Background: In this hospital-based cohort study, we evaluated the accuracy of rapid influenza diagnostic tests (RIDTs) and explored factors affecting their results. Methods: Patients with influenza-like illnesses (ILI) visiting the emergency departments of a tertiary medical center were recruited from November 2012 to June 2015. RIDTs and real-time reverse transcriptase-polymerase chain reaction assay (RT-PCR) were used to test all subjects for influenza. Results: Among the 263 patients with influenza-like illnesses, 36% tested positive by RT-PCR. The sensitivity and specificity of RIDTs were 46.4% and 84.2%, respectively, while physicians’ gestalt showed a sensitivity of 94.3% and a specificity of 42.2%. RIDTs performed better for H3N2 than for H1N1 but showed similar performance across specimens obtained from different body sites. False-negative results from RIDTs were significantly associated with productive cough and muscle soreness compared to true negative results. Conclusions: Given the observed significant association of false-negative RIDTs with symptoms such as productive cough and muscle soreness, it is recommended to confirm patients with these symptoms and negative RIDT results with RT-PCR.