Predictors for the Uptake of Optimal Doses of Sulfadoxine-pyrimethamine for Intermittent Preventive Treatment of Malaria During Pregnancy in Tanzania: Further Analysis of 2015-2016 Tanzania Demographic and Health Survey and Malaria Indicator Survey Data

Background: In Tanzania, the uptake of optimal doses ( ≥ 3) of sulfadoxine-pyrimethamine for intermittent preventive treatment of malaria (IPTp-SP) during pregnancy has remained below the recommended target of 80%. Therefore, this study aimed to investigate the predictors for the uptake of optimal IPTp-SP among pregnant women in Tanzania. Methods: This study used data from the 2015-16 Tanzania demographic and health survey and malaria indicator survey (TDHS-MIS). The study had a total of 4,111 women aged 15 to 49 who had live births 2 years preceding the survey. The outcome variable was uptake of three or more doses of IPTp-SP and the independent variables were age, marital status, education level, place of residence, wealth index, occupation, geographic zone, parity, the timing of rst antenatal care (ANC), number of ANC visits and type of the health facility for ANC visits. Predictors for the optimal uptake of IPTp-SP were assessed using univariate and multivariable logistic regression. Results: A total of 327 (8%) women had optimal uptake of IPTp-SP doses. Among the assessed predictors, the following were signicantly associated with optimal uptake of IPTp-SP doses; education level [primary (AOR: 2.2, 95% CI 1.26–3.67, P = 0.005); secondary or higher education (AOR: 2.1, 95% CI 1.08–4.22, P = 0.029)], attended ANC at the rst trimester (AOR: 2.4, 95% CI 1.20–4.96, P = 0.014), attended ≥ 4 ANC visits (AOR: 1.9, 95% CI 1.34–2.83, P <0.001), attended government health facilities (AOR: 1.5, 95% CI 1.07–1.97, P = 0.017) and geographic zone [Central (AOR: 5, 95% CI 2.08–11.95, P <0.001); Southern Highlands (AOR: 2.8, 95% CI 1.15–7.02, This study was a cross-sectional study utilizing the TDHS-MIS dataset of 2015-16. This was a sixth in a series of nationally representative household surveys conducted in Tanzania. It was conducted 2015 to February 2016 to provide up to date data on the areas of health, population (demographics) and nutrition. It was done under collaboration of Demographic and Health Surveys programme, National Bureau of Statistics, Oce of Chief Government Statistician and Ministries of health in Tanzania Mainland and Zanzibar.


