Determining factors for the prevalence of anemia in women of reproductive age in Nepal: Evidence from recent national survey data.

Anemia or low concentration of hemoglobin (Hb) is a condition in which the number of red blood cells of the body is insufficient to meet physiological needs [ 1 ]. Iron deficiency is thought to be the most common cause of anemia worldwide. Children, women of reproductive age, and pregnant women are at high risk of developing anemia [ 2 , 3 ]. Maternal anemia is associated with maternal and child morbidity and mortality such, as increased risk of miscarriage, stillbirth, prematurity, and low birth weight of the baby [ 4 ]. About 20% of perinatal mortality and 10% of maternal mortality in developing countries is attributed to iron deficiency [ 5 ]. In 2005, globally, 1.62 billion people were estimated to be affected by anemia [ 2 ]. Similarly, as per the 2011 estimates, the global prevalence of anemia among pregnant women is about 38% (32.4 million pregnant women), and non-pregnant women is about 29% (496.3 million non-pregnant women)), and for all women of reproductive age is about 29% (528.7 million women of reproductive age) [ 3 ]. The prevalence of anemia among non-pregnant and pregnant women is relatively low in high-income countries compared with low-and-middle-income countries [ 6 ]. The World Health Organization (WHO) classified the public health significance of anemia based on the prevalence estimated from blood levels of hemoglobin (normal, <4.9%; mild, 5.0–19.9%; moderate, 20.0–39.9%; and severe, ≥40.0%) [ 1 ]. Numerous factors including age, sex, residential elevation (altitude), smoking behavior, and pregnancy status influence hemoglobin concentration [ 1 ]. In addition to being a medical condition, anemia is an important socio-economic issue given its association with decreased physical and cognitive productivity [ 7 ]. A complex interplay of political, ecological, social, and biological factors determines the prevalence and distribution of anemia in a population [ 8 ]. Previous studies have documented several potential causes of anemia among women including rural residency [ 8 ], younger age [ 9 ], pregnancy status [ 10 , 11 ], lower nutritional status [ 9 ], repeated childbearing [ 12 ], lactation/breastfeeding [ 9 , 10 ], poor access to nutritional supplements during pregnancy [ 9 ], and exposure to domestic violence [ 13 ]. Additionally, helminths infection [ 14 ] and malaria [ 15 ] were found to be important causes of anemia. Furthermore, anemia was found to be associated with immunologic disease progression and increased risk of AIDS-related death [ 16 ]. On the contrary, the use of hormonal contraceptives was shown to have a potential protective effect against anemia [ 17 ]. Research on anemia in Nepal has mostly focused on adolescents [ 18 – 21 ], pregnant [ 22 ], and relatively younger women (aged 13–35 years) [ 23 ]. A majority of these studies are small-scale and are limited to a specific region. Moreover, the association of women’s decision-making autonomy and violence with the prevalence of anemia has not been investigated yet. The Nepal Demographic and Health Surveys (NDHS) is the only source of national data on several characteristics of women of reproductive age (15–49 years). Therefore, we designed this study to determine the prevalence of anemia and the factors associated with the risk of developing anemia among women of reproductive age in Nepal. Additionally, we examined the association of women’s decision-making autonomy regarding healthcare and experience of intimate partner violence (IPV) with anemia.

