A New Approach to Deliver Malaria Preventions to Pregnant Women at a Community Level in Uganda

SUMMARY Background: Malaria is the leading cause of morbidity and mortality among pregnant women in Uganda. Although effective tools for prevention and control of malaria exist, their delivery presents a problem. Intermittent presumptive treatment (IPT) with sulfadoxine-pyrimethamine (SP) is effective, yet >60% of women in Uganda do not get it as < 40%, attend antenatal care. Effective ways of delivering IPT with SP to pregnant women at a community level need to be developed. This study assessed whether community based resource persons like traditional birth attendants (TBAs), community reproductive health workers (CRHWs), adolescent peer mobilizers (APMs) and drug-shop owners (DSV) can distribute IPT with SP to pregnant women. Objectives: The objectives of this study were: To assess community based approaches for delivering malaria prevention to pregnant women in Uganda; To assess community perceptions, beliefs and practices associated with malaria treatment and prevention in pregnancy; To assess whether community based resource persons can deliver IPT to pregnant women and reach those most at risk; To assess the impact of IPT on anaemia and pregnancy outcome; To estimate cost-effectiveness of the approaches and assess the acceptability and sustainability of the approaches. Methods: The study was conducted in 5 sub-counties of the Mukono district, situated on the shores of L. Victoria in Central Uganda. The district is hyper-endemic for malaria. 25 parishes with a total population of 75,000 people were used to test the new approaches. Phase 1 obtained qualitative data on community perceptions, beliefs and practices associated with malaria prevention in pregnancy. Phase 2 was an intervention study that assessed distribution of IPT to pregnant women by TBAs, CRHWs, APMs and DSVs compared with health units. Pregnant women of all parities were enrolled. Key resource persons in each parish were identified to sensitise the communities on the intervention. Data was collected regarding: timing of the first dose of SP, proportion of women who complete two doses of SP, birth weight of babies, proportion of low birth-weight babies, and proportion of adolescent pregnancies. The third phase of the study evaluated the sustainability of the approaches. Work Plan: The first phase of the study took two months. The second phase took 14-16 months. Data analysis was expected to take 12 months.

