovementioned foods every day and at least once a week were classified as eating nonvegetarian food [coded as yes (1)], and men who ate them less than once a week or not at all were classified as not eating non-vegetarian food [coded as no (0)] Alcohol consumption Frequency of drinking alcohol having four categories: never (0), less than once in a week (1), once a week (2), daily (3) Currently smokes Men using any of bidi, cigarette, hookah or cigar were coded as '1' yes and 0 'no' otherwise.
Use of smokeless tobacco Men using forms of smokeless tobacco (gutkha with paan, khaini, nassi) were coded as '1' yes and 0 'no' otherwise.
a measure of household wealth that is determined to be reliable based on both expenditure and income metrics [8]. Wealth index was one of the key predictors of this study. BMI was also one of the significant determinants of anaemia among men which was classified into four categories i.e., underweight (<18.5 kg/m 2 ), normal (18.5-24.99 kg/m 2 ), overweight (25-32 kg/ m 2 ), and obese (>32kg/m 2 ). Age, education, social group, religion, and alcohol consumption were other important factors of anaemia in men. These variables could be categorised into four domains namely socioeconomic factors, community variables, health-related variables and behavioural characteristics. All the variables in present study were selected only after extensive review of literature and according to data availability. Fig 2 depicts a conceptual framework that shows the factors affecting anaemia among men.

Statistical analysis
Bivariate statistics was used to analyse the prevalence of anaemia among rural men by their background characteristics. The analysis was weighted for two-stage sampling design. Thus, weighted estimates were presented. Sampling weights (importance weight: iw) were included in the study. The 'svyset' command was used in Stata to account for clustering at the PSU level. Since our dependent variable was binary in nature, we employed binary logistic regression to assess the effects of the predictor variables on the dichotomous dependent variable of the study. Chi-square test was performed to check if the independent variables were associated with the dependent variable. Only those variables that were found statistically significant

PLOS GLOBAL PUBLIC HEALTH
Determinants of anaemia among men in rural India (p<0.05) were included in the regression models. We applied four models, i.e., model 1 included socioeconomic variables; model 2 included statistically significant variables from model 1 variables and geographic variable. Model 3 contained significant variables from model 2 and health-related variables. The final model included significant variables from model 3 and behavioural variables. The equation of a single-level binary logistic regression model can be specified as: Where, P indicates the probability of an event (prevalence of anaemia in this study), β 0 is the intercept on y axis, βi indicates the regression coefficients associated with the reference group, x i is the independent variable.
The results of logistic regression models are presented in the form of odds ratios with p-values and 95% confidence intervals (CI). We calculated Variance Inflation Factors (VIFs) for the final model to check whether multicollinearity among the independent variables existed. The VIFs for all the independent variables were considerably small (below 2.5) indicating that multicollinearity was not a problem for the models (results not shown). Stata 16 was used for analysing the unit level data [29]. ArcMap (version 10.5) was used to create the choropleth maps [30]. Table 2 present the sociodemographic profile of the men in rural India. About 17% men were aged between 15-19 years. One in every seven men had no formal education. About a quarter of all men belonged to SC group, while 12.4% men belonged to ST group. The Hindu faith was practised by the vast majority of men (81%). Nearly one-fourth men belonged to the poorest wealth quintile. Around 30% men were from the eastern region of India. At the time of the survey, roughly two-thirds of men were married. About 18% of men were underweight (BMI less than 18.5 kg/m2). Around 28% of men used smokeless tobacco, while 45% smoked cigarettes. Table 3 depicts the prevalence of anaemia among rural men in the country by various background characteristics. Overall, about one-fourth men in India were found to have anaemia. One out of every five urban men while three out of every ten rural men were anaemic in India (Fig 3). The prevalence of anaemia among rural men was highest in the age group 50-54 (34.1%) followed by the age group 15-19 years (33.8%). Men aged 20-29 years (22.9%) had the lowest prevalence of anaemia. Prevalence of anaemia decreased with increase in education. Men with no education had the highest prevalence of anaemia. ST men (30.9%) showed the highest prevalence of anaemia among the social groups. Anaemia prevalence was significantly higher among Muslim men and lower among Christian men. A steady decline was observed in the prevalence of anaemia with increase in household wealth. About one-third rural men belonging to the poorest households had anaemia.

Differentials in prevalence of anaemia by background characteristics
Anaemia prevalence was highest in the eastern region (34.1%) while lowest in the southern region (18.5%). The north, west, north-west and central regions reported 27.2%, 28.9%, 26.9% and 25% prevalence of anaemia, respectively. Anaemia was inversely related with BMI as prevalence of anaemia was 34.7% among underweight men versus 19.3% among men who were overweight. Men who drank alcohol daily and used smokeless tobacco had slightly higher occurrence of anaemia than who did not consume it.

