Anaemia at antenatal care initiation and associated factors among pregnant women in West Gonja District, Ghana: a cross-sectional study

Introduction Anaemia in pregnancy remains a critical public health concern in many African settings; but its determinants are not clear. The purpose of this study was to assess anaemia at antenatal care initiation and associated factors among pregnant women in a local district of Ghana. Methods A facility-based cross-sectional survey was conducted. A total of 378 pregnant women attending antenatal care at two health facilities were surveyed. Data on haemoglobin level, helminths and malaria infection status at first antenatal care registration were extracted from antenatal records booklets of each pregnant women. Questionnaires were then used to collect data on socio-demographic and dietary variables. Binary and multivariate logistic regression analyses were done to assess factors associated with anaemia. Results The prevalence of anaemia was 56%, with mild anaemia being the highest form (31.0%). Anaemia prevalence was highest (73.2%) among respondents aged 15-19 years. Factors that significantly independently reduced the odds of anaemia in pregnancy after controlling for potential confounders were early (within first trimester) antenatal care initiation (AOR=5.01; 95% CI =1.41-17.76; p=0.013) and consumption of egg three or more times in a week (AOR=0.30; 95% CI=0.15-0.81; P=0.014). Conclusion Health facility and community-based preconception and conception care interventions must not only aim to educate women and community members about the importance of early ANC initiation, balanced diet, protein and iron-rich foods sources that may reduce anaemia, but must also engage community leaders and men to address food taboos and cultural prohibitions that negatively affect pregnant woman.


Introduction
Globally, anaemia affects an estimated 43% of children, 38% of pregnant women, and 29% of non-pregnant women of childbearing age [1]. In low-income countries, anaemia affects 40 to 60% of pregnant women [2,3]. The World Health Organization defines anaemia as decreased concentration of haemoglobin (Hb) level of less than 11g/dL [1]. Anaemia during pregnancy is considered severe when Hb concentration level is less than 7.0g/dL; moderate when haemoglobin level falls between 7.0-9.9g/dL; and mild from 10.0-10.9g/dL [4]. The causes of anaemia during pregnancy are multifactorial, and includes nutritional deficiencies of iron, folate, and vitamin B12 [1]. Economic and socio-cultural factors such as cultural and religious food taboos also significantly contribute to anaemia among pregnant women [3,5]. Other causes of anaemia in pregnancy include parasitic infections like helminths and other conditions such as low intake or poor absorption of iron [4]. Iron deficiency is the most common cause of anaemia in pregnancy in many low-income settings [2,4]. While evidence suggests that most women in low-income countries, including Ghana, enter pregnancy with less than adequate stores of nutrients [6], anaemia in pregnant women could have serious adverse pregnancy outcomes, including high maternal death, impaired mental development in children, increased risk of fetal growth retardation, low birth weight, premature delivery and perinatal mortality [7,8]. Like many countries in Africa, anaemia remains an important threat to safe motherhood and newborn health in Ghana [9,10]. Anaemia is the number two cause of all admissions and the number five cause of death among all admitted patients in Ghana [11]. Indeed, health facility level data suggest that the prevalence of anaemia among pregnant women in Ghana is on the rise, from 34% in 2014 to 37% in 2016 [11]. There are however regional disparities. In the Northern region where this study was conducted, 43.2% of pregnant women attending ANC in 2016 were anaemic [11]. The situation in the specific district (West Gonja District) where this study was conducted is worse: anaemia among ANC attendants rose from 23.4% in 2012 to 43.9% in 2016 [12]. While the potential adverse health consequences of anaemia in pregnancy are widely recognised, few empirical studies have been conducted in Ghana to identify key determinants [10].
Indeed, the lack of evidence on anaemia in many low-income countries is acknowledged as one of the reasons why the fight against anaemia in pregnancy still remains a problem [1]. This study aimed to assess anaemia at antenatal care initiation and its determinants among pregnant women in a local district of Ghana. All pregnant women aged 15-49 years who were attending these two health facilities to receive their first ANC between November 2017 and April 2018 were eligible for the study. However, pregnant women who reported a recent history of blood transfusion (within the past three months) before initiation of first ANC were excluded from the study.

