Determinants of Anemia Among Pregnant Women Attending Antenatal Care in Bale-Robe Town Health Facilities, Bale zone, Southeast Ethiopia: A Case-Control Study

In Ethiopia, the prevalence of anemia in pregnant women differs geographically and ranged from 15.8– 56.8%. The objective of this study is to identify the determinants of anemia among pregnant women attending antenatal care in health facilities of Bale-Robe Town, Southeast Ethiopia. Methods A facility-based case-control study was conducted in Bale-Robe hospital and Baha-biftu health center in Bale-Robe, Southeast Ethiopia. A total of 282 pregnant women participated in the study (141 cases and 141 controls). Cases were pregnant women with altitude-adjusted hemoglobin value < 11.0 g/dl at the rst and third trimesters, and < 10.5 g/dl at the second trimester. Controls were pregnant women with hemoglobin value ≥ 11.0 g/dl at rst and third trimesters and ≥ 10.5 g/dl at the second trimester. A structured and pretested questionnaire was used to collect data. A multivariable logistic regression analysis was applied to assess the determinants of anemia. Determinants were categorized as sociodemographic and economic, obstetric and medical, and dietary intake and behavioral.


Introduction
In Ethiopia, the prevalence of anemia in pregnant women differs geographically and ranged from 15.8-56.8%. The objective of this study is to identify the determinants of anemia among pregnant women attending antenatal care in health facilities of Bale-Robe Town, Southeast Ethiopia.

Methods
A facility-based case-control study was conducted in Bale-Robe hospital and Baha-biftu health center in Bale-Robe, Southeast Ethiopia. A total of 282 pregnant women participated in the study (141 cases and 141 controls). Cases were pregnant women with altitude-adjusted hemoglobin value < 11.0 g/dl at the rst and third trimesters, and < 10.5 g/dl at the second trimester. Controls were pregnant women with hemoglobin value ≥ 11.0 g/dl at rst and third trimesters and ≥ 10.5 g/dl at the second trimester. A structured and pretested questionnaire was used to collect data. A multivariable logistic regression analysis was applied to assess the determinants of anemia. Determinants were categorized as sociodemographic and economic, obstetric and medical, and dietary intake and behavioral.

Conclusion
Housewife occupation, prolonged menstrual bleeding, and malnutrition were the determinants of anemia in pregnant women. Hence, anemia prevention and control strategy in pregnant women should include adequate dietary intake, and strengthening nutritional counseling for pregnant women during antenatal care is also required by the health care provider. Background Anemia in pregnant women is a common public health problem worldwide. It affects 38.2% of pregnant women globally, 48.7% in Southeast Asia, and 46.3% in Africa. 1 Even though, anemia in pregnant women is a global public health problem, its burden is more substantial in developing countries. 1 2 Anemia in pregnant women has multiple adverse outcomes for both mothers and infants, including postpartum hemorrhage, cardiac failure, less exercise tolerability, thromboembolic problems, spontaneous abortion, puerperal infection, placenta previa, maternal mortality, preterm delivery, low birth weight, and prenatal death. 2 Neonates of anemic mothers are born with reduced iron stores and at high risk of developing ironde ciency anemia, which leads to long-term effects such as; neurophysiologic and poor cognitive-motor development of children. 3 Worldwide, it has been reported that anemia contributes to more than 591,000 prenatal deaths and 115,000 maternal deaths in a year. 4 Approximately 20% of maternal death is due to anemia and its majority occur in developing countries. 5 In developing countries, pregnant women start pregnancy with already reduced body stores of iron. This is mainly due to repeated infections, poor nutritional intake, frequent pregnancies, and menstrual blood loss. Besides this, it is also related to socioeconomic conditions, health-seeking behaviors, and lifestyles across different cultures. 6 7 In Ethiopia, several studies showed that the prevalence of anemia in pregnant women varies geographically and ranged from 15.8-56.8%. 8-12 Nearly one-third of pregnant women (31.66%) in Ethiopia has anemia according to a systematic review study conducted in Ethiopia. 8 Currently in Ethiopia, different services are undertaken during antenatal care to prevent anemia during pregnancy which includes, nutritional interventions, iron, and folic acid supplementation and deworming. 13 Despite many efforts made by the stakeholders and government, anemia in pregnant women is still a major public health problem. The prevalence of anemia varies within and between regions. 10 11 14-18 This suggests a need for local data to identify the determinants and address the problem. Therefore, the focus of this study was to identify the determinants of anemia among pregnant women attending antenatal care in the health facilities of Bale-Robe town, Southeast Ethiopia.

