Randomized controlled trial of a rapid diagnostic test in home-based management of endemic malaria in Bugiri District, Uganda

Abstract

Introduction: Malaria is endemic in Uganda, and is the commonest cause of morbidity and mortality of children under-fives. The Government of Uganda is implementing Home Based Management of Fever (HBMF) Strategy. Objective: To evaluate the feasibility and acceptability of the use of Rapid Diagnostic Tests (RDT) in the Home-based Management of fever strategy for the diagnosis of malaria of the under-fives by Community drug distributors (CDDs). Methods: A three phase study was conducted in Bugiri District. Phase one was cross-sectional study targeting 451 heads of household. Phase two was a randomized intervention trial of CDDs in HBMF on the use of RDT involving 350 CDDs. Phase three was a qualitative study to assess the acceptability of RDT in HBMF strategy. Data analysis was carried out using bivariate and multivariate logistic and linear regressions, while qualitative data was coded. Findings: Of the 451 respondents interviewed 48% sought health care from drug shop attendants and 42.4% from nurses working in health units. Caregivers at home treated 39.4% of under-5s for malaria, compared to 65.7% in health units. Few cases (7.8%) were referred to higher health unit. A fifth (20.2%) took prompt action on day one, and half of the population (51.7%) consulted health workers. Fever was the most reported symptom of recent illness (92%). Most drugs (66.5%) were bought without prescription and only a few (6.7%) knew the correct dose. In HBMF access to treatment was very high (96%) but there was also over treatment as prevalence of malaria positivity ascertained by blood slide was only 56%. Treatment within 24 hours increased three fold (from 20% to 60%). Community drug distributors scored highly in use of RDT (72% sd = 33). Of all CDDs 47% returned all RDTs and scored 98%. CDDs’ socio-demographic or background characteristics did not influence performance, except sub-county where CDD lived, (OR = 1.520, 95% C.I. = 1.208 to 1.914, p<0.001). There was demonstrable impact of the use of RDT on HBMF. Before the intervention the referral rates were similar 3.7% in the control group and 3.8% in the intervention (t = - 0.297; p = 0.766). However after the intervention the referral rates were significantly different; 1.7% in the control group to 4.1% in the intervention compared (t = -5.191; p<0.001). This made the intervention group 2.4 times more likely to refer negative cases than the control group. RDT when used by CDDs in HBMF had high sensitivity (91.6%), with low specificity (21.0 %,) and moderate predictive values (positive pv = 59.6%, negative pv = 66.8%). The accuracy of CDDs using the RDTs for malaria diagnosis however was low (Kappa index agreement values = 0.17) though some communities did well (Kappa = 0.809) indicating possibility of improvement. In key informant interviews, the RDT was generally appreciated, and highly recommended before every treatment in HBMF. It was also demanded by both the CDDs and the communities, making HBMF strategy more acceptable. Conclusion: Majority of respondents sought health care from drug shop attendants. Most drugs were bought without prescription and only a fifth took prompt action on day one. Most community drug distributors were capable of using RDT, but this was influenced by age. The use of RDTs by CDDs in HBMF led to improvement in referral rates of negative cases. RDT had high sensitivity (91.6%), low specificity (21.0%), moderate predictive values (+pv = 59.9% and –pv = 66.8%), but low accuracy (Kappa = 0.140, p<0.001) which can be improved upon when used by CDDs in HBMF. Community acceptability of RDT was high. Recommendation: MoH can improve community based distribution of antimalarial drug for treatment of malaria of the under-five by formulating policy to integrate RDT for malaria diagnosis. District Health Team should ensure that criteria for selection of CDDs are adhered to and that CDDs are trained in the use of RDTs to improve malaria management at community level. There is need for cost benefit study to establish optimal use of the diagnostic test.