In light of supply chain failures for reagents and consumables needed for purification of nucleic acid for detection of SARS-CoV-2 RNA by RT-PCR, we aim to verify the performance and utility of a non-extraction protocol for RT-PCR ("direct RT-PCR"). We report improved sensitivity compared to earlier reports of direct RT-PCR testing of swab samples, in particular at the lower limit of detection (sensitivity 93% overall; 100% for specimens with high to moderate viral titre, Ct <34; 81% for specimens with a low viral titre, Ct ≥34). Sensitivity is improved (from 90 to 93%) by testing in duplicate. We recommend swabs are re-suspended in water to minimise PCR inhibition. A cellular target is necessary to control for PCR inhibition and specimen quality. Direct RT-PCR is best suited to population level screening where results are not clinically actionable, however in the event of a critical supply chain failure direct RT-PCR is fit for purpose for the detection of SARS-CoV-2 infection. The results from our study offer front-line laboratories additional reagent options for performing extraction-free RT-PCR protocols.