Study setting and period
The study was conducted from March 20 to June 2, 2019, among pregnant women attending antenatal care in the Bale-Robe Town health facilities. Bale-Robe town is the capital of the Bale zone and is located 430 km from Addis Ababa, the capital of Ethiopia. According to the town administrative health o ce, the total population of the town was 148,089 in the year 2019, of whom 72,860 (49.2%) are females. The town is found at an altitude of 2492 meters above sea level and there is one general public hospital, one public health center and nine private clinics in the town. The routine focused antenatal care service is provided at the public health facilities in the town.

Study design and population
A facility-based case-control study was conducted among pregnant women who attend antenatal care in Bale-Robe town health facilities. Cases were pregnant women with altitude-adjusted hemoglobin value < 11.0 g/dl at the rst and third trimesters, and < 10.5 g/dl at the second trimester. Controls were pregnant women with an altitude-adjusted hemoglobin value ≥ 11.0 g/dl at rst and third trimesters and ≥ 10.5 g/dl at the second trimester. 19 Pregnant women who attend antenatal care in the health facilities of Bale-Robe town during their 1 st visit, irrespective of their trimester, and those who are residents in the town for a minimum of six months were included in the study. However, women who were severely ill during data collection and those already talking the ferrous sulfate were excluded from the study.

Sample size and sampling techniques
The sample size was calculated by Epi Info version 7. 1 software using a formula for an unmatched casecontrol study. The calculation was made for several determinants of anemia among pregnant women from previous studies by considering 95% con dence level, 80% power, a case-to-control ratio of 1:1. The largest sample size was obtained using the proportion of pregnant women who had anemia and not consuming chicken meat at least once per week (91.1%) and an Adjusted odds ratio (AOR) of 2.88. 20 Thus, after considering 5% non-response rate, a total of 282 participants (141 cases and 141 controls) were included in the study.
There are two public health facilities and nine private clinics in Bale-Robe town. Of these, two public health facilities (one general hospital and one health center) were included in the study considering the routine focused antenatal care service provision, which is provided only at the public health facilities. The number of the study subjects were allocated to both facilities proportional to their average rst visit attendants per month by referring to the antenatal care registration books in both facilities. Thus, the number of monthly antenatal care attendants was 230 in the hospital and 60 in the health center.
Accordingly, 238 study subjects were allocated to Bale-Robe General Hospital and 44 participants to Baha-Biftu Health Center. The consecutive sampling method was used to select the study participants.

Data collection procedures
Data were collected using a structured questionnaire adapted from literatures. 5 7 -10 14-20 Data were collected by eight health professionals (four BSc nurses and four BSc laboratory technicians) under the supervision of two supervisors. The study subjects were identi ed as cases and controls based on hemoglobin level measurement, then the supervisor send to the data collectors. The questionnaire was categorized into three groups of characteristics such as; sociodemographic and economic characteristics, obstetric and medical characteristics, and dietary intake and behavioral.

Measurements
Blood hemoglobin level was determined using hemocue 201 following standard operation procedures 19 . The hemoglobin cutoff value was adjusted for the altitude of Bale-Robe town (2492 meters) using WHO criteria. 21 A stool sample was collected from each study subject using a leak-proof stool cup and then stool wets mount as prepared using saline and/or iodine solution. Then, microscopically examined for identi cation of intestinal parasites. HIV test was conducted following the current testing algorism using the rapid test kit. Dietary diversity was measured using the Dietary Diversity Score (DDS), a tool adapted from Food and Agricultural Organization for measuring individual dietary diversity. 22 The DDS was determined by asking each study participant to list all the food items consumed in the last 24 hours. Once the recall was nished, the participant was asked for the food group that was not stated. The food items were categorized into nine food groups. The dietary diversity score is the sum of food groups and ranges from 1-9. 22 The mean upper arm circumference (MUAC) was measured to assess malnutrition in pregnant women using WHO standardized measuring tape. 23 Data quality control The questionnaire was translated to the local language Afan Oromo, and then back-translated to English for consistency (additional le 1). The questionnaire was pre-tested on 5% of the sample size (14 participants) at Goba Referral Hospital and necessary revision was made accordingly. The supervisors and data collectors were trained for two days. The laboratory reagents were regularly monitored for proper storage. On daily basis, the supervisors have strictly checked the consistency and completeness of data collected. Furthermore, the principal investigator checked data for consistency and completeness before data entry.