Estimates from logistic regression analysis for anaemia among men in rural India
The results from the logistic regression models are presented in Table 4. The men within the age bracket 20-29 years and 30-39 years were 33% and 26% less likely to be anaemic than men aged 15-19. Men aged 50-54 years were slightly more likely to be anaemic (OR: 1.10, 95% CI, 1.00-1.21). Men with no formal education were 36% more likely (95% CI, 1.26-1.47) to be anaemic than men who obtained higher education. Men with primary education were a quarter times more likely (95% CI, 1.15-1.34) to have anaemia as compared to men with higher education. Men from to ST category (OR: 1.48, 95% CI, 1.39-1.58) had significantly higher likelihood of being anaemic as compared to men of 'Others' category. The odds of being anaemic were 36% higher among Muslim men (95% CI, 1.27-1.45) and 48% lower among Christian men (95% CI, 0.48-0.57) as compared to Hindu men. The more the wealth, the lesser the risk to suffer from anaemia. Men from the richest wealth quintile were 29% less likely to suffer from anaemia (95% CI, 0.66-0.78) than those from the poorest wealth quintile. The odds of anaemia among men from the east region of the country were 47% (95% CI, 1.39-1.55) higher than those from the central region. Men belonging to the north, west and north-east regions were 27%, 28%, and 24% more likely to suffer from anaemia. However, the odds of anaemia among men were lower by 19% (95% CI, 0.76-0.87) in the south region. Men who were underweight had 36% (95% CI, 1.30-1.43) more likelihood of being anaemic whereas obese men were 23% (95% CI, 0.68-0.86) less likely to suffer from anaemia as compared to men with normal BMI. The risk of anaemia among men using smokeless tobacco was more (OR:1.15, 95% CI, 1.08-1.22) than those not using the same.  unequal distribution of an element within a geographic area is depicted through gradients of the same colour. The higher the value or prevalence, the darker the shade [31]. The states as well as districts of India were classified into six categories according to prevalence of anaemia (%). West Bengal, Tripura, Assam, UT Jammu and Kashmir (>35%) were found to have highest anaemia prevalence among rural men followed by Bihar, Jharkhand, Chhattisgarh, Odisha, Gujarat (28%-35%). South Indian states i.e., Andhra Pradesh, Karnataka, Tamil Nadu, and Kerala,) as well as Manipur and Nagaland showed lowest prevalence.

Spatial analysis
Districts of West Bengal, Bihar, Odisha, Chhattisgarh, Jammu & Kashmir (>50%) showed the highest prevalence of anaemia among rural men followed by some districts of Uttar Pradesh and Madhya Pradesh (40%-50%). The lowest prevalence was found in some districts of Manipur, Nagaland, Karnataka, and Tamil Nadu (<10%).

