Malaria microscopy evaluation and quality assurance in rural clinics of Rarieda and Alego Usonga sub-counties of Siaya County, western Kenya

Background: Although rapid diagnostic tests (RDTs) are widely used for malaria diagnosis, they have notable limitations. Blood smear (BS) microscopy remains the gold standard, yet its reliability in public health facilities (HFs) across malaria-endemic regions of Kenya can be compromised by limited infrastructure, technical capacity, and quality assurance. 

Methods: We assessed the quality of malaria microscopy in 29 HFs in Siaya County, western Kenya, from January–July 2024 by evaluating the concordance of routine HF BS results with expert microscopy. We evaluated the availability and quality of reagents, standard operating procedures, and infrastructure using the National Malaria Control Program (NMCP) technical supervision checklist which follows WHO-certified microscopy standards. Up to 60 participant slides were randomly selected over three visits and two slides were prepared for each participant. Slide 1 was prepared and read on-site as per routine HF practice, then re-examined by an expert microscopist using the HF microscope and again at the Kenya Medical Research Institute (KEMRI). Slide 2 was stained and read at KEMRI, serving as the gold standard. We evaluated factors and characteristics associated with accurate diagnosis using logistic regression. 

Results: Of the 1,494 blood smears examined, 501 (34%) were positive. Concordance between routine microscopy and expert re-reading was 91% (1,289/1,414), ranging from 55% (6/11) to 100% (60/60) across health facilities. Percent agreement between HF slide 1 and slide 2 was 86% (1,276/1,485), with a range from 55% (6/11) to 98% (39/40) by HF. Compared to slide 2, sensitivity and specificity of HF results was 76% and 92%, respectively, resulting in undertreatment of 24% and overtreatment of 8% of patients. Those with parasitemia 1–100 p/μL had lower odds of accurate diagnosis (OR = 0.12; 95% CI: 0.05–0.29; p < 0.001) while specimens with parasite densities > 10,000 p/μL had higher odds of accurate diagnosis (OR = 6.08; 95% CI: 2.22–25.1; p = 0.002), indicating a positive association between parasite density and diagnostic accuracy. Six (21%) of HFs had poor quality BS with debris or contamination. 

Conclusions: While overall concordance was high, the variability in results by HF, limited accuracy at low parasite densities, and challenges with required infrastructure highlight the need for ongoing malaria microscopy quality assurance to ensure proper case management.