Study design and respondents
Study setting: the west Gonja District has an estimated population of 49,386 [13]. Women form 51% of the district's population, with about 1,975 women expected to have become pregnant in 2017 [12].
The main occupations of women in the district are farming and retail trade and services, with few engaged in teaching and nursing [12].
Health service delivery in the district is done through a total of thirteen about the purpose of the study and the study procedures. Those who could read (in English) were immediately provided with information leaflets about the study. Those who could not read were asked if they wanted to receive the information leaflet so that a family member or friend could later read and explain to them. Nearly all such women accepted the information leaflets. The research assistants enlisted the names and contact numbers of all the women approached. Those without personal telephone numbers were requested to provide the numbers of their husband/partner, family member or friend.
Following from this, each woman was given two weeks to decide on their participation. They were each re-contacted via telephone after the two-week period. Where the decision was in favor of participation, interview dates were arranged, usually on the next ANC visit. However, where the decision was against participation (there were 21 such cases), such women were dropped.
In terms of data collection, two methods were employed: data extraction from ANC booklets of respondents and administration of structured questionnaires. First, the following information was The questionnaires were pre-tested at two other smaller health centres located in the district. All necessary corrections were made before actual data collection from November 2017 to April 2018.
Actual data collection occurred alongside recruitment: as women reported to the ANC clinic on weekly basis for the first time, they were approached, recruited and subsequently interviewed. The two research assistants conducted all interviews in a designated small room within the premises of each health facility. Women were interviewed one at a time. English and Gonja (local dialect) were the interview languages.
Data entry and processing: completed questionnaires were manually examined for completeness, then hand-coded and entered into Epi info version 7. The data were independently entered by the two research assistants. The first and second authors then independently compared the two data entries. All errors were discussed and resolved before data were exported into Stata (version 15.0) for further cleaning and analysis.
Variables: the main outcome variable is anaemia, which we defined and measured primarily as a binary outcome. We followed the WHO's definition and categorisation: women whose haemoglobin (Hb) concentration levels were >11g/dl and <11g/dl were classified as 'not anaemic' and 'anaemic' respectively [1]. We re-categorized all anaemic women into mild (10-10.9g/dl), moderate (7-9.9g/dl) and severe (<7g/dl). Several independent variables were also defined and measured, including socio-demographic factors such as age, maternal education, occupation, marital status, religion, husband's occupation, and place of residence as well as maternal and dietary characteristics. Statistical analysis: categorical variables were summarised into frequencies and proportions. Continuous variables were summarised into means and ranges and continuous variables like age were recategorised into age groups. Bivariate analysis was first done using chi-square test of independence to assess association between anaemia and categorical independent variables. Binary logistic regression was used to assess for factors associated with anaemia.
Factors with p-value < 0.05 at 95% confidence level were considered statistically significant and were therefore included in a multiple logistic regression model for further analysis. Odd ratios were estimated.  Table 1 shows the background characteristics of the 378 respondents who took part in the study.

Characteristics of respondents:
The mean age was 26.9, and the majority (29.1%) were aged 25-29 years. Table 2 also shows the maternal characteristics of respondents.
Majority (51.9%) initiated ANC in the second trimester (13 to 24 weeks). Some 13.8% of the respondents tested positive for malaria at their first ANC visit, while 31.2% tested positive for helminths infection. Table 3 describes the dietary characteristics of respondents.
A combined 55.8% of the respondents took meat (including liver) and fish at least three times a week. Some 51.1% also consumed egg 1-2 times per week. Green leafy vegetable consumption was generally high among respondents: 21.4% and 76.7% consumed green leafy vegetable 1-2 times and 3+ times per week respectively.
Prevalence of anaemia: in terms of prevalence of anaemia, 55.8% of the respondents were anaemic (Hb less than 11g/dl), with the mean Hb level being 10.8g/dl and a range of 6.7g/dl to 14.4g/dl. Among the 55.8% who had anaemia, 0.3% had severe anaemia, 24.5% had moderate anaemia, and 31.0% had mild anaemia.
Predictors of anaemia: to determine factors associated with anaemia in pregnancy, chi-square tests of independence were first performed between a total of 24 independent variables and anaemia in pregnancy. From this initial analysis, 11 factors were statistically associated with anaemia in pregnancy. These 11 factors were then pulled into binary and multiple logistic regression models and odds ratios were estimated. The results are shown in Table 4 and Table 4 (suite). After adjusting for potential confounders, two factors significantly independently predicted anaemia in pregnancy: timing of