The baseline demographic characteristics of the study population are shown in Table 1 . A total of 6,414 women aged 15–49 years (mean [±SD] age, 29.2 years [±9.6]) were included in the study. The highest proportion of the women were from the 15–24 years age group (38%), from janajati ethnicity (36%), residing in urban area (63%), had secondary education (35%), was involved in agriculture (47%), belonged to rich family (43%), and used tap water (48%). The majority of the women were fecund (79%), had not given birth in the past three years (78%), were not breastfeeding (78%), and were not using any contraceptive (60%). About one-third of the women (33%) had ≥3 children. Of the 4,572 women with children, the mean (± SD) age at first birth was 19.7 years (± 3.2), and the majority of them (54%) were <20 years when they gave the first birth. Of the 4,651 women who were ever pregnant, the majority (87%) had no adverse pregnancy outcomes in their last/most recent pregnancy. Of the 2,014 women who had a live birth in the past five years, about two of the five women (40%) took iron supplements for >180 days during their last pregnancy. The mean (±SD) BMI was 22.2 (±4.0) and approximately three of the five women (61%) had a normal BMI. The majority of the women (91%) were non-smokers. n: number. %: percentage. Hb: hemoglobin. BMI: body mass index. # The numbers and percentages are adjusted for multi-stage sampling, cluster weight, and sample weight. a (Mean ± SD). b (mild, moderate, and severe anemia combined). c Education level was classified as no education (no years of schooling), primary (up to grade 5), secondary (up to grade 10), and higher (higher than secondary level). d Household wealth status was determined using scores derived from principal component analysis of various household possessions, assets, and amenities. e BMI was categorized as underweight (<18.5 kg/m 2 ), normal (18.5–24.9 kg/m 2 ), overweight (25.0–29.9 kg/m 2 ), and obese (≥30.0 kg/m 2 ). f Hb level adjusted for altitude and smoking. g Includes the sum of mild, moderate and severe anemia. The mean (±SD) hemoglobin level concentration was 12.1 g/dl (±1.4). An overall prevalence of any anemia among women aged 15–49 years was 41% (95% CI, 38.6%–43.0%). Specifically, mild, moderate, and severe anemia was found in 33% (95% CI 31.5%–35.5%), 7% (95% CI 6.2%–7.9%), and 0.3% (95% CI 0.1%–0.5%) of the women, respectively. The prevalence of anemia decreased with the increase in both the age group (i.e., 44% among women aged 15–24 years vs. 37% among women aged 35–49 years), and age at first birth (42% among women who had first birth <20 years compared to 35% among women who had first birth ≥25 years). The highest prevalence of anemia was found among women: residing in rural area (43%), having secondary education (43%), unemployed (43%), of middle-class family (49%), having wells as the sources of drinking water (53%), postpartum amenorrhoeic (48%), had birth in the past three years (46%), breastfeeding (46%), not using any contraceptive methods (42%), underweight (48%), and non-smoker (42%) ( Table 1 ). Table 2 shows the unadjusted and adjusted logistic regression analyses of the factors associated with the risk of developing anemia among women of reproductive age. Having well water as the source of drinking water (aOR 1.93; 95% CI 1.58–2.37) was significantly associated with an increased risk of developing anemia. Compared to women who did not use any contraceptives, women who used hormonal contraceptives (including intrauterine devices) had reduced odds of developing anemia by 37% (aOR 0.63; 95% CI 0.52–0.76) while, having permanent sterilization was associated with a 146% increase in odds of developing anemia (aOR 1.46; 95% CI 1.20–1.78). Overweight/obese women had reduced odds of developing anemia (aOR 0.64; 95% CI 0.51–0.79) compared to women with normal body weight. Similarly, women who were smoker had lower odds of developing anemia (aOR 0.68; 95% CI 0.55–0.84) compared to non-smoker women. Factors such as age, ethnicity, household wealth status, fertility status, having a birth in the past three years, and breastfeeding did not reveal any significant association with the risk of developing anemia despite their significant relationship in the univariate analysis. a Adjusted model consisted of 6,410 observations due to some missing values. b Only variables with p<0.05 are retained in the adjusted multivariate model. OR: odds ratio. CI: confidence interval. All values are weighted for the multi-stage sampling, cluster weight, and sampling weight. Of the 3667 women who were married or in-union women at the time of the survey, more than one-third (37.8%) were not involved in decision-making regarding their healthcare and less than one-fourth (23.9%) have ever experienced physical or sexual IPV. In the unadjusted model, women who took the decision jointly with their partners regarding her healthcare and who experienced IPV were significantly associated with the increased risk of anemia. However, in the multivariate analysis, after adjusting the model for background characteristics (socio-demographic, reproductive, nutritional and behavioral characteristics), women’s decision-making autonomy regarding her healthcare and experience of IPV were not significantly associated with an increased risk of developing anemia ( Table 3 ). n: number, %: percentage. a Adjusted model consisted of married/in-union women (3667). b Final model adjusted for age, ethnicity, household wealth status, drinking water source, maternity status, birth in the past three years, breastfeeding, contraception use, BMI, and cigarette/tobacco smoking. OR: odds ratio. CI: confidence interval. All values are weighted for the multi-stage sampling, cluster weight, and sampling weight.

