Conducting numerous, rapid, and reliable PCR tests for SARS-CoV-2 is essential for our ability to monitor and control the current COVID-19 pandemic. Here, we tested the sensitivity and efficiency of SARS-CoV-2 detection in clinical samples collected directly into a mix of lysis buffer and RNA preservative, thus inactivating the virus immediately after sampling. We tested 79 COVID-19 patients and 20 healthy controls. We collected two samples (nasopharyngeal swabs) from each participant: one swab was inserted into a test tube with Viral Transport Medium (VTM), following the standard guideline used as the recommended method for sample collection; the other swab was inserted into a lysis buffer supplemented with nucleic acid stabilization mix (coined NSLB). We found that RT-qPCR tests of patients were significantly more sensitive with NSLB sampling, reaching detection threshold 2.1±0.6 (Mean±SE) PCR cycles earlier then VTM samples from the same patient. We show that this improvement is most likely since NSLB samples are not diluted in lysis buffer before RNA extraction. Re-extracting RNA from NSLB samples after 72 hours at room temperature did not affect the sensitivity of detection, demonstrating that NSLB allows for long periods of sample preservation without special cooling equipment. We also show that swirling the swab in NSLB and discarding it did not reduce sensitivity compared to retaining the swab in the tube, thus allowing improved automation of COVID-19 tests. Overall, we show that using NSLB instead of VTM can improve the sensitivity, safety, and rapidity of COVID-19 tests at a time most needed.