Background
Malaria in pregnancy is still a major public health problem causing maternal, fetus and neonatal adverse health effects such as spontaneous abortion, maternal anemia, stillbirth, premature birth, low birth weight and maternal death (1). Among the Plasmodium species; Plasmodium falciparum is the leading cause of maternal illness and low birth weight due to malaria especially in Africa (2). The estimates showed that 125 million pregnant women are at risk of acquiring malaria globally with 10,000 maternal death and 200,000 neonatal death annually as a result of malaria in pregnancy (3). Also, malaria accounts for 18% of severe anemia in pregnancy (4). The occurrence of malaria in pregnancy has been associated with low gestational age, low maternal age, low parity, non-uptake of SP during pregnancy, level of knowledge on malaria prevention, place of residence of a woman and household wealth status (5,6).
The World Health Organization recommends interventions such as the use of intermittent preventive treatment in pregnancy with sulfadoxine-pyrimethamine (IPTp-SP) and folic acid supplementation as a part of antenatal care services to prevent and treat malaria and anemia during pregnancy in areas with moderate to high malaria transmission. However, the IPTp-SP should be combined with the use of insecticide treated nets (ITNs) and effective case management of malaria (2). SP uptake starts in the second trimester and pregnant women should receive at least 3 doses with an interval of one month during the pregnancy (7). The evidence shows that the scale-up of malaria prevention interventions (IPTp and ITNs) in 25 malaria endemic African countries has lowered neonatal mortality and low birth weight by 18% and 21% respectively (3).
Tanzania is one of the malaria-endemic countries with approximately 1.7 million pregnant women at high risk of malaria infection (8). The trend of malaria prevalence among pregnant women in Tanzania has been uctuating with the highest of 10.3% in 2014 and the lowest of 6.8% in 2017 (9). It was observed that despite the low prevalence of malaria in some areas such as in Zanzibar Island, 0.8% of the pregnant women had malaria parasitemia during delivery (10).
Tanzania decided to opt for IPTp-SP policy in 2001 and for the revised policy of using ≥ 3doses of IPTp-SP in 2013. The Tanzania demographic and health survey and malaria indicator survey of 2015/16 has reported the uptake of sulfadoxine-pyrimethamine (SP) among pregnant women to be 68% for the rst dose, 35% for the second dose and 8% for the third dose (11). The uptake of optimal doses (at least three or more doses) of SP among pregnant women in Tanzania is still low compared to required optimal coverage of at least 80% (3). The low coverage of IPTp-SP was also observed in sub-Saharan Africa with the median coverage of 64%, 38% and 23% for rst, second and third doses respectively (12). The use of less than recommended doses of SP is less bene cial in prevention of maternal and fetus/neonatal adverse health effects as it has been proven that the use of three or more doses to have bene cial effects against malaria related health effect to both of maternal and fetus/neonatal (1).
Globally, studies have reported potential determinants associated with the uptake of at least two doses of SP to be; age of a woman, place of residence, education level of a woman, household social-economic status, knowledge of a woman about malaria and IPTp-SP, number and timing of antenatal care (ANC) visit (12)(13)(14)(15)(16)(17). It was observed that educated and wealthier women were knowledgeable about IPTp and more likely to receive the SP (18). Also, the uptake of SP was higher for women who attended 3 to 4 ANC visit and started ANC visit in their rst or second trimester compared to those who started at third trimester (13,16,17). The predictors for the uptake of optimal (≥ 3) IPTp-SP doses for malaria prevention in Tanzania have not been fully investigated. Therefore, this study aimed to analyze the predictors for the uptake of the optimal doses of SP among pregnant women who gave birth 2 years preceding the 2015-16 TDHS-MIS. The ndings shall inform the policymakers and programmes implementers on the necessary changes to be done to improve the uptake of SP to optimal doses as recommended by the WHO.

Study design and data source
This study was a cross-sectional study utilizing the TDHS-MIS dataset of 2015-16. This was a sixth in a series of nationally representative household surveys conducted in Tanzania

Study Population And Sample Size
The study population was women of reproductive age (15-49) who had a live birth in the two years before the survey. The data was extracted from women individual recode le (TZIR7BFL) which had a total sample size of 4,128 women who gave birth two years preceding the survey. However, 17 women were dropped because of incomplete information on the outcome variable. Therefore, the total sample size used for this study was 4,111 women (unweighted sample) and 4056 women (weighted sample).

Sampling Technique
The 2015-16 TDHS-MIS employed strati ed two-stage cluster sampling. The rst stage involved the selection of clusters contained enumeration areas in which 608 clusters were selected. The second stage involved the systematic selection of the households from the 608 selected clusters in which 22 households were selected in each cluster. This sampling technique produced a probability sample of 13,376 households whereby only 12,767 households were occupied. In the occupied households, a total number of 13,266 women were interviewed and out of the interviewed women, 4,128 gave birth two years before this survey. Hence, they were eligible for inclusion in this study (11).

Measurement Of Variables
The variables associated with uptake of IPTp were extracted from women's data set based on the literature review. The outcome variable was uptake of three or more doses of IPTp-SP. The explanatory variables were socio-demographic characteristics (age, marital status, level of education, place of residence, geographical zone of residence, wealth status) and obstetric characteristics (parity, timing of rst ANC visit, number of ANC visits and type of health facility used for ANC services). The summary of variable de nitions and categories are shown in Table 1.

