Are malaria rapid diagnostic test results stable over time to support verification of surveillance data?

Background Rapid diagnostic tests (RDTs) have improved malaria case management by enabling point-of-care confirmation of infection, particularly in low-resource settings. In addition to clinical use, RDT results recorded in health facility registers are a critical component of national malaria surveillance systems. Recently, national programs have explored using stored RDT cassettes to validate register data. However, manufacturers caution that results should be read within 15 - 30 minutes, raising concerns about result validity after this period. This study evaluated the stability of RDT results over a one-month period to assess whether stored cassettes can reliably reflect initial test outcomes. Methods We conducted a prospective, observational study in 48 health facilities across Benin, Nigeria, and Uganda from June to September 2023. A digital artificial intelligence (AI)-powered RDT reader (HealthPulse, Audere, Seattle WA USA) was used to photograph RDTs immediately after interpretation by health workers and again at one week and one month. Images were independently interpreted by a trained panel, with results categorized as positive, negative, invalid, or indeterminate. Only RDTs with valid interpretations at all three time points were included in the final analysis. Positive and negative predictive values (PPV and NPV) were calculated to measure the accuracy of results from stored RDTs relative to the initial interpretation. Results Out of 54,251 RDTs captured, 45,155 (83.2%) met inclusion criteria. At one month, 95.1% of initially positive RDTs remained positive, and 95.3% of initially negative RDTs remained negative. The PPV of a positive result at one month was 96.3% (95% CI 96.1, 96.5), while the NPV of a negative result was 93.8% (95% CI 93.4, 94.1). Most result changes occurred within the first week. Faint lines were associated with higher rates of change in both directions; 26.8% changing from positive to negative and 48.1% changing from negative to positive. Stability of results also varied across RDT products and specific test lines. Conclusions Stored RDT cassettes maintain high result stability over one month and can serve as a reliable reference to verify health facility records. Result changes were linked to premature interpretation, faint lines or product- or line-specific characteristics. Adherence to manufacturer-recommended read times may reduce the proportion of RDTs that change from negative to positive. These findings support the utility of stored RDTs in improving data quality and rational antimalarial use in malaria-affected settings.