Predictors of Anaemia prevalence among pregnant women in urban Ghana: a cross-sectional study

Background Anaemia during pregnancy is a major public health concern. Despite its wide scope and adverse effects including increased maternal and perinatal morbidity and mortality, and long-term adverse effects in the new born, extensive interventions using upstream approaches to public health have largely not been implemented. This study investigated the prevalence and associated factors of anaemia in pregnant women in four health facilities in the Kwabre East Municipality of Ghana.Method A cross-sectional survey with a two-stage sampling technique was conducted on 220 pregnant women who attended antenatal care at the selected health facilities. Interview-based structured questionnaires were used. Bivariate and multivariate logistic regression were used to identify predictors.Results The prevalence of anaemia was 11.4%. Few women (25) were anaemic and morphologically, 14 had normocytic normochromic (56%) anemia and 9 had Microcytic hypochromic (36) anaemia. Iron de�ciency was reported in 19 (8.6%) pregnant women. Iron sulphate intake (AOR [95% CI] = 3.16 [1.15, 7.37], ANC follow-up during pregnancy (AOR [95% CI] = 3.07 [1.59, 7.99], household size of ≥ 5 (AOR [95% CI] = 3.58 [1.75, 9.52], folic acid intake (AOR [95% CI] = 5.29 [2.65, 12.39] and the period in pregnancy AOR [95% CI]= ≥ 36 weeks 3.2 (1.3–4.5) were independent predictors of anemia.Conclusion Though anaemia prevalence has been low in urban areas as previously reported, collaborated healthcare measures that aim at eradicating the menace are encouraged. Maternal health care interventions including the administration of folic acid, regular iron sulphate intake and intensive education on early ANC are recommended.


Background
Anaemia is a major public health problem globally among pregnant women in both low and middle-income countries (WHO, 2015;Gupta and Gadipudi, 2018).The menace exerts serious consequences on the health of pregnant women and the unborn baby (WHO, 2008;McLean et al., 2009).In a report released in 2011 by the World Health Organisation (2015), the global anaemia prevalence rates for pregnant women and all women of reproductive age were 38.2% and 29.4% respectively.The prevalence rates of anaemia are more alarming in the regions of South-East Asia, Eastern Mediterranean and Africa, which range from 37.7% to 41.5% for non-pregnant women and 38.9% to 48.7% for pregnant women in these regions, respectively (WHO, 2015).It is disheartening that the highest prevalence of anaemia is among children under ve years and pregnant women (Balarajan et al., 2011;Salhan et al., 2012;WHO, 2015).
Anaemia during pregnancy is a major cause of maternal morbidity (Balarajan et al., 2011), and is estimated with a haemoglobin concentration of less than 10 gm/dl.Any patient with haemoglobin of less than 11 gm/dl to 11.5 gm/dl at the start of pregnancy will be treated as anaemic.Anaemia during pregnancy has both maternal and foetal consequences, hence, a major cause of morbidity and mortality among pregnant women especially in developing countries (Akhtar and Hassan, 2012;Elzahrani, 2012;Ali et al., 2019).It is in this regard that the WHO requires blood testing to detect syphilis and severe anaemia as major components of antenatal care among pregnant women (GSS, 2011;WHO, 2015).
The causes of anaemia during pregnancy is multidimensional, however, in developing countries, it is predominantly nutritional de ciencies in iron, folate and vitamin B12.Other causes of anaemia are loss of blood due to bleeding, failure of bone marrow to produce su cient number of RBCs causing aplastic anaemia, and parasitic diseases (Salhan et al., 2012;Chaparro et al., 2019).Iron de ciency is the most common cause of anaemia and undermines the cognitive development of children and low productivity among adults (Baig-Ansari et al., 2008).Anaemia infections among pregnant women is a public health problem in Ghana (WHO, 2008).However, among the regions in Ghana, anaemia prevalence ranges from as low as 2 percent in the Eastern region to 19 percent in the Northern, Upper East and Upper West regions.There is also a large variance in anaemia prevalence by ecological zone (GSS, 2012).Anaemia was three times higher in the Savannah zone (15%) than in the Coastal (4%) and Forest (5%) zones (GSS, 2012).
In the Kwabre East Municipality, the trend of anaemia pregnancy cases from 2009-2012 has been progressively encouraging.In 2009, there were 1,851 (37.8%) reported cases, 2010 had 1,428 (29.6%) cases, 2011 witnessed 1,115 (24.3%) cases and in 2012, 1,032 (23.2%) cases were reported.In 2013, 825 pregnancy anaemia cases were recorded with 430 reported in 2014 (KEMHD, 2015).There have been downward trend in anaemia prevalence among pregnant women in the municipality.Anaemia prevalence among pregnant women is a major indicator used to assess the effectiveness of interventions focused on maternal healthcare.According to the Municipal Health Directorate (2015), the key reason for low anaemia prevalence is the implementation focused ANC by the government of Ghana, which helps pregnant women to be taken care of by a particular healthcare provider through the period of pregnancy.Despite the intervention, the prevalence still exists, though in lower rates.There is therefore, the need to undertake a study to determine the factors contributing to anaemia prevalence in the municipality six years after the implementation of the focused ANC.The ndings from this study therefore, would be useful in informing maternal health policy and interventions in Ghana.

