Geospatial Analyses of Childhood Malaria Following Repeated Village-Wide Ivermectin Administrations: Secondary Analyses for the RIMDAMAL Pilot Study

Malaria has long been a major public health concern, with historic roots dating back thousands of years. This febrile disease is caused by a parasite that is transmitted among vertebrates by mosquitoes. Over the past century, global eradication programs have focused on minimizing populations of the insect vectors, and administering treatments to people infected, especially young children and pregnant women, as they are the most vulnerable to suffering severe complications. Overall, these programs have decreased the geographic distribution and global disease burden; however, malaria remains a major problem in regions where these efforts have been unsuccessful. In 2015, there were an estimated 214 million cases throughout the world, resulting in approximately 438,000 deaths; however, over 3 billion people are living at risk of becoming infected with malaria. Widespread use of the few available effective insecticides and anti-malarial drugs has conferred resistance in both parasitic and mosquito species, decreasing the effectiveness of current interventions. As anti-malarial resistance and insecticide resistance spread, the need for novel malaria interventions becomes more urgent. One novel approach to combatting malaria was pilot-tested by researchers in the Department of Microbiology, Immunology and Pathology at Colorado State University. The Repeated Ivermectin Mass Drug Administration to control Malaria, or the RIMDAMAL study, evaluated the safety and effectiveness of repeated village-wide administrations of an anti-parasitic drug to prevent malaria in children ≤ 5 years old. The RIMDAMAL study was a randomized trial carried out in Burkina Faso, a small tropical country in West Africa. Ivermectin (IVM) is a common anti-parasitic used around the world to prevent and treat parasitic diseases. Recent evidence has demonstrated that IVM is toxic to malaria-transmitting mosquitoes, and can inhibit the propagation of some life stages of malaria parasites. Initial analyses of the RIMDAMAL data found significantly fewer childhood malaria cases in intervention villages that received repeated IVM administrations, compared to control villages. This study is a geospatial analysis of the RIMDAMAL data to provide further insight as to how this intervention could be implemented. There were two study aims for this research: 1) identify significant clustering of high and low childhood malaria incidence within each study village; and 2) identify significant clustering of high and low childhood malaria incidence throughout the entire study region. In total, eight villages were enrolled in the study, four of which served as controls, while the other four received the intervention. Residents of each village live in concessions, or compounds of extended family. Geospatial coordinates were collected for each concession within a study village, along with data on the participants within each concession. Using this data, incidence density of malaria among children 5 years old or younger was calculated at the concession level. Concessions were mapped, and spatial clustering of incidence density values was evaluated using the Getis-Ord Gi* (G-I-star) spatial autocorrelation statistic. To evaluate within village clustering, each of the eight study villages were analyzed individually, and between village clustering was evaluated by analyzing the entire study region. Within each village, several “hot spots,” or statistically significant clusters of high malaria incidence density values were recognized during analyses with max clustering, at the 95% confidence level. Statistically significant clusters of low incidence density were identified in one study village during the analysis with max clustering. The proportion of concessions identified as significant clusters varied by village, ranging from 12% to 91.3%. There seems to be no trend in clustering patterns seen within each village; some villages had randomly distributed hot or cold spots, while others appeared more clustered. The spatial clustering patterns in the whole study region are more telling. Max clustering occurs in a bimodal pattern with two peaks; at 2,100 meters and 10,000 meters. The clustering patterns that occur indicate regions of similar malaria incidence. The proximity and locations of these villages may imply the RIMDAMAL protocol has regional impacts. Additional research is needed to evaluate how to most effectively implement this intervention to protect against malaria