Discussion
This study is one of the few to assess anaemia prevalence and associated factors among pregnant women attending ANC services in Ghana. Results suggest that the prevalence of anaemia among pregnant women in the study is quite high (56%), with mild anaemia being the highest (31.0%) form. Two factors significantly independently predicted anaemia in pregnancy after adjusting for other factors, namely timing of ANC initiation and egg consumption per week. Several aspects of these results deserve further reflection on. The prevalence of anaemia in this study is highest (73.2%) among respondents aged between 15-19 years. This is consistent with findings from Mangla & Singla's study [6]. A number of factors could contribute to high anaemia in this age group. One of the important causes of anaemia is iron deficiency, and studies suggest that the 15-19year age band is a period of intense physical and mental growth, with a higher demand for iron and other nutrients [4,10]. Pregnancy and childbirth during this age group could place further demands on the already inadequate iron stores in teenage mothers. This could easily predispose pregnant teenagers to anaemia. Apart from the fact that young girls may be unprepared biologically, they may also be unprepared emotionally and economically to deal with pregnancy. This is particularly likely because in many contexts in Ghana, sexual and reproductive health topics remain taboo subjects for most parents to discuss with their adolescent children, and teen pregnancy is often not welcome [14]. This could easily undermine social and economic support for teenage mothers, which could in turn affect their nutritional status. This would suggest a need to intensify early sexual and contraception education and counselling for female adolescents at home and in school as well as self-efficacy training and skills acquisition to help them negotiate peer-pressures to initiate sex early and to protect themselves during sexual intercourse. The role of parents and guardians in providing sexual and reproductive health education needs to be encouraged given that early sexual debut and childbearing among female adolescents is a widely reported phenomena in Africa [15][16][17]. Apart from interventions to stop or reduce early sexual debut and childbearing, interventions to encourage teen mothers to seek early ANC together with targeted nutritional counselling and support services, would also be essential.
The timing of ANC initiation emerged as an important predictor of anaemia among first time registrants. Compared to women who initiated ANC in the first trimester, the odds of having anaemia in pregnancy were still significantly higher among pregnant women who initiated ANC in the second trimester (AOR=2.71; 95% CI=2.09-5.81; P<0.01) and third trimester (AOR=5.01; 95% CI=1.41-17.76; p=0.013). This is also consistent with what has been reported in Bangladesh [18], and in a regional health facility study in South Africa [8], where anaemia in pregnancy was higher among women who registered in the second and third trimesters compared to those who registered within the first trimester. That late ANC initiation is associated with anaemia in pregnancy is however not surprising. This is not only because majority of women in our study initiated ANC either in the second or third trimester, but also because late ANC initiation means that many of the interventions and services routinely offered to pregnant women at ANC clinics to prevent anaemia in pregnancy such as IFA supplementation, provision of LLINs, and IPT dosing, as well as laboratory investigations (e.g. Hb check and stool tests) to diagnose early anaemia in pregnancy and offer early treatment, are delayed for such women. This would suggest a need for both health facility and community-based preconception and conception care interventions to educate women and community members on the importance of early antenatal care initiation and the need to seek ANC services early. In doing this, efforts must be made to address health system barriers such as long distances to service centres as well as engage community members (men and mothers-in-law in particular) to address socio-cultural barriers such as the need to perform traditional pregnancy-related rituals before permission is granted for pregnant women to access services as shown in previous research in northern Ghana [19,20]. Some aspects of dietary characteristics were also significantly associated with anaemia in pregnancy. While meat/fish and green leafy vegetable consumption did not surprisingly show significant statistical association with anaemia in pregnancy as we expected, frequency of egg consumption did show strong statistical association with anaemia such that women who consumed eggs three or more times in a week were less likely to be anaemic in pregnancy compared to those who did not consume eggs at all. This is similar to findings by Gebre & Mulugeta in northern Ethiopian where frequency of egg consumption per day was strongly associated with anaemia in pregnancy [3]. Our results here support existing evidence that highlights the nutritional benefits of egg consumption (at least in the context of anaemia) during pregnancy and suggest a need for pregnant women to incorporate eggs into their diet. We acknowledge that poverty and socio-cultural food laws and taboos could deny

Conclusion
The main objective of this study was to assess anaemia prevalence and associated factors among pregnant women attending ANC services in the West Gonja District. The study revealed a relatively high (56%) anaemia prevalence among the study respondents.
Timing of ANC and regular egg consumption were the strongest predictors of anaemia in pregnancy. These findings and discussion together suggest that awareness and knowledge about anaemia among pregnant women attending ANC alone may not even be sufficient to bring about reduced prevalence. Therefore, interventions need to go beyond awareness and knowledge creation through information provision to focusing on other important dietary, economic and cultural factors that may impact negatively on the possibility of getting anaemia in pregnancy. In this regard, health facility and community-based preconception and conception care interventions must not only aim to educate women and community members on the importance of early ANC initiation, balanced diet and sources of iron rich foods that may reduce anaemia, but must also engage community leaders to address issues related to food taboos and prohibitions during pregnancy that could expose pregnant women to adverse health outcomes, including anaemia.

What is known about this topic
 Anaemia affects 40-60% of pregnant women in lowincome countries;  Anaemia is a significant contributory factor to adverse pregnancy outcomes, including high maternal death, impaired mental development in children, increased risk of fetal growth retardation, low birth weight, premature delivery and perinatal mortality;  But its determinants are not exactly clear. We are grateful for this support. Table 1: demographic and economic characteristics