This study demonstrates the factors associated with the prevalence of anemia in Nepal. It is quite evident from the study findings that the general health status of the women of reproductive age in Nepal needs substantial improvement. This study found that about 41% of the women of reproductive age was anemic. Thus, anemia could be considered as a severe public health problem. The higher prevalence of anemia revealed in this study compared to the past survey highlights the needs to reconsider the existing nutritional policy in Nepal. The prevalence of anemia was higher among underweight women, while obese women were less likely to become anemic. Both, anemia and lower BMI are the indicators of poor nutritional status. Poor diet quality may result in the lower bioavailability of dietary iron, thus increasing the risk of developing anemia. An overall higher prevalence of anemia suggests the necessity of developing appropriate nutrition interventions; iron supplementation could be the immediate measure to cure anemia while periodic deworming, food diversification, and food fortification might be the long-term measures for the management of anemia. Given that a larger proportion of households are using well-water as the source of drinking water, appropriate measures to protect the well-water such as protecting the wells and disinfection of water before use may decrease the likelihood of parasite infestation, and in turn, reduce the prevalence of anemia. Mobilization of community health workers and volunteers can be a strategic measure to reach a large section of adolescents including women of reproductive age. Anemia surveillance and interventions programs specific to targeted groups and high-risk populations are recommended for reduction of anemia. Further research studies are advised to understand the prevalence and determinants of anemia.

In this study, we used the data from the 2016 NDHS, a nationally representative cross-sectional survey conducted as a part of Demographic and Health Surveys (DHS) by New ERA under the guidance of Ministry of Health, Nepal, which aims to provide up-to-date estimates of the basic demographic and health indicators [ 24 ]. The dataset of this survey was accessed publicly from ‘The DHS Program’ website upon subsequent registration and authorization [ 25 ]. Details of the questionnaires, procedures, and methodology used in the survey can be found elsewhere [ 24 , 25 ]. Briefly, the 2016 NDHS was conducted based on the multi-stage cluster sampling technique. For sampling, each of the seven provinces of Nepal was divided into rural and urban area yielding 14 sample strata. The 2016 NDHS sample was then selected in two and three stages in rural and urban areas, respectively. In the first stage, 383 wards (urban, 184; rural, 199) were selected using the probability proportional to size sampling technique. In rural areas, wards were smaller and thus, served as the primary sampling unit (PSU); in urban areas, the wards were larger, and therefore, one enumeration areas (EAs) was randomly selected form each sampled ward in the second stage of sample selection. In the final stage, a fixed number of 30 households per cluster were selected using an equal probability systematic selection technique. The 2016 NDHS used the following six questionnaires to collect data: household questionnaire, woman’s questionnaire, man’s questionnaire, biomarker questionnaire, fieldworker questionnaire, and the verbal autopsy questionnaire. The household questionnaire intended to provide information on the household characteristics of all members of the household. The women’s questionnaire was used to collect information from all women aged 15–49 years. The biomarker questionnaire was used to record height, weight and hemoglobin concentration and was provided only to the women in the subsample of the households selected for the men’s questionnaire. The data on anthropometric measures (height and weight) and hemoglobin concentration were first recorded on the paper during data collection and then entered into the computer-assisted system. The fieldworker questionnaire was used as a tool for examining data quality, and the verbal autopsy questionnaire was used for recording neonatal deaths. Blood samples for the hemoglobin test were collected from women who voluntarily provided their consent to undertake the test and otherwise excluded. Following a finger-prick, blood was drawn into a microcuvette for on-site analysis using a