Data Management And Analysis
The data was extracted, cleaned and analyzed using STATA version 14 (STATA Corp, College Station, Tx, USA). Descriptive statistics analysis was rst conducted on socio-demographics and obstetric characteristics to obtain frequencies and proportions. Chi-square test was carried out to describe the association between explanatory variables (socio-demographic/obstetric variables) and the outcome (the optimal uptake of IPTp-SP) at a signi cance level of 5%. Then, binary logistic regression (univariate) was conducted to all variables with statistical signi cant in chi-square test and results were reported as crude odds ratio. All explanatory variables with a P-value < 0.25 on univariate analysis were subjected to multivariable logistic regression for further analysis of association to obtain adjusted odds ratios. To account for the differences in sampling probabilities across the clusters and strata, we used the sample weighted to adjust for the cluster sampling design.

Ethical Consideration
This study used secondary data without involving any human subjects. Therefore, no formal ethical approval was required. However, the Tanzania Demographic and Health survey was conducted after approval from national and international review boards including the National Institute of Medical Research, Zanzibar Medical Research Ethical Committee, Institutional Review Board of Inner City Fund and the Centers for Disease Control and Prevention in Atlanta. All women interviewed were requested to provide verbal informed consent before the commencement of the study. The permission to use the IPTp-SP data was sought and obtained from the DHS program.

Socio-demographic characteristics of the study respondents
This study included 4,111 women, majority (27.3%) were between the ages of 20-24 years. The majority of the respondents (75.9%) lived in rural areas and 83% were in a union. More than half of the respondents (59.6%) had attained primary education and 51.4% were self-employed. Also, the highest number of respondents came from the regions of Lake Zone (27.4%) and belonged to the poorest wealth quintile (22.5%) as shown in Table 2. There was a signi cant association between the optimal uptake of SP doses (3 or more doses) and woman's education level (p = 0.0003), place of residence (p = 0.0002), wealth index status (p = 0.0002), occupation (p = 0.0326), and geographical zone (p < 0.0001). The optimal uptake of IPTp-SP doses increased by level education ranging from 3.9% for women with no formal education to 10.7% for women with secondary and above the secondary level of education. For the place of residence, more women in the urban areas (11.6%) took optimal doses of IPTp-SP as compared to women in rural areas (6.7%). For the wealth index, the results show that the richest women (13.1%) took more optimal doses of IPTp-SP as compared to the poorest women (6%). For the geographical zone, the uptake of optimal doses of IPTp-SP was lowest in the Western zone (2.1%) and highest in the Eastern zone (15.3%) as presented in Table 3. Distribution of obstetric characteristics and ANC service provider by optimal uptake of IPTp-SP This study revealed a signi cant association between optimal uptake of IPTp with parity (p = 0.0337), timing of 1st ANC visit (p < 0.0001), number of ANC (p < 0.0001), and type of health facility that a woman was attending for ANC (p = 0.0113). Women with one child and two children (9.3% and 9.8% respectively) received optimal doses as compared to women with three or more children (6.8%).Women who started ANC in the rst trimester (12.6%) and attended four or more ANC visits (11.7%) had higher uptake of optimal doses compared to those who started ANC late and attended 1-3 ANC visits. Also, the uptake of optimal doses of IPTp-SP was high for women who attended government health facilities (10.4%) as compared to those who attended non-government facilities (7.4%) during their ANC visits (Table 4).

