Testing for SARS-CoV-2 is crucial to tracking and controlling the pandemic. In particular, rapid testing in settings such as the emergency department (ED) could improve time to diagnosis and promote proper infection control measures. Early in the COVID-19 pandemic, we implemented the Abbott ID NOW COVID-19 method for screening symptomatic ED patients. However, due to concerns of suboptimal sensitivity, samples with a negative result were reflexed to the lab for confirmatory testing by the TaqPath COVID-19 Combo RT-PCR method. This study analyzed 6773 ID NOW results from April 2020 to September 2020 in the ED, of which 10% (n = 673) were positive and reported directly. The rest 90% (n = 6100) were negative and reflexed to RT-PCR. Among them, 3% (n = 175) turned positive on RT-PCR while 97% (n = 5925) of the results were consistently negative. The cycle threshold (Ct) values of the false-negative samples (n = 175) showed 90% (n = 158) of them with relatively low viral loads (Ct > /= 30) with median Ct value at 35, while a number of samples (n = 17) had low Ct values (Ct < 30) and no clear explanation for false-negative results. Our study demonstrates that the Abbott ID NOW, despite it's sensitivity limitations, was capable of providing near real-time results for 10% of symptomatic patients presenting to the ED allowing for improved management and workflow. However, our study findings emphasize the need to reflex negative specimens to a higher sensitivity method when prevalence is high and false-negative results are intolerable.