battery-operated portable HemoCue analyzer [ 24 ]. According to the WHO, for non-pregnant women aged ≥15 years, any anemia was defined as blood hemoglobin level <12.0 g/dL, which was further categorized as mild (11.0–11.9 g/dL), moderate (8.0–10.9 g/dL), and severe anemia (<8.0 g/dL); for pregnant women, any anemia was defined as blood hemoglobin level <11 g/dL, and further categorized as mild (10.0–10.9 g/dL), moderate (7.0–9.9 g/dL), and severe anemia (<7.0 g/dL) [ 1 ]. Hemoglobin concentration was adjusted for cigarette smoking and altitude (1000 m above sea level) [ 24 ]. For analysis, the severity of anemia was categorized as any-anemia and no-anemia. Predictors of anemia were chosen based on the array of literature pertaining to the risk of development of anemia among women in low-and-middle income countries including Nepal [ 8 , 9 , 11 – 13 , 26 – 30 ]. The independent variables in this study included socio-demographic factors (age, ethnicity, place of residence, education level, occupation, household wealth status, and sources of drinking water); reproductive characteristics (fertility status, age at first birth, total number of children ever born, adverse pregnancy outcome for last/most recent pregnancy, birth in the past three years, breastfeeding status at the time of the survey, contraception use at the time of the survey, and iron supplementation during the most recent pregnancy in the past five years); nutritional and behavioral factors (body mass index (BMI) and cigarette/tobacco smoking); women’s autonomy in decision-making regarding her healthcare; and experience of any intimate partner violence (IPV; physical and/or sexual). In this study, the self-reported age of the women, a continuous variable, was categorized into the following age groups: 15–24 years, 25–34 years, and 35–49 years [ 9 ]. Ethnicity was categorized as Brahmin/Chhetri, Janajati/Indigenous (includes Newar), Dalit and other castes (includes all other recorded ethnicities). Education level of women was categorized as no education (no schooling), primary (up to grade 5), secondary (up to grade 10), and higher (higher than secondary level). Household wealth status was determined by using the scores derived from the principal component analysis of several possessions, assets, and amenities the households own [ 31 ]. this derived variable was already included in the dataset as five quintiles ranked as poorest, poorer, middle, richer and richest each comprising 20% of the population [ 24 ]. In this study, wealth status was re-categorized as poor (includes poorest and poorer), middle, and rich (includes richer and richest) [ 32 ]. Occupation was categorized as did not work, professional/service (included technical, managerial, sales, and clerical), agriculture, and manual (included skilled and unskilled manual work) [ 33 ]. The sources of drinking water were grouped as tap water (public and private tap), well (tube well, dug well, borehole), surface water (river, dams, ponds, streams, springs) and others. Fertility status of women was categorized as infecund/menopausal, fecund, pregnant, and postpartum amenorrhoeic. Women who were pregnant at the time of the survey were categorized as pregnant women. Women whose period has not returned since the last birth were categorized as postpartum amenorrhoeic women. Similarly, women who were not pregnant and not postpartum amenorrhoeic, and did not have a period in the last six months were categorized as menopausal women. Likewise, women were categorized as infecund, if they were not menopausal, not postpartum amenorrhoeic, not pregnant, and had no birth in the past five years. All other women who were not included in the former categories were categorized as fecund. Adverse pregnancy outcome for the last/most recent pregnancy the women had was categorized as no (live birth) and yes (stillbirth, miscarriage, abortion). Contraception use was categorized as not using any method, hormonal (included pills, injections, implants, IUD, emergency contraception), female sterilization, male contraception (included male condoms and male sterilization), and traditional (withdrawal, abstinence, and other traditional methods). Iron supplementation during last pregnancy (for women with a live birth during the last five years of the survey) was categorized as not taken at all, took for <180 days, and took for ≥180 days. Women are recommended to take iron/folic acid supplements daily for a minimum of 180 days during pregnancy and