Nov 13 - 17, 2016


Dianna B. Hergott1 , Christopher Schwabe1 , Guillermo Garcia2 , Wonder Philip Phiri2 , Megan Perry1 , Jose Luis Segura2 , Immo Kleinschmidt3 , Jackie Cook3

1 Medical Care Development International, Silver Spring, MD, United States, 2 Medical Care Development International, Malabo, Equatorial Guinea, 3 London School of Hygiene & Tropical Medicine, London, United Kingdom

The Bioko Island Malaria Control Project (BIMCP) uses a GIS-based Campaign Information Management System (CIMS) that uniquely identifies each household based on geographical location. 2014 spray data revealed Island wide coverage of just 57% with high refusal rates, well below the WHO recommended coverage for effectiveness. Due to this, as well as budget constraints that did not support long term sustainability of Island-wide IRS, the BIMCP stratified households and communities targeting spraying at the most vulnerable populations only, relying on LLINs in other areas. To stratify households by vulnerability, we utilized community-level parasite prevalence and risk of importation derived from the 2014 MIS, quality of housing and 2014 IRS coverage from the CIMS. As the 2014 MIS sampling frame was based on sentinel sites, MIS values for prevalence and importation in communities not included in the MIS sample were imputed. Based on this information, a risk score (R) and a risk-acceptability score (Ra) were derived for each community. Communities with the highest R and Ra values were selected for IRS- targeting, given the available budget, 30% of all households. Spray coverage in targeted communities was high, with an overall coverage of 80%. As such, in 2016, the BIMCP re-ran the stratification but removed previous IRS coverage as a criteria utilizing only prevalence, importation and housing quality - our measures of vulnerability. In addition, the 2015 MIS sampling frame was all communities in the Island with 20 households or more, eliminating the need to impute data. MIS data from 2014, 2015, and 2016 will be analyzed in communities that were stratified for IRS and in those that were not, to determine if the stratification of IRS and replacement of LLINs as the main control strategy in all communities resulted in a differential change in prevalence of infection across communities. An initial analysis of 2014 and 2015 data revealed prevalence of infection decreased both in communities stratified for IRS and those where LLINs alone were deployed, indicating that the withdrawal of IRS from less vulnerable communities did not lead to adverse consequences.