Background: Intermittent preventive treatment of malaria in pregnancy with 3+ doses of sulfadoxine-pyrimethamine (IPTp-SP) reduces maternal mortality and stillbirths in malaria endemic areas. Between December 2014 and December 2015, a project to scale up IPTp-SP to all pregnant women was implemented in three local government areas (LGA) of Sokoto State, Nigeria. The intervention included community education and mobilization, household distribution of SP, and community health information systems that reminded mothers of upcoming SP doses. Health facility IPTp-SP distribution continued in three intervention (population 661,606) and one counterfactual (population 167,971) LGAs. During the project lifespan, 31,493 pregnant women were eligible for at least one dose of IPTp-SP. Methods: Community and facility data on IPTp-SP distribution were collected in all four LGAs. Data from a subset of 9427 pregnant women, who were followed through 42 days postpartum, were analysed to assess associations between SP dosages and newborn status. Nominal cost and expense data in 2015 Nigerian Naira were obtained from expenditure records on the distribution of SP. Results: Eighty-two percent (n = 25,841) of eligible women received one or more doses of IPTp-SP. The SP1 coverage was 95% in the intervention LGAs; 26% in the counterfactual. Measurable SP3+ coverage was 45% in the intervention and 0% in the counterfactual LGAs. The mean number of SP doses in the intervention LGAs was 2.1; 0.4 in the counterfactual. Increased doses of IPTp-SP were associated with linear increases in newborn head circumference and lower odds of stillbirth. Any antenatal care utilization predicted larger newborn head circumference and lower odds of stillbirth. The cost of delivering three doses of SP, inclusive of the cost of medicines, was US.93-ce:para.20. Conclusions: It is feasible, safe, and affordable to scale up the delivery of high impact IPTp-SP interventions in low resource malaria endemic settings, where few women access facility-based maternal health services.
From the HRS database, 25,572 women who were eligible for SP between April and November in the three intervention LGAs, were identified. Based on actual numbers, 94% of pregnant women in the intervention LGAs became eligible for SP1 between April and November. In the intervention LGAs, the number of pregnant women eligible to receive SP1, was used as a denominator to calculate SP1, SP2 and SP3+ coverage (Table 4). Determining the number of women eligible for SP1 between April and November 2015, and differences between project mapping and official projections aMean percentage of women becoming eligible for SP1 in the intervention LGA’s bEstimated using the mean percentage of pregnant women that became eligible for SP1 in the intervention LGAs In the counterfactual group, the number of pregnant women eligible to receive SP1 over the life of the project was estimated. Using the official population estimate for Yabo LGA, from the Sokoto State Government (167,971) and assuming that 5% of the population was pregnant, there were an estimated 8399 pregnant women in Yabo, in 2015. Prorating that number for the 8 months of the project, there were an estimated 6299 pregnant women in Yabo between April and November 2015. Assuming Yabo would have the same percentage of women eligible for SP1, as found in the intervention areas (94%) it was estimated that there was a total of 5921 women eligible for SP1 during the project lifespan (Table 4). This estimate as the denominator to calculate SP1 and SP2 coverage. Univariate and bivariate analyses were performed to compare intervention and counterfactual LGAs on the number of SP doses and source of SP. Analyses were conducted in Excel spreadsheet®. Programmatic data extracted from outcome forms were available for 9241 live births in intervention and counterfactual LGAs between April and November. Head circumference (analyzed in mm) of live newborns measured within 7 days of birth, was available for 6720 (73%) of live births. Head circumference data were missing for 2521 live newborns. Of these, 1721 (19%) mother-newborn dyads missed initially, were identified during the data quality review in October. For the remaining 800 newborns, head circumferences were not measured within 7 days postpartum. Independent variables used in this analysis were guided by prevailing epidemiological evidence base and by what was feasible to collect by CBHV supervisors. These included the sex of newborn, gravidity of mothers, successive doses of SP, exposure to ANC, month of delivery and gestational age at birth. Globally, female newborns have smaller head circumferences than male newborns [15]. Primigravida women are at higher risk for placental malarial infection, [35]. Table 1 presents