Background
Integrated community case management (iCCM) of malaria complements and extends the reach of public health services to improve access to timely diagnosis and treatment of malaria. Such community-based programmes rely on standardised test-and-treat algorithms implemented by community health workers using malaria rapid diagnostic tests (RDTs). However, due to a changing epidemiology of fever causes, positive RDT results might not correctly reflect malaria-disease in all malaria-endemic settings in Africa. This study modelled diagnostic predictive values for all malaria-endemic African regions as an indicator of the programmatic usefulness of RDTs in iCCM campaigns on malaria.
Methods
Positive predictive values (PPV) and negative predictive values (NPV) of RDTs for clinical malaria were modelled. Assay-specific performance characteristics stem from the Cochrane Library and publicly available data on the proportion of malaria-attributable fevers among African febrile children under five years of age were used as prevalence matrix.
Results
Average country-level PPVs vary considerably: Ethiopia had lowest PPVs (HRP2-assay: 17.35%; pLDH-assay: 39.73%) and Guinea the highest PPVs (HRP2-assay: 95.32%; pLDH-assay: 98.46%). On the contrary, NPVs were above 90% in all countries (HRP2-assay: ≥94.87%; pLDH-assay ≥93.36%).
Conclusions
PPVs differed considerably within Africa when used for screening of febrile children indicating unfavourable performance of RDT-based test-and-treat algorithms in low-PPV settings. This suggests that the administration of antimalarials alone may not constitute causal treatment in the presence of a positive RDT result for a substantial proportion of patients particularly in low-PPV settings. Therefore, current iCCM algorithms should be complemented by information on other setting-specific major causes of fever.