Abstract
Preventive chemotherapy with praziquantel for schistosomiasis morbidity control is commonly done by mass drug administration (MDA). MDA regimen is usually based on prevalence in a given area, and effectiveness is evaluated by decreases in prevalence and/or intensity of infection after several years of implementation. Multiple studies and programs now find that even within well-implemented, multiyear, annual MDA programs there often remain locations that do not decline in prevalence and/or intensity to expected levels. We term such locations "persistent hotspots." To study and address persistent hotspots, investigators and neglected tropical disease (NTD) program managers need to define them based on changes in prevalence and/or intensity. But how should the data be analyzed to define a persistent hotspot? We have analyzed a dataset from an operational research study in western Tanzania after three annual MDAs using four different approaches to define persistent hotspots. The four approaches are 1) absolute percent change in prevalence; 2) percent change in prevalence; 3) change in World Health Organization guideline categories; 4) change (absolute or percent) in both prevalence and intensity. We compare and contrast the outcomes of these analyses. Our intent is to show how the same dataset yields different numbers of persistent hotspots depending on the approach used to define them. We suggest that investigators and NTD program managers use the approach most suited for their study or program, but whichever approach is used, it should be clearly stated so that comparisons can be made within and between studies and programs.