Assessment of Extrinsic Validity of HIV Oral Self-Test in A Low-Prevalence Population: A Bayesian Approach

Background: HIV oral self-test (HIVOST) is used for screening, particularly in diverse populations. While these tests offer convenience and privacy, their accuracy in low-prevalence settings remains a concern. A key measure of test reliability is extrinsic validity, which considers test sensitivity, specificity, and disease prevalence to determine the probability of true positivity given a positive test result. This statistical investigation aimed to evaluate the extrinsic validity of an HIVOST in a low-prevalence population. Methodology: We applied Bayes' theorem to calculate the extrinsic validity of the HIVOST, incorporating test sensitivity, specificity, and the prevalence of HIV infection. Important Findings: The analysis yielded an extrinsic validity of 7.70% (95% CI: 1.8% to 25%), meaning that in this low-prevalence setting, the probability of an individual truly having HIV after testing positive was relatively low. The participant who tests negative has a 99.97% (95% CI: 99.94% to 99.98%) chance of truly not having HIV, indicating high reliability of a negative result in this low-prevalence setting. Sensitivity analysis shows that PPV increases with prevalence, while NPV remains high, confirming the reliability of negative results across scenarios. Conclusions: Although a positive HIVOST result strongly suggests HIV infection, the low extrinsic validity underscores the need for confirmatory testing, particularly in low-prevalence settings.