This study aimed to determine if sampling of oropharyngeal secretions (OSs) helps improves detection of SARS-CoV-2 RNA by nucleic acid amplification testing of potential patients with COVID-19. The first prospective study consisted of 75 patients with COVID-19 who were ready for discharge and who had 2 consecutive negative results per nucleic acid amplification testing (NAAT) of viral samples retrieved with nasopharyngeal swabs (NPSs). Because of detection of potential false negatives in that cohort, the NAAT results of paired OS and NPS samples from 50 additional recruits with COVID-19 during their recovery stage were used in a second prospective study to compare the diagnostic values of the 2 viral RNA sampling methods. For identification of the frequency of inconsistency between the sampling methods, the McNemar’s test was used for difference analysis and the kappa statistic for consistency analysis. OSs obtained from 2 of the 75 participants in the first study yielded positive results for SARS-CoV-2 nucleic acid. Both were male and aged >60 y. Subsequent chemiluminescence enzyme immunoassays indicated that they were positive for the SARS-CoV-2 IgM and IgG antibodies. For parallel NAAT of OS and NPS samples in the second study, McNemar’s test indicated that the difference between the frequencies of inconsistent parts of OS and NPS was statistically significant (P = 0.021). Cohen’s kappa coefficient for OS and NPS was 0.244, which is indicative of fair consistency. The NPS test has a risk of sending home more patients (59%) who still have the infection, while the OS test will make such an error in fewer patients (14%). Although OS sampling improves the accuracy of SARS-CoV-2 nucleic acid testing, it has to be emphasized that this conclusion is based on a very small sample size. Detection of viral RNA from a patient’s secretions is not confirmative of viral infectivity.