Menu
Menu
Menu
Menu
Background: Despite massive anti-malaria campaigns across the subcontinent, effective access to intermittent preventive treatment (IPTp) and insecticide-treated nets (ITNs) among pregnant women remain low in large parts of sub-Saharan Africa. The slow uptake of malaria prevention products appears to reflect lack of knowledge and resistance to behavioural change, as well as poor access to resources, and limited support of programmes by local communities and authorities. Methods. A recent community-based programme in Akwa Ibom State, Nigeria, is analysed to determine the degree to which community-directed interventions can improve access to malaria prevention in pregnancy. Six local government areas in Southern Nigeria were selected for a malaria in pregnancy prevention intervention. Three of these local government areas were selected for a complementary community-directed intervention (CDI) programme. Under the CDI programme, volunteer community-directed distributors (CDDs) were appointed by each village and kindred in the treatment areas and trained to deliver ITNs and IPTp drugs as well as basic counseling services to pregnant women. Findings. Relative to women in the control area, an additional 7.4 percent of women slept under a net during pregnancy in the treatment areas (95% CI [0.035, 0.115], p-value < 0.01), and an additional 8.5 percent of women slept under an ITN after delivery and prior to the interview (95% CI [0.045, 0.122], p-value < 0.001). The effects of the CDI programme were largest for IPTp adherence, increasing the fraction of pregnant women taking at least two SP doses during pregnancy by 35.3 percentage points [95% CI: 0.280, 0.425], p-value < 0.001) relative to the control group. No effects on antenatal care attendance were found. Conclusion: The presented results suggest that the inclusion of community-based programmes can substantially increase effective access to malaria prevention, and also increase access to formal health care access in general, and antenatal care attendance in particular in combination with supply side interventions. Given the relatively modest financial commitments they require, community-directed programmes appear to be a cost-effective way to improve malaria prevention; the participatory approach underlying CDI programmes also promises to strengthen ties between the formal health sector and local communities. © 2011 Okeibunor et al; licensee BioMed Central Ltd.
The study used a pre-post parallel group design, with group assignment implemented at the local government area level (LGA). Each LGA comprises an average population of approximately 130,000 individuals and an estimated average number of 6500 pregnancies per year. Three LGAs (Eket, Esit Ekit, Onna) were assigned to the treatment group, and three LGAs (Ikot Abasi, Mbo, Mkpat Enin) were assigned to the control group. The assignment of areas to treatment and control was not random, but rather determined prior to the beginning of the study with the objective to guarantee balanced samples with respect to health facility infrastructure across the two study arms. Independent random samples of women with pregnancies over the six-month period preceding the survey were interviewed pre- and post-intervention to measure changes in effective access to malaria prevention in pregnancy. Target population of the programme was all pregnant women residing in the six programme LGAs. As Figure Figure11 illustrates, 1,280 women aged 15-49 with recent pregnancies in the study areas were randomly selected for an interview at baseline. A second, independent sample of 1,380 women was randomly selected for a follow-up interview in February 2010. Target populations of both surveys were women who had given birth within six months prior to the survey. Sampling and time line. The study was located in the Eket Senatorial Zone in Southern Nigeria. Eket Senatorial Zone is one of three senatorial zones within Akwa Ibom State, and the principal area of operation of ExxonMobil in the region. Akwa Ibom State covers a landmass of approximately 8,000 square kilometers and is currently home to an estimated population of 3.9 million [26]. The discovery and extraction of crude oil in the area has led to massive in-migration over the last decades, resulting in a rapidly growing ethnically diverse population. Local climate is tropical, with a dry season between November and March and a wet season between April and October. The average temperature ranges from 23°C – 31°C, providing an ideal climate for malaria transmission throughout the year, and placing Akwa Ibom State among the areas with the highest malaria transmission in the whole region [27]. Even though only 2.5% of Nigeria's population lives in the state, Akwa Ibom accounts for over 11% of malaria-linked maternal mortality and 12-30% of under-5 mortality due to malaria in the country [28]. As illustrated in Figures Figures22 and and3,3, the target area of the study covers six LGAs within the Eket Senatorial zone along the coastal belt of Akwa Ibom State: Eket, Esit Eket, Ikot Abasi, Mbo, Mkpat Enin and Onna. Each of the six local government areas is served by at least one state government hospital, and all LGAs have at least four additional health clinics or health posts. Overall, there are 20 Local Government Primary Health Care facilities in the treatment area, and 19 health facilities in the control area. Nigeria and Akwa Ibom State. Target areas within Akwa Ibom State. In order to minimize equity concerns, increased resource and training support was provided to both treatment and control areas. All public health clinics in the programme area were provided with drugs, functional equipment and other supplies such as tracking sheets to enhance their functionality. Health workers received additional training focusing on the delivery of ANC services at public health clinics in general, and on improving malaria in pregnancy (MIP) services in particular. MIP performance standards were developed to improve service delivery and workers from health facilities in the control and intervention arms were trained to achieve these standards. In addition to these common interventions, a CDI programme was implemented in the three treatment LGAs. The main objective of the CDI was to select and train local delivery agents (CDDs) to support