Abstract
Abstract Background: COVID-19 has symptoms similar to several other respiratory and non-respiratory diseases, which makes differentiating them a challenging task and could lead to unnecessary use of realtime reverse transcriptase polymerase chain reaction (RT-PCR) resources. Aims: The study aimed to assess self-reported symptoms as predictors for RT-PCR positivity in suspected COVID-19 cases. Methods: This was a cross-sectional study. We retrospectively reviewed the database of COVID-19 care centres in the eastern district of Tripoli, Libya, from May to December 2020. Presenting symptoms and RT-PCR test data were extracted. Results: Of the 4593 subjects, 923 (20.1%) had positive RT-PCR result. Sensitivity for COVID-19 disease diagnosis was very low (≤ 18.2%) for all symptoms, except for myalgia (82.1%). Specificity was high for all symptoms (90.7–99.8%), except for myalgia (11.0%). Loss of taste and smell had the highest positive likelihood ratio (LR) for RT-PCR positivity (LR+ = 3.59, 95% CI: 2.95–4.37). In the multiple logistic regression, three symptoms maintained significant contribution to RT-PCR positivity; these were loss of taste and smell (odds ratio (OR) = 3.90, 95% CI: 3.04–4.99), sore throat (OR = 1.50, 95% CI: 1.02–2.19), and myalgia (OR = 0.65, 95% CI: 0.49–0.85). Other significant predictors were history of contact with a COVID-19 case (OR = 0.50, 95% CI: 0.39–0.62), and being female (OR = 1.33, 95% CI: 1.15–1.55). Conclusion: The findings of this study do not support the use of self-reported symptoms for the confirmation of COVID-19 disease in suspected cases because of their poor diagnostic properties