Discussion
This study used data from TDHS-MIS 2015/16 to analyze the predictors for the uptake of optimal doses of SP (three or more doses) among pregnant women. The uptake of three or more doses of SP was reported to be 8% countrywide which is still low than the recommended coverage of 80% from WHO and Roll Back Malaria (RBM) benchmark target (3). This was also observed in several studies conducted in other sub-Saharan countries (20)(21)(22)(23). Hence, the urgent need to plan effective strategies to improve IPTp-SP coverage and uptake in sub-Saharan Africa.
The predicators for the uptake of optimal doses of SP were; geographical zones, education level (primary, secondary or higher education), attending ANC in the rst trimester of pregnancy, attending ANC visit more than four times and attending government health facility for ANC services. Pregnant women who attained at least primary education were likely to receive optimal doses of SP compared to those with informal education. This is because educated pregnant women could be aware and knowledgeable on the importance and bene ts of using SP for malaria prevention during pregnancy. Similarly, the ndings from Nigeria (24), Malawi (21,25), Ghana (22), and Zimbabwe (23) showed that the knowledge on the SP and on the consequences of not taking IPTp-SP as a facilitator toward the uptake hence the association between education level and the likelihood of the uptake of three or more doses of SP for malaria prevention during pregnancy.
Pregnant women who registered and attended ANC clinics in their rst trimester received optimal doses of SP compared to those attended ANC clinic in third trimester. The possible explanation could be; attending ANC clinics in rst trimester give the room for pregnant women to attend ANC for more than 4 times hence higher chances for start taking SP doses in their second trimester as required. Also, it has been predicted in several studies conducted in Zimbabwe, Sierra Leone, Malawi, Nigeria and Uganda that early booking and attending of rst ANC in rst or second trimester has an association with receiving optimal doses of SP while late attending to ANC clinic results in lower uptake of SP doses (6,15,23,24,26).
The signi cant relationship between number of ANC visit and uptake of optimal doses of SP was observed in our study. The pregnant women who attended at least four ANC visit received optimal doses of SP compared to those with few attendance. The more the pregnant women attend to clinic, the higher the exposure toward health information on IPTp-SP hence the higher likelihood of receiving optimal doses of SP. The ndings are consistent with the studies conducted in Malawi, Ghana and Cameroon (21,27,28). Also, attending ANC visits only once or at late such as after 36 weeks where SP cannot be administered were observed to be a barrier towards the uptake of optimal doses of SP (25). Therefore, the urge to raise awareness among pregnant women on the importance of early and adequate attendance to ANC clinics so as to receive optimal doses of SP for malaria prevention is important.
Attending government health facilities ANC was found to in uence the uptake of optimal SP doses among pregnant women compared to those who attended private clinics. The plausible explanation could be sensitization of the SP uptake under direct observation therapy (DOT) and seriousness on following SP administration protocol. It was noticed that in some private clinics pregnant women were allowed to take the drugs at home hence compromise the optimal uptake of SP doses. The ndings are in accordance to a study conducted in Ghana which found poor adherence to DOT in private health facilities as one of the obstacles towards the uptake of optimal doses of SP (29).
Geographical zones were also the predictors for the optimal uptake of SP doses. Being a resident of regions that belong to Central, Eastern, Southern, Lake, Southern highlands, and South west highlands was signi cantly associated with the optimal uptake of SP doses compared to the residents of Zanzibar and Northern zones. This might be contributed by level of malaria endemicity in different zones. In the zones with a high or moderate level of malaria transmission, possibly the awareness and emphasis on SP uptake could be higher due to higher risk of contracting malaria that's why pregnant women in those zones had higher odds of taking optimal doses of SP compared to those residing at Zanzibar and Northern zones where there is a low level of malaria transmission. The observed ndings are in lines with another study conducted in Tanzania which showed that pregnant women residing in Eastern and Coastal regions had higher odds of optimal uptake of SP (26).
This study had the following limitations; the data analysis was limited only on the variables captured on demographic and health survey questionnaire, some of the important variables that could in uence uptake of optimal doses of IPTp-SP were not captured for example socio-cultural factors, knowledge of health care providers and availability of SP in ANC clinics hence hindered full exploration of other important variables. Response (recall) bias was another limitation, the data collection was based on selfreported experiences of the past two years hence due to response bias there was a possibility of over-orunderestimation of the responses.

Conclusions
The uptake of optimal doses of SP among pregnant women in Tanzania is still below the WHO recommendations. The identi ed predictors for optimal uptake of SP were primary and secondary or higher education level, attending ANC in the rst trimester, attending ≥ 4 ANC visits, attending a government health facility for ANC services and being a resident of any geographical zone except for Northern and Zanzibar zones. The alarming ndings indicate the urgent need to improve the uptake of optimal doses of SP among pregnant women in Tanzania. Therefore, we recommend health education, social and behavior change interventions with the emphasis on the earlier attendance to ANC clinics and on the optimal use of IPTp-SP doses. The mentioned interventions will help to improve the awareness and knowledge of the optimal use of IPTp-SP among pregnant women in Tanzania. Also, private ANC clinics should adhere to the implementation of the DOT policy to improve the uptake of optimal doses of SP among pregnant women who attend private ANC clinics.