Quantitative reverse transcription polymerase chain reaction (RT-qPCR) is regarded as the gold-standard for diagnostic testing. However, the detection of residual viral RNA genome fragments is affecting several percent of recovered patients, which unnecessarily pro-longs quarantines or delays clinical procedures. To minimize the detection of such fragments, we introduced a single modification in the COVID-19 RT-qPCR to distinguish between infectious and non-infectious viral RNA. After validation of the assay using UVC inactivation of infectious virus, we analyzed positive COVID-19 clinical samples from two different countries. We find that after 15 days of the onset of symptoms, the modified RT-qPCR protocol leads to significantly fewer positive diagnoses in persistently positive samples compared to the standard RT-qPCR test, without compromising diagnoses within 5 days of the onset of symptoms. The method may improve test-to-release protocols and expand the tools available for clinical diagnosis.ImportanceMolecular tests can be used to detect RNA virus infections. The RT-qPCR test is currently regarded as the gold-standard, but its sensitivity to residual viral RNA genome fragments can lead to “incorrectly-positive” RT-qPCR results. Such results are different from false-positive RT-qPCR results, which can be generated due to in vitro cross-reactivity or contaminations. However, the detection of RNA fragments leads to similar incorrect conclusions about the presence of infectious virus long after a patient has recovered from a viral infection and thus false-positive diagnoses. We here modified a commercial RT-qPCR kit to make it less sensitive to residual viral RNA genome fragments, reducing the likelihood for such results in recovered COVID-19 patients. The method may improve test-to-release protocols, expand the tools available for clinical testing, and help reduce hospital encumbrance.