Abstract
BACKGROUND: Rapid diagnostic tests (RDTs) and microscopy are routinely used for the diagnosis of malaria in Ghana. DNA-based polymerase chain reaction (PCR) is not yet used routinely. We compared diagnostic methods and tested the sensitivities of different malaria diagnostic methods against PCR. MATERIALS AND METHODS: Study participants from four hospitals with a suspicion of malaria donated finger -prick blood for RDT and blood film examination. In addition, a blood spot was collected for PCR analysis, prior to treatment. Retrospective species-specific PCR was performed on all samples collected. RESULTS: Using PCR we found an overall malaria prevalence of 39% among the 211 evaluable blood spots (83/211) and this ranged between 6-61% across the four hospitals. Of the 164 participants with RDT data, malaria prevalence was 57% (94/164), ranging from 3-100% from the four hospitals. Microscopy was the least sensitive with a parasite prevalence of 21% (25/119) of the evaluable 119 participants, varying from 9 to 35% across three health facilities. By comparison, we found the sensitivities and specificities of RDT results when compared to PCR to be slightly higher than microscopy compared to PCR. These were 56.4% versus 41.7% and 90% versus 81.9%, respectively, but generally lower than expected. Ninety-five percent of the PCR-detected infections were P. falciparum, while 4% were mixed species infections of P. falciparum and P. malariae, with the remaining being a mono-infection of P. malariae. CONCLUSIONS: While using PCR as a gold standard, we found RDT to be more reliable in diagnosing malaria than microscopy. In addition, a majority of malaria-treated cases were not supported by PCR diagnosis, leading to possible overtreatment. Pragmatic strategies are needed to ensure suspected malaria cases are accurately diagnosed before treatment.