Table 2.

Characteristics of published studies of home- and community-based treatment for malaria in Africa^a

Location	Epidemiology	Drug distribution	Incentive	Outcomes measured	Results
Kenya 1981–83 (Spencer et al, 1987a, 1987b)	• Rural • Hyper- to holo-endemic	• CHWs provided presumptive CQ treatment for free	• Volunteer CHWs supported by the village	• Overall and malaria-specific mortality • Birth and fertility rates • Parasite rates	No obvious effect of providing CQ for treatment of malaria on mortality, fertility, or parasite rates
The Gambia 1982–87 (Greenwood et al, 1988; Menon et al, 1990)	• Rural • Seasonal transmission	• CHWs sold CQ for presumptive treatment	• Volunteer CHWs supported by the village	• Overall and malaria-related mortality • Frequency of clinical malaria • Packed cell volume, parasite rates, splenomegaly	Treatment alone had no significant effect on morbidity and mortality from malaria
Zaire (DRC) 1985–87 (Delacollette et al, 1996)	• Rural • Meso-endemic • Continuous transmission with seasonal fluctuations	• CHWs sold CQ at cost for presumptive treatment	• CHWs received "symbolic monetary reward"	• Malaria morbidity and mortality • Parasitological indices • Proportion of fever episodes receiving antimalarial treatment, proportion receiving treatment at home, and source of treatment	No impact on malaria mortality, but two-fold reduction in malaria prevalence and incidence
Burkina Faso 1994–95 (Pagnoni et al, 1997)	• Rural • Seasonal transmission	• Mothers trained to recognize illness and make decision to treat • CHWs sold pre-packaged CQ for presumptive treatment	• CHWs kept 0.6 US cents for each package sold	• Proportion of under-5 malaria cases recorded as severe in health centres • Mothers' care-seeking practices • Availability and use of drugs at peripheral level, community awareness of educational messages	The proportion of severe cases decreased in the first year of the program; in the second year, the proportion decreased only in health facilities with drug coverage ≥50%
Ethiopia 1996–98 (Kidane and Morrow, 2000)	• Rural • Seasonal transmission	• Mother coordinators provided presumptive CQ treatment for free	• None mentioned	• Malaria-related mortality in children under age 5 years	Intervention associated with 40.6% reduction in overall under-5 mortality (95% CI 29.2–50.6, p < 0.003)
Burkina Faso 1998–99 (Sirima et al, 2003)	• Rural • Hyperendemic • Seasonal transmission	• Mothers trained to recognize illness and make decision to treat • CHWs sold pre-packaged CQ for presumptive treatment	• Drugs sold with 10% incentive margin for CHW • Incentive provided to some drug store managers	• Proportion of malaria cases progressing to severe (as reported by mothers in annual cross-sectional surveys) • Proportion of cases receiving correct dose of CQ	Risk of progression to severe malaria lower in children treated promptly with pre-packaged CQ (5%) than not (11%) (RR 0.47, 95% CI 0.37–0.60, p < 0.0001)

^aCHW = community health worker; CQ = chloroquine; RR = risk ratio