until 45 days after childbirth [ 34 ]. Height and weight of all eligible women aged 15–49 years were measured by trained field staffs. BMI was calculated as the ratio of weight (in kilograms) to the square of height (in meters); it was recorded as a continuous variable and was available in the 2016 NDHS dataset [ 24 ]. For this study, BMI was categorized as underweight (less than 18.5 kg/m 2 ), normal (18.5–24.9 kg/m 2 ), overweight (25.0–29.9 kg/m 2 ), and obese (≥30.0 kg/m 2 ) [ 35 ]. For regression analysis, overweight and obese categories were merged into a single category as overweight/obese. The 2016 NDHS collected information on domestic violence and women’s participation in household decision-making, specifically among currently married women. Women’s decision-making autonomy regarding her healthcare was categorized as high autonomy (if she takes a decision independently), medium autonomy (if she takes a decision with the partner), and low autonomy (if she is not involved). Women’s experience of intimate partner violence (IPV) was measured by analyzing two forms of violence, physical and sexual IPV. Physical IPV was determined based on women’s responses to seven questions by asking them whether their husbands ever did the following: 1. pushed, shook or thrown something at her; 2. slapped her; 3. twisted her arm or pulled her hair; 4. punched her with fist or something that could hurt her; 5. kicked, dragged or beat her; 6. tried to choke or burn her on purpose; and 7. threatened or attacked her with knife, gun, or any other weapon. Similarly, sexual IPV was determined based on women’s responses to three questions by asking them whether their husbands ever did the following: 1. physically forced her to have unwanted sexual relationships with him; 2. physically forced her to perform any other unwanted sexual acts; and 3. forced her with threats and any other way to performs unwanted sexual acts. A composite dichotomous summary was developed from 10 questions (physical IPV, 7 questions; sexual IPV: 3 questions) with yes/no answers to capture the women’s experience of any IPV and categorized as ever experienced IPV (‘yes’ response to at least one of the ten questions) and not experienced IPV (‘no’ response to all the ten questions). The Cronbach’s alpha for any IPV (10-item scale) was 0.876. The outcome variable of this study was the presence of anemia among women. a binary variable was created as any-anemia (mild, moderate, and severe) and no-anemia. Data analysis was carried out on SPSS 25.0 (IBM Corp., Armonk, NY). Due to the non-proportional allocation of the sample in the 2016 NDHS, the data were adjusted before any statistical analysis for sampling weights, stratification, and multistage sampling procedure to provide population-level estimates [ 36 ]. Data were analyzed using descriptive statistics to describe the characteristics of the study population. Frequencies and proportions (weighted) were reported. Continuous variables were reported using mean and standard deviation (SD). Chi-square test (χ 2 ) or Fisher’s exact was used to assess the frequency distribution and the relationship between independent variables and anemia status. Similarly, bivariate logistic regression analysis was used to determine the individual effect of each factor on anemia status. Finally, multivariate regression analysis model was performed to determine the adjusted effect of each factor on the dependent variable. The results of regression analysis were presented by crude/unadjusted odds ratio (OR) and adjusted odds ratio (aOR) with 95% confidence intervals (CIs). A variable with a p-value <0.05 in the bivariate analysis were considered statistically significant and thus, included in the final regression model. Prior to the multivariate regression analysis, the independent variables were checked for multicollinearity by examining the variance inflation factor (VIF); no serious issues (VIF≤ 2) were found [ 37 ]. Additionally, a separate multivariate regression analysis model was generated to find out the impact of women’s decision-making and experience of IPV on anemia status; a similar data analysis procedure was followed. As the study was based on secondary data from the 2016 NDHS, no separate ethical approval was sought. Nevertheless, the 2016 NDHS was reviewed and approved by the ICF International Review Board and the Nepal Health Research Council (NHRC). Details on ethical procedures used in the DHS survey can be found elsewhere [ 38 ].