Re-analysis and interpretation of 2011 Liberia malaria indicator survey data to provide project-specific estimates for select malaria indicators

Malaria is one of many components addressed by the USAID-funded Rebuilding Basic Health Services (RBHS) project implemented in 5 out of Liberia's 15 counties. In late 2012, USAID drew RBHS's attention to intermittent preventive treatment coverage of pregnant women (IPTp) reported by the 2011 Liberia Malaria Indicator Survey (MIS) report. According to the report, only 50% of the women received two doses of IPT during their last pregnancy. RBHS was reporting project coverage of 80%. The discrepancy led to this reanalysis of 2011 MIS data collected in RBHS facility catchment areas to data collected in non-RBHS facility catchment areas to assess differences in IPT coverage. The analysis was expanded to include prevalence, diagnosis and prompt treatment of children with fever, and mosquito net usage by children under 5 years of age and by pregnant women. However, the 2011 Liberia MIS was administered using cluster sampling to generate robust estimates at the regional and national levels, not at the RBHS project level, resulting in limitations in generalizing findings. Nonetheless, the exercise proved to be a beneficial and cost-effective tool to help evaluate project outcomes. The process involved mapping the project target areas over those captured in the 2011 MIS using Arc-GIS, identifying the appropriate comparison groups, and performing a series of chi-square tests of independence using SAS. This presentation will describe the process for designing and conducting the analysis, and the benefits and limitations in using secondary survey data to provide project-level coverage estimates.