Data analysis procedures
The questionnaire was coded and entered using Epi Data. The data cleaning and analysis were done by SPSS version 23 software. Kolmogorov-Smirnov test was used to check the normal distribution of continuous variables. The outcome variable which is hemoglobin level was adjusted for Bale-Robe town altitude by subtracting 1.3g/dl from the observed hemoglobin level. 21 Descriptive statistics were computed, and the categorical variables were cross-tabulated with the outcome variable, and the ndings were described by frequencies and proportions for the cases and control group. The reliability of items of the scale for dietary diversity score was assessed using Cronbach's alpha.
Bivariable logistic regression was performed and variables with P-value less than 0.25 were transferred to multivariable logistic regression to identify the determinants of anemia. Variables with P-value<0.05 in multivariable logistic regression were considered statistically signi cant. The strength of association was described using an adjusted odds ratio along with its 95% con dence interval. Multicollinearity test was done using variance in ation factor and no collinearity exists between explanatory variables. The model goodness of the test was checked by Hosmer-Lemeshow goodness of the ttest (P-value=0.81).

Operational de nition
Altitude adjusted hemoglobin level Hemoglobin value adjusted for altitude < 11.0 g/dl at rst and third trimesters and < 10.5 g/dl at second trimester were used to de ne anemia. 19 Dietary diversity Adequate dietary diversity is considered when the DDS is ≥ 4 and inadequate dietary diversity is the DDS below 4. 22

Malnutrition
The mean upper arm circumference measurement below 23 centimeters. 23 Alcohol consumption Current alcohol drinker is considered if a woman consumed more than 14 standard drinks of alcohol per week, and ever alcohol drinker is considered if they had consumed alcohol at least once in their life time. 24

Sociodemographic and economic characteristics
A total of 281 pregnant women (141 cases and 141 controls) were recruited in the study. The response rate was 100%. The median (interquartile range) age of the participants was 28 years (IQR=25-31 years), and ranged from 18-39 years for cases and 19-40 years for controls. Fifty-three (37.6%) of the cases and 63 (44.7%) of the controls were found between 25-29 years. More than three-fourth the cases, 109 (77.3%) and 111 (78.7%) of controls were attended formal education. The occupational status of nearly to two-third of the cases, 89 (63.1%) and only 60 (42.6%) of the controls were housewives. The average monthly family income for three-fourth of the cases, 106 (75.2%) and 106 (75.2%) of the controls were above 2000 Ethiopian birr (Table 1). ) of the controls were in the second trimester pregnancy. Fifteen (10.6%) of the cases and 6 (4.3%) of the controls had history of blood loss during the current pregnancy. Intestinal parasites were identi ed in only nine (6.4%) of the cases and ve (3.5%) of the controls. Among the participants, only 6 (4.3%) of the cases and 6 (2.8%) of the controls reported history of chronic illness (Table 2).

Bivariable Analysis
Variables having P-value < 0.25 in bivariable logistic regression were transferred to multivariable logistic regression. Accordingly, religion and occupational status from sociodemographic and economic characteristics (Table 1), the duration of menstrual ow and history of blood loss during pregnancy from obstetric and medical characteristics (Table 2), drinking coffee immediately after meal, dietary diversity and nutritional status from dietary and behavioral characteristics (Table 3) were selected as candidate variables for multivariable logistic regression.

Determinants of anemia among pregnant women
In multivariable logistic regression, three variables (occupational status, duration of menstrual ow, nutritional status) were signi cantly associated with anemia among pregnant women. Pregnant women whom their occupational status was housewives were two times more likely to have anemia compared to the employed women (AOR=2.1, 95% CI =1.12-3.92). The odds of developing anemia among pregnant women who had prolonged previous menstrual period for more than 6 days were two times higher compared to those mothers with less than six days menstrual bleeding (AOR=2.33, 95% CI=1.38-3.92). Similarly, pregnant women who had malnutrition were four times more likely to have anemia during pregnancy as compared to well-nourished (AOR=4.03, 95% CI=1.38-11.83) ( Table 4).