Study setting
The study area is one of the 32 newly carved Municipalities by the government of Ghana in 2018.The Kwabre East Municipality was carved out of the former Kwabre Sekyere District in 1988.It is located almost in the central portion of the Ashanti region and within latitudes 60 44° North and longitudes 10 33° to 1044° West (GSS, 2012).The Municipality has a total land area of 246.8 square kilometres constituting about 1.01% of the total land area of Ashanti Region.The population of the Municipality according to the 2010 Population and Housing Census stands at 115,556 with 55,106 males 60,450 females.Data from the Kwabre East Municipal Health Directorate (2015), reveal that ANC utilisation level is high and most pregnant women make use of the service purely on the intense education the health directorate has embarked towards achieving the WHO standards of service use.

Study Design and Sampling
This study forms part of a larger research that investigated the socioeconomic determinants of antenatal care utilisation of pregnant mothers in selected health facilities in the Kwabre East Municipality.The study adopted a facility-based cross-sectional survey design to investigate the predictors of anaemia prevalence among pregnant women in the Municipality.These health facilities are publicly owned and were purposively sampled based on the administrative divisions of the Municipality and the free maternal health policy introduced by the Ghana government in 2003.
The study used a three-stage sampling technique to select the respondents.At the rst stage, purposive sampling technique was used to select the four health facilities located in Asonomaso, Mamponteng, Antoa, Sakra Wonoo in Municipality based on administrative divisions.Using the Municipal Health Information Management System software, a total of 1010 pregnant women constituted the sampling frame from the four health facilities.Consequently, the proportional sampling technique was used to select 220 pregnant women aged between 15-49 were selected for the study (Table 1).Structured questionnaires were used to collect data from study participants.Data collected from respondents included demographic characteristics, health status, family planning services, social support systems, quality of service, choice of facility, level of utilisation, and views on caregivers.
The sample size was obtained by using the fractional approach, which suggests that any fraction of an intended population that is 10% or more is deemed representative especially if the sample is scienti cally selected (Monette, 2005).In a related study by Bush and Burns (2000), it was revealed that a sample size of 5% of a population of study is deemed to achieve its purpose if it was scienti cally chosen.Hence, the sample size of 22.6% of the total population from the selected health facilities was deemed to be scienti cally chosen.The proportional strati ed sampling technique was used to determine the sample sizes using the formula; n = (ni/N)S, where ni = facility population, N = total population and S = sample size.The same formula was used based on the variance in population of study respondents from selected facilities.As a result, the margin of error for the determination of the sample size was 0.02 and the proportion of the study population likely to agree that pregnancy anaemia is a major public health concern was assumed as 10% (0.10).The gures were then substituted into the formula outlined by Moser and Kalton (2007) for determining sample size.The formula is given as: [Due to technical limitations, this equation is only available as a download in the supplemental les section.]