how each of these variables was coded in the analyses. Variables used to understand the impact of SP interventions on head circumference among 6720 live newborns, born between May and November 2015 Univariate analyses tested for any associations between a given independent variable and the mean head circumference of live newborns. Unadjusted t tests were used to assess any differences within each predictor variable and newborn head circumference. Unadjusted tests of correlations between mean head circumference and doses of IPTp-SP, and month of birth, were performed. Mean newborn head circumferences were used to assess correlations in the number of IPTp-SP doses over the project period in the intervention and counterfactual LGAs. A multivariate linear regression model was used to test for the impact of SP doses and other variables on head circumference. Analyses were performed with Excel® and SAS v. 9.4. Data extracted from outcome forms were available for 9453 term births in both the intervention and control LGAs between April and November. To examine the impact of IPTp-SP doses on the incidence of stillbirths, all confirmed pregnancies that ended in miscarriage and abortion, or were delivered before 8 months of gestational age (n = 99) were excluded. If the newborn was stillborn, it was coded as “1”; if it was a live birth, it was coded as “0.” Stillbirth rates (SBR), per 1000 births, and correlations between SBR and doses of SP were calculated. Unadjusted and adjusted logistic regression modelling was used to predict the odds with 95% confidence intervals, of having a stillbirth among women who ingested different doses of SP, according to exposure to intervention, those who attended at least one ANC visit, gravidity of mothers, those who gave birth in a facility vs those who gave birth at home, and those who gave birth later (July–November) vs. those who gave birth earlier (April–June). Multivariate logistic regression analysis was used to assess whether these associations would hold after controlling for other variables in the model. Analyses were conducted in Excel® and SAS v9.4. Table 2 presents each variable as coded in the analyses. Variables used to understand the associations between SP interventions and stillbirths between April and November 2015 (n = 9453) Nominal cost and expense data in 2015 Nigerian Naira (NGN) directly related to community and facility distribution of SP in the intervention and counterfactual LGAs were obtained from project records and other sources. The cost estimates obtained are what it would cost the state government and LGAs as de jure providers of primary health care in Nigeria, to deliver SP-related services at both the community and facility level, including start-up costs. Estimates were limited to a 12-month horizon. Different degrees of contributions by each service component at facility and community levels, towards the delivery of SP at facility and community levels, were assumed. Table 3 lists the assumptions about the magnitude of contributions by each level of care to SP distribution. For this purpose, six cost centers were included in the analysis: health facility, LGA technical administration, CBHV supervisors, WDC, CBHV, and logistics for SP distribution. Table 3 provides a summary of each cost center, and their relative contribution towards SP distribution the activities involved; these were costed. Twenty-two work days per month were assumed. Published government salary schedules were used to compute government officers’ salary costs. Governments officers’ salary costs included time spent at monthly LGA level review meetings in each LGA attended by representatives of wards. At the community level, costs were attributed to WDCs and to CBHVs. There was a WDC and one CBHV supervisor in each ward. WDCs supported CBHVs in the distribution of SP and supervised CBHV, as well as ad hoc community meetings that were called to tackle issues that could undermine the demand or supply sides of the programme. CBHV-related costs also included level of effort, transportation for SP distribution and monthly rentals of meeting rooms for ward-level CBHV review meetings also attended by LGA officials. Costs associated with the transportation and distribution of SP to 42 health facilities were captured in central storage costs. Cost centers by level of care and magnitude of their component costs associated with SP distribution Two ratios were calculated: cost per dose and cost per woman served, disaggregated by number of SP doses in the intervention and counterfactual group. Ratios were obtained from annualized costs derived in each LGA, intervention and counterfactual, the total as well as the disaggregated number of SP doses distributed, and from the total number of women served.