SMS photograph-based external quality assessment of reading and interpretation of malaria rapid diagnostic tests in the Democratic Republic of the Congo

Background: The present External Quality Assessment (EQA) assessed reading and interpretation of malaria rapid diagnostic tests (RDTs) in the Democratic Republic of the Congo (DRC). Methods: The EQA consisted of (i) 10 high-resolution printed photographs displaying cassettes with real-life results and multiple choice questions (MCQ) addressing individual health workers (HW), and (ii) a questionnaire on RDT use addressing the laboratory of health facilities (HF). Answers were transmitted through short message services (SMS). Results: The EQA comprised 2344 HW and 1028 HF covering 10/11 provinces in DRC. Overall, median HW score (sum of correct answers on 10 MCQ photographs for each HW) was 9.0 (interquartile range 7.5 - 10); MCQ scores (the % of correct answers for a particular photograph) ranged from 54.8% to 91.6%. Most common errors were (i) reading or interpreting faint or weak line intensities as negative (3.3%, 7.2%, 24.3% and 29.1% for 4 MCQ photographs), (ii) failure to distinguish the correct Plasmodium species (3.4% to 7.0%), (iii) missing invalid test results (8.4% and 23.6%) and (iv) missing negative test results (10.0% and 12.4%). HW who were trained less than 12 months ago had best MCQ scores for 7/10 photographs as well as a significantly higher proportion of 10/10 scores, but absolute differences in MCQ scores were small. HW who had participated in a previous EQA performed significantly better for 4/10 photographs compared to those who had not. Except for two photographs, MCQ scores were comparable for all levels of the HF hierarchy and non-laboratory staff (HW from health posts) had similar performance as to laboratory staff. Main findings of the questionnaire were (i) use of other RDT products than recommended by the national malaria control programme (nearly 20% of participating HF), (ii) lack of training for a third (33.6%) of HF, (iii) high proportions (two-thirds, 66.5%) of HF reporting stock-outs. Conclusions: The present EQA revealed common errors in RDT reading and interpretation by HW in DRC. Performances of non-laboratory and laboratory staff were similar and dedicated training was shown to improve HW competence although to a moderate extent. Problems in supply, distribution and training of RDTs were detected.