Diagnostic accuracy of routine malaria tests and antimalarial prescription practices at health facilities in a endemic setting in Uganda

Accurate malaria test results are a prerequisite for effective malaria case management. The accuracy of routine Rapid Diagnostic Test [RDTs] and microscopy at lower level health centers (HC) in Uganda was determined. Patient exit interviews were conducted at all HC IV and IIIs and 20% of HC IIs in 11 districts of Busoga sub-region. A total of 210 febrile outpatients of all ages were enrolled from facilities in each district. Data on routine malaria test results (RDT or microscopy) and prescribed antimalarials were obtained from patient medical records. Additionally, a blood sample was collected from each patient for repeat RDT testing and thick blood smear examination by, experts, and which served as the reference standard for comparison with corresponding routine test results. Agreement was measured using percentage agreement and Cohens Kappa (κ). Standard measures of diagnostic accuracy for binary diagnostic tests were calculated. Clustering at the facility level was adjusted for. Sampling weights were applied to account for the survey design. A total of 3981 patients were included in the analysis. Overall, for routine RDTs, κ was 0.77 (95% CI 0.72, 0.81) and was similar across facility and age categories. κ was lowest at Private for Profit (PFP; 0.48, 95% CI 0.09, 0.87) facilities. Sensitivity and specificity was 92.9% (95 CI 89.8%, 95.0%) and 83.6% (95% CI 78.4%, 87.8%) and remained the same across facility and age categories. Sensitivity and specificity estimates were lowest at PFP facilities; 76.0% (95 CI: 41.4, 93.4) and 75.0% (95% CI 39.7%, 93.2%), respectively. The overall area under the curve (AUC) was high 0.87 (95% CI 0.81, 0.92) but lowest at PFP facilities 0.76, (95% CI 0.67, 0.86). Routine microscopy, demonstrated low agreement (κ 0.11, 95% CI 0.01, 0.20), sensitivity (69.7%, 95% CI 52.3, 82.8%). and specificity (52.7%, 95% CI 44.0%, 61.2%) and remained low across facility types and age groups. Agreement levels and diagnostic accuracy of routine RDTs was relatively good and that of routine microscopy very poor. Performance of both tests was lowest at PFPs. There is an urgent need to improve performance of malaria tests, especially for microscopy and for both tests at PFP facilities.