OBJECTIVES OF THE STUDY: The objectives of this study were: To assess new community based approaches for delivering malaria prevention to pregnant women in Uganda; To assess community perceptions, beliefs and practices associated with malaria treatment and prevention in pregnancy; To assess whether TBAs, CRHWs, APMs and DSVs can administer IPT with SP to pregnant women at a community level; To assess whether TBAs, CRHWs, APMs and DSVs can reach women at most risk of malaria in pregnancy; To evaluate the impact of community based distribution of IPT with SP on anaemia in pregnancy and birth outcome; To assess the acceptability and sustainability of delivering IPT with SP to pregnant women at a community level; To estimate the cost-effectiveness of the new community based approaches for delivering malaria prevention interventions to pregnant women. The study was conducted in the Mukono district, Central Region, Uganda. The district is situated 50-150kms east of Kampala. The district has an estimated population size of 1,128,500. The district lies at altitudes from 1,158m to 1,219m and is bordered by Kayunga district in the North and Lake Victoria in the South. It has high temperatures and heavy rainfall during the months of March to May and October to November. The district is divided into 6 counties, 28 sub-counties and 145 parishes. A sub-county usually has a population of about 30,000 people while a parish has approximately 3,000 people. The Baganda are the majority tribe in the district. Other tribes include: Basoga, Badama, Banyole, Banyarwanda, Barundi, Balulu, Lugbar and Jaluos. There is a high fertility rate, of 7.2 births per woman and an annual population growth rate of 2.3%. The district is hyper- to holoendemic for malaria. The population is served by 4 hospitals, 5 health centres IV, 23 health centres III and 25 health centres II. There are numerous drug-shops, TBAs, private midwives and CHRWs who provide health care in the district. Drug shops are licensed and registered at the District Health Office. Five rural sub-counties from the lake- shore region with the same level of malaria transmission were selected. Within each sub-county, 5 parishes were selected based on: availability of a health centre II, III or IV, and trained TBAs, DSVs, CRHWs and APMs. Study Population: The study population included all the pregnant women who live in the study area. PHASE ONE: Meetings and discussions were held with civic leaders, women leaders, and community resource persons in the study areas to present the study and solicit consent from the community and opinion leaders. A formative study was carried out to explore the communities' ideas about approaches to deliver malaria prevention and to assess perceptions, beliefs and practices associated with malaria prevention in pregnancy. Focus group discussions (FGDs) and in-depth interviews targeting pregnant women, opinion leaders, and men and women in different age groups were conducted in each sub-county. FGDs had 6-7 participants within the same age range and sex. In-depth interviews were held with DSVs, TBAs, opinion leaders, traditional healers, CRHWs, APMs and health workers in each sub-county. A survey was conducted in the 25 parishes to get a register of the number of women in the reproductive age group and the number of women currently pregnant. The total population in the 25 parishes is approximately 75,000 people of which 43% are in the reproductive age group. Pregnant women were given cards with personal identification numbers (PINs) and a GPS map was constructed for easy tracking. PHASE TWO: An intervention study testing community based approaches for delivering IPT with SP to pregnant women was conducted. The study had four delivery approaches: TBAs, DSVs, CRHWs and APMs were trained to give a two dose regimen of SP to pregnant women. These approaches were compared with routine services at health centres II to IV. The intervention was coupled with community mobilisation and sensitisation to ensure that all women were informed on the intervention. Information materials and posters were distributed in the study areas to indicate the benefits of malaria prevention and where to get SP. All pregnant women who visited DSVs, TBAs, CRHWs, APMs and health centres and consented to participate in the study were recruited and given IPT with SP once during the second and third trimester. Baseline data were obtained on Hb levels, parasitaemia, history of previous illnesses, obstetric history and socio-demographic data. Appointments were made for the pregnant women to have a blood sample taken for investigations before receiving the first dose of SP. TBAs, CRHWs, APMs and DSVs were trained to keep appropriate calendars of scheduled visits for the second dose of SP. Pregnant women received health education, iron and folic acid supplementation, de-worming and information on proper nutrition. All pregnant women used their card with PIN number for identification wherever they went to get care. The women were encouraged to deliver at health units. TBAs, CRHWs, APMs and DSVs were supplied with baby weighing scales for babies who are delivered at home (within 5 days of delivery). Pregnant women who had a history of allergy to sulphonamide containing drugs were excluded. Women who consented to participate were told about the necessity of taking a blood sample to check Hb levels and malaria parasites. Local laboratory assistants were hired, trained and supervised for quality control. Women were given appointments to visit the laboratory assistants or go to a laboratory in the study area. Two blood samples were taken, one at recruitment and another at 36 weeks of gestation. Outcome variables: Timing of the first dose of SP. Proportion of women who complete two doses of SP. Birth weight of babies. Proportion of low birth-weight babies. Maternal Hb and malaria parasite counts. Proportion of adolescent pregnancies enrolled. Proportion of pregnant women who agree to take part in the study. Proportion of scheduled visits completed by each delivery approach. The costs per low birth weight of babies avoided. The costs per anaemic pregnant woman avoided. Mothers' perceptions on the four different SP delivery outlets. The sample size calculations were based on Pockock, 1983 and three outcome variables were used: Hb level of 11.0 gm/dl, standard deviation (SD) 1.5 gm/dl; mean birth weight of 3000gms, SD 500gms; and proportion of women who take malaria prophylaxis in Uganda, currently estimated at 34%. To detect differences in Hb of 0.5 gm/dl, birth weight of 150gms and a change of 12% in the proportion of women using malaria prophylaxis; between the control and any delivery approach of the study, the sample sizes required are 170, 211, and 310 for the three variables respectively, allowing for a 20% loss of follow up. For this study, the sample size of 310 was taken for each arm giving a total sample size of 1,550. Cost component of the intervention study: For all of the approaches an attempt was made to collect information from a societal point of view so that all costs could be included irrespective of who pays for them. The approaches incurred costs falling into three main categories: patient related costs, supplier costs and health facility based costs. Patient related costs due to malaria illness in pregnancy was collected as part of the overall intervention study. A questionnaire was developed to capture costs borne by the pregnant women and the families as a result of malaria related morbidity (e.g. number of visits to health centres and hospitals to seek ANC as well as treatment for malaria illnesses, transport costs, transport time, number of SP pills consumed, time lost by the pregnant woman due to malaria illness and time lost by family members to provide care. The questionnaire was administered to all women immediately after birth. Supplier costs referred to the costs of enabling TBAs, DSVs, CHRWs and APMs to distribute SP (e.g. training costs of suppliers as well as delivery costs). Interviews with TBAs, CRHWs and APMs were conducted to assess time spent on average per week distributing SP. The total number of SP tablets distributed was recorded. Regarding health facility based costs, the patient questionnaire identified the average number of health facility visits per pregnant woman for ANC and malaria related morbidity. Two hospitals and six health centres were randomly selected from the study areas and visited to collect the relevant costing information. Bottom-up costing was used to estimate the costs of patient specific inputs such as drugs prescribed and laboratory tests taken for individual patients. This was done in the hospital through a review of a sample of malaria patient notes and in health centres through examination of a sample of OPD cards for malaria patients. All remaining costs were determined utilizing standard step-down costing. As the main measures of effect, the number of low birth weight babies per pregnancy avoided and the number of anaemic women per pregnancy avoided was used. PHASE THREE: Following the intervention study, data was collected from a sample of women who participated in the study and the community resource persons to get their views on sustainability of the delivery system. This was a qualitative study using FGDs and in-depth interviews. DATA GENERATED BY THE STUDY: The formative study generated qualitative data on perceptions, beliefs and practices associated with malaria prevention in pregnancy; data on costs of the interventions; and data on acceptability and sustainability of the community based approaches. DATA ANALYSIS: After completing and validating field forms, data was double data entered and validated using EPINFO version 6.0. Analysis was done using SPSS PC for Windows Version 10. Descriptive statistics was computed. Analysis of variance (ANOVA) will be used to compare differences in means. Logistic regression analysis will be used to identify factors associated with access, utilisation and impact of malaria prevention interventions. Qualitative data will be analysed thematically to explain some of the findings and help in interpretation of quantitative results. DISSEMINATION OF RESULTS: The results will be disseminated at the national level to policy makers, program managers and the scientific community in Uganda. Dissemination will take place within the study district and the participating communities. Data will be published in peer-reviewed journals. ETHICAL CONSIDERATIONS: The research proposal was be submitted to the Ethics Committee in Denmark and the Uganda National Council of Science and Technology for approval. Informed consent (verbal) was obtained from all pregnant women participating in the intervention. Consent was obtained from community leaders, women and opinion leaders. Sterile needles and syringes were used to draw blood. The women were informed about laboratory examination results. Women received explanations on the following issues: advantages and disadvantages of the intervention; and possible adverse effects with use of SP in pregnancy. An explanation was given to the women that they were free to withdraw from the epidemiological study any time and they had access to SP if they wished. All malaria cases and other diseases that arose during pregnancy were treated and referral of complicated cases made.

Inclusion Criteria: Pregnant women of all parities and gestational week 16 Exclusion Criteria: Pregnant women with a history of allergy to sulphonamide containing drugs