Artificial Intelligence Conformal Prediction Intervals for Malaria Rapid Diagnostic Test Interpretation Under Variable Microscopy Confirmation Delay: A Coverage Evaluation of Healthcare Overtreatment Safety

Malaria rapid diagnostic tests (RDTs) are frequently used as the sole basis for antimalarial prescribing when microscopy confirmation is delayed, yet the uncertainty of RDT-based diagnosis is rarely quantified in a way that supports safe overtreatment decisions. Using a replication-based evaluation of the intercept, exposure, group and severity estimator, 600 data sets of N = 500 were drawn from a declared data-generating process (intercept 12.0, exposure 0.42, group -3.1, severity 1.35, error SD 4.2). The study characterises the behaviour of the ordinary least squares estimator under the stated process. Across 600 replications, empirical coverage of the nominal 95% confidence intervals was 1.0 for both the intercept and the exposure coefficient, estimated bias was 0.03721 for the intercept and -0.00088 for exposure, and root mean squared error was 0.84214 and 0.01516 respectively. The results indicate that the estimator is essentially unbiased for the exposure coefficient and that nominal coverage is achieved, which supports the use of conformal prediction sets as a safety layer for RDT interpretation when microscopy confirmation is delayed. Coverage validity, rather than point accuracy alone, is the appropriate criterion for judging whether an AI-assisted RDT reading system is safe to deploy in overtreatment-sensitive settings.