malaria prevention efforts. To ensure local support, the selection of CDDs was delegated to each kindred within a given community. Similar to the kindred definition described by Katabarwa et al for the Ugandan context [29], "kindreds" are extended family units or clans with a common ancestry, commonly referred to as "ekpuk" by the local Ibibios, and as "Umunna" in neighbouring areas of Nigeria [30,31]. Typically, Ibibio villages held about five hundred people, and were divided into physically distinct divisions dominated by separate patrilineages. Priority in the selection of CDDs was given to women from each local kindred with prior childbearing experience in order to minimize communication barriers between pregnant women and the CDDs. Overall, 700 volunteer community-directed CDDs were trained from more than 450 kindred groups. With an estimated 86,000 women between 15 and 49, this implies that approximately one CDD was trained for every 120 women of childbearing age in the treatment areas. With a estimated general fertility rate of 194 per 1000 women of childbearing age [12], this implies that each CDD covered on average 23 births per calendar year. CDDs were trained to deliver ITNs and SP (IPTp1 and IPTp2) to pregnant women in the treatment arm communities, and also to provide basic health counseling services. To make sure the CDI programme would not undermine usage of the existing health system, CDDs were instructed to refer pregnant women to health facilities for additional ANC services. Conditional on successful completion of the training, CDDs were equipped with counseling cards, IPTp drugs and ITNs, as well as village register, referral forms and monthly tally sheets to report service statistics. All training and supervision of the CDDs was conducted by staff of the nearest health facility within the treatment area to ensure full collaboration between health facility employees and CDDs. Baseline surveys were conducted from October 1st to October 15th 2007. The rollout of the interventions started in July 2008; interventions continued throughout 2010. The follow-up survey interviews were conducted between February 11th and February 26th 2010. The primary objective of the intervention was to increase effective access to malaria prevention among pregnant women; specific targets were increasing ITN use, and increasing access to IPTp in the form of two doses of SP during pregnancy. The effectiveness of the intervention is evaluated through five health access indicators: the probability of a woman (1) reporting to access ANC services during her pregnancy at least once; (2) reporting to have slept under a ITN during pregnancy; (3) reporting to have slept under a ITN the night before the interview; (4) reporting having taken any malaria prevention drugs; (5) reporting having taken at last two doses of SP (each dose of SP consists of three tablets). A randomly-selected sample of 1,280 women were interviewed at baseline. Data analysis was restricted to women with complete information, which results in a final baseline sample of 1,274 respondents. An independent random sample of 1,380 women were interviewed at endline. Two women had missing information on at least one key variable, resulting in an endline sample of 1378 observations. The sample size for the household surveys was calculated in order to be able to detect a 10 percentage point increase in utilization of government ANC services from a baseline of 25% of pregnant women reporting at least one ANC visit at a government health facility. Primary endpoints of the study were the fraction of pregnant women sleeping under ITNs, as well as the fraction of women taking the recommended two doses of SP. Secondary outcomes included ANC visits, ITN use post-pregnancy, as well as other malaria prevention efforts. Unconditional pre-post group mean differences are presented as a first step, before showing multivariate regression results. In order to facilitate coefficient interpretation, heterogeneity-adjusted linear probability models of the following functional form are estimated: y is the outcome of interest for woman i in local government area j and period t, and POST is a binary indicator, which equals 0 for baseline (PRE) observations and 1 for end line (POST) observations. The POST term captures the average improvements in both groups, while the POST * CDI interaction term captures differences in changes across the two groups, and thus represents the treatment effects of interest. X is a vector of control variables, which includes respondent's age, education, marital status, ethnicity, religion, occupation and household wealth. In order to capture time-invariant characteristics specific to the six local government areas, LGA fixed effects δj are included in the empirical models. Given that all dependent variables are binary, Huber-White standard errors are applied to adjust for the non-normal distribution of the error terms. To ensure correct causal inference in the presence of spatial correlation within the treatment areas, all standard errors are clustered at the LGA level. In order to ascertain that the main results are not driven by differences in access at baseline, a series of small-sample estimates representing pair-wise comparisons of the two most similar LGAs for each outcome of interest are also reported. All empirical analysis was conducted using the Stata© 10 statistical software package. The study was registered with the Federal Ministry of Health in Abuja. Prior to the rollout of the programme, a consensus-building meeting was held with senior members of the State Ministry of Health to agree on the services CDDs would provide. In addition, health workers engaged with community stakeholders in local meetings to seek their approval, sensitize them towards the importance of their role in promoting maternal health, encourage shared learning and create a supportive environment for the programme. Ethical approval was obtained through the Committee on Human Research at Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, USA; and the National Malaria Control Programme (NMCP), Abuja approved the study protocol for implementation. A formal MOU was set up between Jhpiego and local as well as State authorities. Informed and written consent was obtained from all persons who voluntarily agreed to be interviewed.
N/A
