1. INTRODUCTION

Anemia not only causes growth retardation of fetus, puts pregnant women at high risk of complications, causes poor growth and development among children, but also affects adults with a significant effect on productivity and economic growth [1]. A study that analyzed the cost of iron deficiency in 10 developing countries found that the median value of physical productivity losses per year due to iron deficiency was about US$2.32 per capita, or 0.57% of the gross domestic product and combining the effect of iron deficiency on cognition, total economic loss was estimated to be US$16.78 per capita or 4.05% of the GDP [2]. As per the World Health Organization’s (WHO) estimates globally 12.7% of the adult males suffer from anemia [3]. In India, the fourth round of National Family Health Survey (NFHS-4) reported that 22.7% of the males in the age group 15–49 years had anemia [4]. The prevalence of anemia among males of Haryana in this age group had shown an increase from 17.6% in 2005–2006 to 20.1% in 2015–2016 [5]. Despite high burden and increasing trend of anemia, both in terms of prevalence and economic impact, there is paucity of information on anemia among adult males; and consequently, this age group is often not included in the anemia control programs.

Dietary deficiency plays a major role in causation of anemia. Dietary intake is influenced by cultural factors as well as perceived benefits of a particular diet. Studies on pregnant women and adolescent girls reported that though they were aware that poor quality or lack of diet causes anemia, their knowledge regarding specific dietary elements in preventing anemia was limited [6,7]. Economic constraints on dietary improvement were cited as the primary barrier to a food-based approach for alleviating symptoms of anemia. In many cases, respondents were also not aware of the serious consequences of anemia [7]. However, there is limited information about the attitude and beliefs of rural adult population, especially men, regarding anemia and the role of various dietary factors in its causation. Hence, this study was planned to estimate the prevalence of anemia among adult males (≥18 years) in a rural area of the Ballabgarh block of Haryana, India; its association with various sociodemographic, clinical, and dietary factors; and to explore the knowledge, attitude, and beliefs of adult males toward anemia.

2. MATERIALS AND METHODS

This study was conducted in Ballabgarh block of district Faridabad in the northern Indian state of Haryana. The total population of the study area was 98,000 in the year 2016. About 65% of the population in the study area was aged 18 years or above. A detailed description of the study area is available elsewhere [8]. The study was conducted between August 2016 and December 2016. Participants who refused to give consent for hemoglobin estimation were excluded from the study.

Mixed methods approach was adopted. A community-based cross-sectional study was conducted to assess the prevalence of anemia and to collect information on various sociodemographic, clinical, and dietary factors. Participants were recruited through simple, multistage random sampling. The sample size was calculated to be 1200 based on assumed prevalence of anemia as 20% [5], relative error of 20%, and a design effect of 3. Four of the 28 villages in the study area were selected randomly, and within each village an equal number of participants was selected by simple random sampling. Villages in the study area are similar in sociodemographic profile, and a sample was selected randomly within the village that was likely to be heterogeneous, an arbitrary design effect of 3 was taken as no information was available on the intracluster coefficient [9]. A Health Management Information System (HMIS) is maintained at Comprehensive Rural Health Services Project (CRHSP), Ballabhgarh with demographic and other details of all individuals in the area. A list of inhabitants and their details were obtained from HMIS. Trained graduate interviewers made home visits and interviewed the selected participants. Interviewers were trained for a day by one of the authors. However, no formal assessment of adequacy of training was undertaken. If the selected participants were not found at home in two visits, they were excluded from the study.

A pretested questionnaire in the local language, Hindi, was used to collect information on various sociodemographic, clinical, and dietary variables. Dietary history was assessed through a pretested questionnaire, which was adapted from the WHO Stepwise Approach to Surveillance tool for noncommunicable diseases, and included the frequency of consumption of green leafy vegetables, fruits, and dairy products. Dietary assessment could not be performed in one of the four villages due to logistic reasons. Current tobacco and alcohol users were defined as those who reported consuming any tobacco or alcoholic product during the last 7 days. The questionnaire was first prepared in English, which was then translated into the local language Hindi in which it was to be administered. The questionnaire was translated back into English to remove any error and to assess the quality of translation. Although there is no information on the prevalence of intestinal nematodes in the study area and among adult men in a similar setting, a study performed in a nearby area found the prevalence among school children to be 29.7% [10].