Discussion
An estimated one-fourth of Indian rural men aged 15 to 54 years were found to be anaemic. In the last four years, the prevalence of anaemia has only risen [8]. The findings suggested that the prevalence of anaemia varies by sociodemographic characteristics among rural men. Rural men who were 49-54 years old, had no formal education, belonged to ST group, were Muslim, or from the poorest wealth quintile, lived in the eastern region, were underweight, and consumed alcohol and smokeless tobacco on a daily basis were more likely to have anaemia. At both the state and district levels, there was significant geographical variation in the prevalence of anaemia.
Older men, aged 50-54 years, were more likely to be anaemic, followed by adolescents (15-19 years). However, the prevalence of anaemia was lowest in the 20-29 age group. Older men are more vulnerable to anaemia, possibly as a result of suffering from other chronic diseases such as diabetes, chronic kidney diseases, hypertension [32]. Previous research has yielded similar results [26,33], although none of these studies have been conducted specifically on rural men. Education level emerged to be a significant determinant of anaemia. Rural men with no education were most vulnerable to anaemia. Men with a higher level of education were less likely to develop anaemia. Previous studies also corroborate the same [19,34]. Disease awareness and knowledge, as well as the necessity of sanitation and health care, are raised through education. It also encourages people to listen to and accept the advice of health professionals [35,36].
Men in the ST category were more likely to be anaemic than men from other social groups. STs have a long history of being marginalised, with the majority of them still living in remote areas of the country. As a result, poor diet and a lack of access to healthcare services could be linked to their increased risk of anaemia [37]. Various studies have found that SCs and STs generally have poor health outcomes, underscoring the importance of caste prejudices. Despite affirmative actions by the Indian government post-1947, people from SC/ST groups remain deprived in a variety of areas, including health [38][39][40].
Muslim men had greater risk of anaemia, which is similar to the findings from an earlier study [20]. This study also found that Christian men living in rural India were at significantly lower risk of anaemia, which was not highlighted previously by any scientific study and requires further investigation. However, these results need to be cautiously interpreted as the sample size for Muslim and Christian population was small and sample distribution was highly skewed.  Household wealth was strongly associated with anaemia. The wealthier the household, the lesser the risk of anaemia. Several researches have explored the linkage between poverty and malnutrition. Poverty makes it difficult for people to eat a healthy diet and get health care [41,42]. Low socioeconomic position can exacerbate the prevalence of anaemia in a variety of ways, including a poor living and working environment, unhealthy habits such as smoking and limited access to health care, and a lack of health literacy [40]. In developing countries, people from poorer households are more likely to suffer anaemia than those from wealthy homes [43] due to factors such as substandard housing, hunger, and increased disease exposure.
A significant geographical variation in the prevalence of anaemia among rural men was also noted in this study. The likelihood of being anaemic was maximum among men belonging to the eastern region. A recent study on anaemia among men also pointed out higher prevalence of anaemia among men in the eastern India [21]. A previous study on anaemia among children also highlighted that people from central and eastern region were associated with higher risk of being anaemic [44]. However, this study found that men from the north, west and north-east region were more likely to suffer from anaemia compared to men in the central region. The differentials in population composition from one region to another could play a vital role in spatial variation in anaemia. Anaemia was significantly associated with BMI in this study. Rural men with lower BMI, who were underweight, had a higher risk of anaemia. It is a well-researched fact that underweight persons have higher likelihood to be anaemic, as low BMI is caused by a lack of a balanced and healthy diet [45]. Men who drank alcohol on a regular basis had a higher risk of anaemia than men who did not consume alcohol. A population based study on India offered a similar finding [19]. Frequent consumption of alcohol leads to deterioration of health and a number of other chronic illnesses which are linked with anaemia.
There was no significant relationship found between anaemia prevalence and media exposure, blood sugar, and non-vegetarian food consumption. Previous studies have found a link between anaemia and these variables. It has been noted that as blood sugar levels rise, the likelihood of being anaemic also rises [46]. Another study on women in Afghanistan found a strong negative correlation between anaemia risk and meat consumption frequency [47]. Since the above-mentioned study was conducted on women, we should exercise caution while comparing the results of the current study with the findings from other studies. It was also found that women who took iron tablets or syrup at regular intervals had a lower risk of anaemia [48]. Thus, further research is needed to investigate the effect of such interventions in men.
The Indian government has devised a number of programmes and policies aimed at reducing the prevalence of anaemia in the country. Almost all of them, however, primarily target women of reproductive age and children. The National Nutritional Anaemia Control Program (NNACP), for example, was established in 1970 with the goals of encouraging regular consumption of iron-rich foods, providing iron and folate supplements to susceptible groups, and identifying and treating severely anaemic patients [49]. The 12-by-12 Initiative (2007) was launched in collaboration with the Ministry of Health and Family Welfare, the World Health Organization (WHO), UNICEF, and the Food and Agriculture Organization of the United Nations, with the goal of every child having a haemoglobin level of 12 grams by the age of 12 years by 2012 [50]. The National Iron Plus Initiative (NIPI) was launched in 2013 with the goal of providing free iron and folic acid supplements to adolescent boys and girls (10-19 years) as well as women of reproductive age, such as pregnant and breastfeeding women. The Union Ministry of Health and Family Welfare started the Weekly Iron and Folic Acid Supplementation (WIFS) initiative in 2013 under the National Health Mission to prevent anaemia among teenagers (NHM). Another initiative, Anaemia Mukt Bharat (2018), aimed to cut anaemia in young children, teenage boys and girls, pregnant and breastfeeding women, and women of reproductive age by half [50]. It is therefore recommended that the government and policymakers expand the scope of programs such as Anaemia Mukt Bharat and Iron supplementation programs to include sub-strata of rural aged men who are more susceptible to anaemia. Additionally, more research is needed to design such health interventions for men, keeping the sociodemographic and cultural context in sight.
This study has a few limitations that should be mentioned. First, because the data for this study comes from a cross-sectional survey, the relationship between dependent and independent variables demonstrated in this paper should be interpreted as association, and not as causality. Second, the model in this study uses only those variables that were available in the dataset. Some predictor variables may have been left out which may have resulted in what is known as omitted variable bias. We were unable to include folate, vitamin B12, or vitamin A intake as predictor variables in the model due to a lack of data. Third, the nationwide representative survey measured haemoglobin concentrations with a battery-operated portable Hemo-CueHb 201+ analyser, which may have underestimated the results when compared to laboratory testing (Didzun et al., 2019). Future research should take these limitations into account to get a more accurate and comprehensive picture of the prevalence of anaemia among rural men in India.

Conclusion
Anaemia among rural men, just like it is among women and children, is a serious public health concern in India. Anaemia was found in three out of ten rural men. High-risk groups were older men, men without education, Muslim and STs, men from the poorest households, and men who were underweight. The benefits of existing programs and policies related to anaemia eradiation should be extended to men as well. In addition, targeted interventions among susceptible groups of rural men are advised as a way to reduce the prevalence of anaemia. Men's haemoglobin levels should be checked on a regular basis and for that purpose appropriate screening facilities should to be made available closer to their residences so that they can be screened easily. When developing policies, it is important to keep geographical regions with high anaemia prevalence in mind. A comprehensive strategy based on the aforementioned proposals could be beneficial to reduce burden of anaemia among men in rural India.