Discussion
Anemia in pregnant women increases the risk of both maternal and child morbidity and mortality, including impaired cognitive development of the children, and decrease work productivity in adults. 25 Prevention and control of anemia among pregnant women are key measures to reduce the adverse effects of anemia in mothers and children. The nding of this study showed that housewife occupational status, prolonged menstrual period and malnutrition were signi cant determinants of anemia among pregnant women.
Housewife occupational status is found to be signi cantly associated with anemia in pregnant women.
Pregnant women whose their occupational status was housewives were two times more likely to have anemia compared to employed women. This nding is consistent with a similar study conducted in Durame, Southern Ethiopia, 26 and a study in Gondar, Northern Ethiopia, 27 which reported a signi cant positive association between housewife occupational status and anemia in pregnant women. This might be due to the workload on housewives and nancial shortage, which result in di culty of obtaining food.
Mothers with a shortage of income due to occupational status will not able to ful ll the daily dietary intake needed for pregnant women and as a result, develop anemia. 28 In contrast, the study conducted in North Shoa, Ethiopia found a negative association between housewives' occupational status and anemia during pregnancy. 14 This discrepancy might be due to the variation in study design. The implication of this nding is the importance of women engaging in an income-generating occupation to ful ll dietary requirements during pregnancy.
Another nding of this study is the prolonged previous menstrual bleeding, which is signi cantly associated with anemia. The odds of developing anemia among pregnant women who had prolonged previous menstrual bleeding for more than 6 days were two times higher compared to those mothers with less than six days of bleeding. This nding is in line with the studies conducted in Durame town, Southern Ethiopia, 26 and in Mekele town, Tigray Ethiopia, 29 which showed a signi cant association between heavy menstrual bleeding and development of anemia during pregnancy. Likewise, our nding is supported by similar studies conducted in Hawassa and Yirgalem, 30 Wolayita, 18 Mizantepi, 12 and East Hararghe zone of Ethiopia 5 . This is due to the fact that prolonged menstrual bleeding results in blood loss and reduces the number of circulating red blood cells. In addition, prolonged blood loss may decrease iron levels enough to increase the risk of iron de ciency anemia. This nding implies the women with the prolonged previous menstrual bleeding need to take iron-rich food and iron supplementation during pregnancy.
Undernutrition during pregnancy is found to be signi cantly associated with anemia. Pregnant women who had malnutrition during pregnancy were four times more likely to have anemia as compared to wellnourished. This is consistent with the nding of studies conducted in Asosa, Western Ethiopia, 31 and study in Gondar, Northern Ethiopia. 28 Similarly, the studies conducted in Jamaica, 32 and Tanzania, 33 reported a positive association between undernutrition during pregnancy and anemia. Nutritional de ciency is the major contributing factor for anemia among pregnant women and this is due to undernutrition is related to iron, folate, and vitamin A and other micronutrients de ciencies. 12 27 34 This is due to the fact that poor dietary intake and micronutrient de ciencies due to malnutrition result in anemia because pregnancy is the most nutritional demanding. In addition, malnutrition during pregnancy is results in iron de ciency anemias as a result of inadequate dietary iron intake. This is further supported by the WHO report, which stated in developing countries, inadequate intake of dietary iron is the main cause of anemia during pregnancy. 35 Our nding suggests that taking adequate dietary intake during pregnancy is required for the prevention of anemia among pregnant women.

Limitations Of The Study
The temporal relationships between explanatory variables and anemia cannot be determined, as it is a case-control study. Social desirability bias and recall bias are also other limitations of our study regarding dietary diversity, alcohol consumption and monthly income. The dietary intake could also be affected by seasonal variation. Moreover, the ndings of the current research cannot be generalized to the whole community because of its institution-based nature.

Conclusions
Different factors were studied to identify the determinants of anemia among pregnant women.
Housewife occupation, prolonged menstrual period and malnutrition were the determinants of anemia among pregnant women. Therefore, pregnant women with prolonged previous menstrual bleeding need to take iron-rich food and iron supplementation. Anemia prevention and control strategy in pregnant women is required to include adequate dietary intake for the prevention of anemia. Strengthening nutritional counseling for pregnant women during antenatal care is also required by the health care provider. Finally, further longitudinal studies which determine the cause of anemia in pregnant women are recommended for the researcher.

Strengths And Limitations Of This Study
This study used hemocue 201 to determine blood hemoglobin levels following standard operating procedures.
This study included several detailed measures of dietary, behavioral, obstetric and medical characteristics, and identi ed determinants that will be important in the prevention and control of anemia in pregnant women.
The temporal relationships between explanatory variables and anemia could not be established as we have used a case-control study design.
Social desirability bias and recall bias might be introduced during measuring dietary diversity, alcohol consumption and monthly income.

Declarations
Contributions: MHB involved in the study design, data collection, analysis and write-up. FSK performed analysis, interpretation of data and drafted the manuscript. MSA contributed to the design, data collection and revision of the manuscript. All authors approved the nal version.
Funding: This study no speci c nance from any funding agency.
Competing interests: None declared. Patient and Public Involvement: Written informed consent was obtained from each participant after explaining the purpose of the study. The con dentiality of information obtained from the participants was assured and the study subjects were informed that, they had full right to participate to the study. The pregnant women who had anemia and those investigated positive for intestinal parasite were linked to the health facilities antenatal care unit for treatment.
Data sharing statement: No additional data are available.