Focus Group Discussions (FGD) with adult male participants were conducted. Purposive sampling was performed to recruit participants for FGDs. Attempts were made to keep the groups homogeneous with regard to age and socioeconomic status. An FGD guide was used to elicit participant’s knowledge, attitude, and beliefs regarding anemia and their dietary determinants. An FGD guide was prepared by the study investigators who facilitated the discussion. Domains for discussion included knowledge regarding anemia, its causes, symptoms, severity, role of diet and specific dietary elements in preventing anemia, and barriers and facilitators in dietary consumption of nutrients that may prevent anemia. FGD guide was pre-tested in the non-study area. A total of six FGDs were conducted and was facilitated by study investigators.

HemoCue^® Hb 201 DM System (HemoCue AB, Sweden) was used to estimate the concentration of hemoglobin in capillary blood. Hemoglobin determined by the HemoCue method has been reported to be comparable with that determined by the other methods and has been recommended for on-the-spot estimation [11]. A single drop of blood was collected from a finger prick after removing the first two drops of blood to ensure that the sample was fresh capillary blood, and was placed into a cuvette for measurement. A hemoglobin level <13 g/dL was defined as anemia; that between 11 and 12.9 g/dL as mild anemia; between 8 and 10.9 g/dL as moderate anemia; and a hemoglobin level less than 8 g/dL as severe anemia [12]. All investigators were trained for a day in various aspects of the data collection. Data collection was performed under the supervision of a senior staff. Double data entry was done for 5% of the forms. Appropriate checks were performed in the data entry platform. All HemoCue instruments were recalibrated before the study.

Quantitative data were entered in Epi Info version 7.2 (CDC Atalanta, Georgia). Forms were reviewed for completeness before data entry. Data analysis was done using Stata version 12 (StataCorp, Texas, USA). Prevalence of anemia was reported as its proportion along with 95% CI. Age standardization of the prevalence of anemia was performed using the age structure of rural population of district Faridabad from the year 2011 [13]. Prevalences of mild, moderate, and severe anemia were reported for various subgroups. Possession of household consumable items, land and house ownership, source of drinking water, type of toilet facility, type of cooking fuel used, and type of electricity connection were combined into a composite wealth score using principal component analysis. It was transformed into wealth scores by computing the scores of the first component, which explained 41% variance. Wealth index constructed from these variables has been used in large-scale demographic and health surveys in India such as NFHS-4 [14]. Data from earlier NFHSs with similar variance from the first factor have been used to construct the socioeconomic indicator [15]. The study sample was divided into five wealth quintiles. Association of anemia with various sociodemographic, clinical, and dietary factors was assessed by logistic regression. In dietary factors and comorbidities, different categories were merged to improve the statistical power. Factors that had p-value <0.3 in bivariate logistic regression were adjusted through multiple logistic regression to assess the association between various factors and anemia. P-values of linear predicted value and squared linear predicted value were estimated to assess model assumptions. Pseudo R² and Hosmer–Lameshow test was used to test the goodness-of-fit. The level of significance for all analysis was set at 0.05. All FGDs were tape recorded. Recorded information was transcribed using standardized transcription forms. Qualitative data were categorized in relevant domains and themes were identified from the transcribed information.

Written informed consent was obtained from each subject. Information collected was kept confidential. Analyses were performed after removing identifiers. The result of hemoglobin testing was provided to the participant on the spot. Those found anemic or with other morbidity were offered treatment through local primary health centers. Ethical approval for the study was obtained from the Institute Ethics Committee of All India Institute of Medical Sciences, New Delhi (web address: https://www.aiims.edu/en/about-us.html, http://14.139.245.45:8082 approval No: IEC-378 dated August 1st, 2016).