Data Collection and analysis
The selection was based on the level of antenatal care attendance at the selected prenatal outpatient departments of selected facilities.Using the Municipal Health Information Management System software (KEMHD, 2015).In effect, a list ofpregnant women aged between 15-49 was obtained from the antenatal care units of the sampled health facilities.Subsequently, simple random sampling technique was used to select 220 pregnant women for the study.Data collected from respondents included demographic characteristics, health status, family planning services, social support systems, quality of service, choice of facility, level of utilisation, and views on caregivers.
A face-to-face interview using structured questionnaire was employed to obtain data about sociodemographic characteristics, dietary habit, ANC followup, multivitamin intake, taking meat and animal products, sulphate intake, folic acid intake, sleeping under insecticide treated nets (ITNs), use repellents/coils/sprays and nutrient supplementation at their pregnancy period at the time of data collection.Blood pressure, height and weight and body mass index (BMI) of respondents were calculated in the forms of (weight (kg)/height (m²)).Serum ferritin levels using the DPC Immulite 1 ferritin assay (Diagnostic Products Corporation, Los Angeles, USA) were taken.Again, erythrocyte sedimentation rate (ESR) were measured using the Sediplast system (LP Italiana Spa, Milan, Italy); Diagnostic Products Corporation, Los Angeles, USA); complete blood counts (CBC) were performed using Sysmex KX-21 Hematology Analyzer (Sysmex Corporation, Kobe, Japan) were run on venous blood samples.The WHO criteria on anaemia (haemoglobin concentration b11 g/dL, (Beutler and Waalen, 2006)) were strictly used in this study.Also, urine and stool samples using iodine were also obtained to determine the presence of pathogens in the intestines of study participants.
Raw data for the study were rst entered into SPSS version 21 and then imported into STATA version 20 statistical package for analysis.Descriptive statistics in the form of standard deviation, means, and percentages were carried out using tables.Again, the binary logistic regression analysis was to evaluate the difference in anaemia prevalence in relation to the relevant variables.To assess the strength of association and statistical signi cance in bivariate analysis, odds ratio, and 95% CI for odds ratio were calculated to determine their signi cant levels.

Characteristics of Study Participants
A total of 220 pregnant women with a mean (±SD) age of 30.46 ± 6.16 were included in the study.43% (95) of study respondents were at the rst trimester, followed by 38% (83) in their second trimester with 19% (42) of respondents in their third trimester (Figure 1).Table 2 summarises the characteristics of the study participants.
The dietary pattern and nutritional assessment of the respondents revealed that, 40.5%, 68.2%, 32.3%, 73.2%, 90.9% and 68.8% took folic acid, fruit after meals, green leafy vegetables, meat and other animal products, iron sulphate and multivitamin, respectively.Among study respondents, 31.8% did not take fruit after meals during pregnancy, 67.7% did not take leafy vegetables, 13.2 did not take multivitamin, 26.8 did not take meat and other animal products, and 59.5 did not take folic acid in their pregnancies (Table 3).
On the health characteristics on haemoglobin levels of pregnant women, participants were split into two groups, early; ≤24 weeks and late; ≥36 weeks.Using the stepwise algorithm on multivariate analysis, the odds of being iron de cient among pregnant women who were ≥36 weeks of pregnancy (AOR 3.2) were more likely to be increased than their counterparts in the ≤24 (AOR 1) weeks group (Table 4).Again, the odds of being anaemic were increased if women were iron de cient (AOR 2.4).The study presents low prevalence of anaemia and iron de ciency, which are 11.4% and 8.6% respectively.
The study parameters considered age, household size, occupation status and education of pregnant mothers if they in uenced in uenced anaemia prevalence among study participants.These variables were then added in the multiple regression analysis to determine their level of signi cance.Their results show that the z statistic presented marginal effects at variance at 10%, 5% and 1% (Table 6).This indicates that anaemia prevalence among older pregnant women compared to their younger counterparts is at variance among a particular population group.

