ABSTRACT
The inSCALE cluster randomised controlled trial in Uganda evaluated two interventions, mHealth and Village Health Clubs (VHCs) which aimed to improve Community Health Worker (CHW) treatment for malaria, diarrhoea, and pneumonia within the national Integrated Community Case Management (iCCM) programme. The interventions were compared with standard care in a control arm. In a cluster randomised trial, 39 sub-counties in Midwest Uganda, covering 3167 CHWs, were randomly allocated to mHealth; VHC or usual care (control) arms. Household surveys captured parent-reported child illness, care seeking and treatment practices. Intention-to-treat analysis estimated the proportion of appropriately treated children with malaria, diarrhoea, and pneumonia according to WHO informed national guidelines. The trial was registered at ClinicalTrials.gov (NCT01972321). Between April-June 2014, 7679 households were surveyed; 2806 children were found with malaria, diarrhoea, or pneumonia symptoms in the last one month. Appropriate treatment was 11% higher in the mHealth compared to the control arm (risk ratio [RR] 1.11, 95% CI 1.02, 1.21; p = 0.018). The largest effect was on appropriate treatment for diarrhoea (RR 1.39; 95% CI 0.90, 2.15; p = 0.134). The VHC intervention increased appropriate treatment by 9% (RR 1.09; 95% CI 1.01, 1.18; p = 0.059), again with largest effect on treatment of diarrhoea (RR 1.56, 95% CI 1.04, 2.34, p = 0.030). CHWs provided the highest levels of appropriate treatment compared to other providers. However, improvements in appropriate treatment were observed at health facilities and pharmacies, with CHW appropriate treatment the same across the arms. The rate of CHW attrition in both intervention arms was less than half that of the control arm; adjusted risk difference mHealth arm -4.42% (95% CI -8.54, -0.29, p = 0.037) and VHC arm -4.75% (95% CI -8.74, -0.76, p = 0.021). Appropriate treatment by CHWs was encouragingly high across arms. The inSCALE mHealth and VHC interventions have the potential to reduce CHW attrition and improve the care quality for sick children, but not through improved CHW management as we had hypothesised. Trial Registration:ClinicalTrials.gov (NCT01972321).
ABSTRACT
BACKGROUND: Integrated community case management (iCCM) relies on community health workers (CHWs) managing children with malaria, pneumonia, diarrhoea, and referring children when management is not possible. This study sought to establish the cost per sick child referred to seek care from a higher-level health facility by a CHW and to estimate caregivers' willingness to pay (WTP) for referral. METHODS: Caregivers of 203 randomly selected children referred to higher-level health facilities by CHWs were interviewed in four Midwestern Uganda districts. Questionnaires and document reviews were used to capture direct, indirect and opportunity costs incurred by caregivers, CHWs and health facilities managing referred children. WTP for referral was assessed through the 'bidding game' approach followed by an open-ended question on maximum WTP. Descriptive analysis was conducted for factors associated with referral completion and WTP using logistic and linear regression methods, respectively. The cost per case referred to higher-level health facilities was computed from a societal perspective. RESULTS: Reasons for referral included having fever with a negative malaria test (46.8%), danger signs (29.6%) and drug shortage (37.4%). Among the referred, less than half completed referral (45.8%). Referral completion was 2.8 times higher among children with danger signs (p = 0.004) relative to those without danger signs, and 0.27 times lower among children who received pre-referral treatment (p < 0.001). The average cost per case referred was US$ 4.89 and US$7.35 per case completing referral. For each unit cost per case referred, caregiver out of pocket expenditure contributed 33.7%, caregivers' and CHWs' opportunity costs contributed 29.2% and 5.1% respectively and health facility costs contributed 39.6%. The mean (SD) out of pocket expenditure was US$1.65 (3.25). The mean WTP for referral was US$8.25 (14.70) and was positively associated with having received pre-referral treatment, completing referral and increasing caregiver education level. CONCLUSION: The mean WTP for referral was higher than the average out of pocket expenditure. This, along with suboptimal referral completion, points to barriers in access to higher-level facilities as the primary cause of low referral. Community mobilisation for uptake of referral is necessary if the policy of referring children to the nearest health facility is to be effective.