3. RESULTS

A total of 1226 participants were approached, of which nine (0.7%) refused to give blood for hemoglobin estimation. The mean age of study participants was 39 years (SD: 15.6 years); most of the participants (53.9%) were in the age group 25–49 years. Elderly participants (aged ≥60 years) constituted 14.3% of the total study population. More than one-third of the study participants were educated for 11 or more years, and 38.8% were employed in a government or private organization Table 1. About 4% of the participants reported ever being diagnosed with anemia. About 27.2% reported presence of a comorbidity, 6.5% of which had renal disorder, liver disorder, malignancy, or external bleeding. More than half of the study participants (51.2%) were currently using tobacco whereas 39.3% reported using alcohol. Washing hands after defecation and before having food was reported by 92.2% and 89.7% of the participants, respectively. Whereas 82% of the participants consumed milk or other dairy products daily or more than once weekly, only 43% consumed fruits daily or more than once weekly.

Unadjusted prevalence of anemia among study participants was 28.8% (95% CI: 26.3–31.4%); 22% had mild, 6.6% had moderate, and only 0.2% had severe anemia. Age adjusted prevalence of anemia was 27.9% (95% CI: 27.4–28.4%); age adjusted prevalences of mild, moderate, and severe anemia were 21.3%, 6.4%, and 0.2%, respectively. Prevalence of anemia was 46.8% (95% CI: 39.4–54.3%) in participants of age 60 years or more and was 23.7% (95% CI: 20.9–26.5%) in participants aged 18–49 years. It was found to be 22.1%, the lowest among those who were 11 years or more educated and 25.5% in participants who were self-employed Table 1. About 45% of the participants who reported being diagnosed either with renal disorders, liver disorders, malignancy, or bleeding per rectum were found to be anemic. One-quarter of the study participants who consumed milk daily or more than once weekly were found to be anemic Table 2.

Anemia was found to be positively associated with age [OR: 3.7 (2.4–5.8)] in the age group 60 or more as compared with 18–24 years), and self-reported history of renal disorder [OR: 3.2 (1.2–8.7)], and malignancy [OR: 2.5 (1.0–6.3)] on bivariate analysis. However, on multiple logistic regression, age was found to be significantly associated with anemia with odds ratio of 1.6 (1.1–2.3) in the age group 25–49 years, 3.1 (1.8–5.3) in the age group 50–59 years, and 4.1 (2.5–6.7) in age group ≥60 years as compared with age group 18–24 years and presence of one of the four chronic morbidities – renal disorder, liver disorder, malignancy, or bleeding per rectum [OR: 1.9 (1.2–3.2)]. p-values for linear predicted value and squared linear predicted value were found to be <0.01 and 0.41, respectively, suggesting that there was no specification error in the model. Pseudo R² for the model was 0.056 and p-value for Hosmer–Lemeshow test was 0.55. Anemia was also found to be significantly associated with age when age was used as a continuous variable; odds ratio in the bivariate analysis was 1.03 (1.02–1.04) and when adjusted using variables in the above model odds ratio remained the same, i.e., 1.03 (1.02–1.04) Table 3.