Prevalence and Predictors of Anaemia
According to the Ghana multiple Indicator Cluster Survey (GSS, 2011), iron de ciency is one of the major causes of anaemia among pregnant women.This study revealed iron and anaemia prevalence rates of 11.4% and 8.6% among study respondents.These rates are considerably lower than most of those previously reported including those of Adesina et al., (2011) reported 33.1% andNybo etal., (2007) reported 20.0%.Again, Thirukkanesh and Zahara (2010) reported anaemia prevalence of 42.3% and Aikawa et al., (2006) reporting anaemia prevalence as 43.2%.In a related study, Lassey et al., (1999) indicated that 56% of urban pregnant women in Ghana had haemoglobin levels below the 11 g/dL threshold while Mockenhaupt et al., (2000) reported on the prevalence of anaemia and iron de ciency as 54% and 46% respectively on 540 rural pregnant women.Again, the ndings of this study is in line with the work of Engmann et al., (2008) on anaemia and iron de ciency in pregnant Ghanaian women from urban areas who reported anaemia and iron de ciency prevalent rates as 34% and 7.5, respectively.The possible reason for the low rates in anaemia and iron de ciency anaemia among pregnant women in the present study could be due to the progressive improvement of ANC services alongside the focused ANC adopted by the Kwbre East Municipal health directorate.
The variations in anaemia prevalence among these studies could also be attributed to the differences in broader health systems and health responses targeted at the well-being of pregnant women.The difference in anaemia prevalence in the current study could further be attributed to improvement in the standard of living, with personal hygiene and the planned initiative of the government of Ghana to meet the sustainable development goal 3 by 2030.In a study by Akowuah et al., (2018), it was found that the Ghana health services has instituted focused antenatal care which enables individual pregnant mothers to receive health care from a particular health staff in the entire period of pregnancy.The ndings indicate that the predictors of anaemia among pregnant women are iron sulphate, low use of ANC during pregnancy, large household size, and folic acid.This supports observations from Ghana (GSS, 2010;Fuseini et al., 2010;Nonterah et al., 2019), in Tanzania (Ali et al., 2019) and Nicaragua (Lubbock and Stephenson, 2008) which revealed that older pregnant women are more susceptible to be anaemic compared to their younger counterparts.In 5 addition, a comparative study between Indonesia and Ghana by Mocking et al., (2018) revealed similar study outcomes.
In the study, logistic regression analysis indicated that age was a key predictor of anaemia among pregnant women with beta coe cient and z statistic of 1.44 and 8.27 respectively.This observation is in agreement with Melku at al., (2014), who in their work on the prevalence and predictors of maternal anaemia during pregnancy in Gondar, Northwest Ethiopia, found that older pregnant women were more prone to anaemia than their younger counterparts.Possible reason for age as a predictor of maternal anaemia might be due to the poor feeding pattern of older pregnant women and the experience on safe multiple pregnancies, which do not make most of them attend ANC at early stages of pregnancy.The study demonstrated that mothers who have family size of more than ≥4 members were more likely to be anaemic (AOR = 3.58, 95% CI: 1.75, 9.52) as compared to their counterparts with ≤2 low family sizes (AOR = 1).This nding is in agreement with the work of Ali et al., (2019) which indicates that the higher the family size of pregnant women, the less are they are able to fully utilise ANC services partly because of nancial constraints they may beset with.Thus, such women easily become anaemic.However, the ndings of Al-Mehaisen et al.'s, (2011) study in Jordan showed no signi cant difference of anaemia prevalence and household size.This variation could partly be due to geographical differences in study areas as in the case of Jordan.

Limitations of the Study
One of the limitations of this study is that the number of pregnancies and childbirths of respondents in relation to history of anaemia prevalence was not enquired.Again, the study could not determine which of the predictors preceded the other, because of the cross-sectional study design used.The nal limitation of the study was the exclusion of both tertiary and primary health facilities as the study was executed in secondary care facilities that were in primarily urban centres.

Conclusion
The study reports a relatively low prevalence of anaemia and iron de ciency among pregnant women in urban Ghana compared to those previously reported.The focused ANC organised by the Municipal Health Directorate for pregnant women could be among the reasons for the low prevalence.In terms of morphological types of anaemia, the predominant types reported by the study was normocytic normochromic, microcytic hypochromic and macrocytic normochromic.The main predictors of maternal anaemia were age of pregnant women, household size, ANC follow-up during pregnancy, folic acid intake in present pregnancy, the sleeping under ITNs.In order to control the prevalence levels of anaemia and iron de ciency, the broader healthcare system is encouraged to make available folic acid in facilities and pregnant women must be encouraged to take more folic acid in addition to fruits at all times especially during pregnancies.Intense education on the need to fully utlise ANC must be strengthened to ensure early detection of potential pregnancy-related complications.Again, education on maternal health promotion should include sleeping in ITNs.In addition, programmes and policies aimed at reducing pregnancy anaemia should target mothers with large household sizes since such families tend to be nancially handicapped.Table 1:  Table 2: Sociodemographic characteristics of pregnant women on anaemia prevalence ( = 220).