Study setting
The study was conducted among women attending antenatal clinics at Mkuranga district hospital and Irene-Kilimahewa health centre, coastal region, Tanzania. The two facilities were selected based on their location and number of villages that they are serving. Mkuranga district hospital is located at Mkuranga centre(town) while Irene kilimahewa is located at 36 km from Mkuranga centre. Mkuranga district hospital RCH is providing services to 7 villages within the districts and also it is a referral centre for all health centres in the district while Irene-Kilimahewa serves 8 villages and it receive referral from ten dispensaries within district. All two facilities are providing blood transfusion services.
Study design and sampling method
The study population were pregnant women of reproductive age (that is women ages 15 – 49 years). Sample size was estimated by using Fisher’s formula [17] n = Z2 P (1-P)/ Ꜫ2. Where n is estimated minimum sample size; Z is confidence level at 95% (standard value is 1.96); P is proportion (prevalence of anemia during pregnancy 53% TDHS, 2010); Ꜫ is precision at 95% CI = 0.05. The minimum sample that was required for this study was 399 pregnant women. A 5% non-response rate used to give a total sample size as 418 pregnant women.
Inclusion and exclusion criteria
The study included pregnant women attending ANC from first visit and above at Irene Kilimahewa health centre and Mkuranga district hospital RCH-ANC.The study excluded pregnant women that did not start first visit at Irene Kilimahewa health centre and Mkuranga district hospital (relocate). In addition, pregnant women who were not able to express themselves in either Kiswahili or English were excluded to participate in the study.
Data collection
Pregnant women aged 15-49 years attending ANC at Mkuranga district hospital and Irene Kilimahewa health center RCH-ANC were included. The study involved 418 participants who were conveniently sampled. Data were collected through structured questionnaire and blood sampling. The purpose of the study was explained to all eligible individuals. Those who accepted to participate were asked to sign the consent form.
Blood sample collection
The recently calibrated Celltac Es Nihon Kohden was used for haematology analyzation. To ensure the accuracy of the machine, laboratory technician run control test every day before starting the actual sample testing. The procedure of collecting sample for full blood picture test was as follow: Participants were required to go to the laboratory where they were instructed to sit upright on a chair and rest their arm face up on an elevated armrest. The laboratory technician tied a strap tourniquet around the top of their arms to temporarily restrict the blood flow from the arm back to the heart. This made the vein inside of client elbow “pop out,” and therefore easier to find. The area where the needle was inserted wiped with a sterile alcohol wipe to reduce any risk of infection. A needle inserted into the vein and a small amount of blood (4cc) drawn into the vial attached to the needle. After the procedure, laboratory technician press a small wad of cotton on the entry point to stop the flow of blood. The cotton wad strapped on with a band aid. Last, the participants were instructed that the cotton was only needed to remain on for a couple of minutes.
The findings obtained were recorded in the participant questionnaires. Based on World Health Organization (WHO) guideline, Hb level less than 11g/dl considered as anemia. Blood sample from Irene-Kilimahewa health centre were collected and kept in the cool box with the ice packs of 2-40c then transported to Mkuranga District Hospital where the test was conducted. The samples were tested within eight hours after been collected.
Tools and questionnaires
Questionnaire with structured questions used for collecting data that was assessing factors associated with anemia in pregnancy. Outcome variable of the study was anemia in pregnancy where by all pregnant women that was found with haemoglobin level <11g/dl were considered to be anaemic [6]. According to WHO, Anemia in pregnancy is categorized into three groups where by those with haemoglobin level of 10.0g/dl- 10.9g/dl considered to have mild anemia, 7.0g/dl- 9.9g/dl moderate anemia and <7g/dl severe anemia[18]. Independent variables of the study were socio- economic and socio- demographics variables that was assessed by 34 questions adopted from Tanzania demographic health survey[9].
Household food insecurity, was assessed by using the tool adopted FANTA and WHO[19]. The tool has 9 questions that required the women to recall her eating experience in the past one month time basing of the nine item questionnaire. The average Cronbach‟s alpha reliability coefficient for the instrument was 0.76. The lowest and highest values were 0 and 27 respectively. The scores were grouped into four categories; Food secure, mildly insecure, moderately insecure, and severely food insecure as recommended by the developer based on cut-off points. The tool was validated in developing countries including Tanzania [20].
Dietary diversity; pregnant women were asked to identify the type of food they took in the past 24 hours. A list of common food was adopted from tool developers. A list of 10 food groups provided by FANTA [21] was used to calculate the women dietary diversity score (WDDS). In this study, the WDDS had a mean score of 4.70 ±1.41 SD. Minimum dietary diversity was defined as it was instructed by the tool developers. The women who consumed five meals and above, considered to have minimal adequate dietary diversity. Also, the tool was validated[22]
Burden of disease; the impact of anemia in relation to other health conditions such as malaria was assessed through questions that was adopted from TDHS/MIS 2016. The tool was validated by the previous users within the country.
Data analysis
The analysis was conducted using STATA version 15. All probabilities were two-tailed and independent variables with p values <0.05 were regarded as significantly related with anemia. Descriptive statistics involving cross-tabulations was used to analyze categorical variables and results were presented in the form of frequency and percentage, while mean and standard deviation were presented for continuous variables. Logistic regression analysis was applied to determine factors associated with anemia among pregnant women. Bivariate regression was first fitted for each study variable to identify the independent variables that were associated with anemia. Variables that were significant in bivariate analysis with (P= 0.05) were then included in a multivariate analysis to obtain the adjusted factors associated with anemia. The results of the model were presented using odds ratios (OR) and 95% confidence interval (CI).