4. DISCUSSION

This was one of the few studies that attempted to assess the prevalence of anemia and its association with selected sociodemographic, clinical, and dietary factors among adult males. Data on the burden of anemia among adult population is limited and conflicting. A study conducted by the National Nutrition Monitoring Bureau in 2006 in rural areas of nine states of the country with a much larger sample size reported a comparatively higher prevalence of anemia, i.e., 54.8% among men aged 20 years or more [16]. This study used cyanmethemoglobin method for estimating anemia. This study also found a positive association between various sociodemographic constructs such as type of house, electrification status, source of drinking water, and type of cooking fuel used. We combined all these constructs as well as other measures of wealth such as household possession of consumer items, type of toilet used, and land ownership in a composite wealth index. However, we did not find statistically significant association of wealth index with anemia despite prevalence being higher among participants in the lower wealth quintiles. One of the reasons could be the limitation of the wealth index in this case to accurately decipher household inequalities as the first factor explained only 41% variance. Another study in a rural area in Haryana among men aged 16–70 years conducted almost two decades back in the year 1994–1995, wherein hemoglobin was estimated using the cyanmethemoglobin method, reported the prevalence of anemia to be 44.3% [17]. The study also reported anemia prevalence to be higher among people with lower educational status, smokers, and of a lower socioeconomic status. Another study that assessed anemia among adult males aged 15–45 years in the desert areas of Rajasthan, India reported anemia prevalence by way of clinical examination to be 43.9% [18]. However, another study in rural Andhra Pradesh in the year 2007 reported the prevalence of anemia among adult males to be only 14.8% using Hemo Control photometer, which measures hemoglobin using a similar principle as HemoCue [19]. This study also reported a higher prevalence of anemia among people engaged in farming and found a positive association between age and anemia. NFHS-4 reported the prevalence of anemia among males aged 15–49 years in Faridabad district to be 14.7% [20], which is lower than that found in our study that reported the age adjusted prevalence in age group 18–49 years to be 23.2%. This difference could be because of a lower prevalence of anemia in urban population, which was included in NFHS-4.

Our study corroborated evidence from elsewhere regarding the association between chronic morbidities and anemia. Although hand-washing has been shown to reduce the prevalence of anemia among adolescents and children [21,22], we did not find any association between hand-washing and anemia, possibly as helminthiasis does not seem to play a major role in causation of anemia among adults. A diet rich in phytates and calcium, which slow down iron absorption and predispose humans to anemia [23]. However, we did not find any association between the intake of green leafy vegetables, fruits, and milk products and anemia. This is possibly because of the qualitative assessment of their intake, which was likely to be imprecise and was a major limitation in finding the association. We also did not consider gathering detailed information on dietary history, which would have allowed more accurate assessment of their nutrient intake. Studies on pregnant women and adolescent girls have found lack of knowledge regarding specific dietary elements in preventing anemia [6,7]. Economic constraints to dietary improvement were cited as the primary barrier to a food-based approach to alleviating symptoms of anemia [7]. In our study also, the knowledge and attitude toward dietary factors of anemia were conflicting. Although study participants were aware of the importance of good diet and green leafy vegetables, they also said that a diet rich in dairy products is important to prevent anemia. Easy availability of dairy products and unaffordability of green leafy vegetables and fruits were cited as the main reasons for low consumption of green leafy vegetables. Qualitative assessment of dietary intake and assessment of morbidity by self-reporting were few of the major limitations of our study. Furthermore, dietary estimation could not be done in one of the villages that might affect the estimates. Definition used for current smoker and alcohol use might overestimate their prevalence.

5. CONCLUSION

Despite a decline in prevalence of anemia among rural population over the years, anemia remains a significant public health problem, as defined by the WHO [24]. Elderly population was found to be at a higher risk of anemia. There were misconceptions among the study participants about the role of specific dietary factors in causation of anemia and poor purchasing capacity was cited as the main reason for not adhering to iron-rich diet.

CONFLICTS OF INTEREST

None declared.

AUTHORSHIP DECLARATION

Shashi Kant, Sumit Malhotra, Ravneet Kaur, Partha Haldar declares conceptualization, protocol and tools development, writing and approval of the manuscript. Rakesh Kumar declares conceptualization, protocol and tools development, data analysis, writing and approval of the manuscript.

All authors are in agreement with the manuscript and declare that the content has not been published elsewhere.

FUNDING SOURCE

This study was not funded.

ETHICAL APPROVAL

Institute Ethics Committee, All India Institute of Medical Sciences, New Delhi, India.

ACKNOWLEDGEMENT

We acknowledge the support of participants of this study and students and residents of CRHSP, Ballabgarh in conducting this study.