Skip to main content
. 2023 Mar 13;90:104519. doi: 10.1016/j.ebiom.2023.104519

Table 2.

Classification model performance for post-COVID, NDs and different cohort subdivisions.

Cohort 1 Cohort 2 Accuracy Sensitivity Specificity
Normal donors Post-COVID 90.1% ± 1.1% 90.6% ± 1.1% 80.0% ± 3.3%
Normal donors Post-SYMP COVID 92.5% ± 1.0% 92.8% ± 1.0% 87.5% ± 5.6%
Normal donors Post-ASYMP COVID 90.7% ± 2.1% 97% ± 1.5% 75.0% ± 6.5%
Normal donors PACS suspected 87.1% ± 1.5% 86.6% ± 1.7% 92.5% ± 3.8%
Normal donors PACS not suspected 90.0% ± 3.2% 94.3% ± 3.2% 82.5% ± 6.5%
Post-ASYMP COVID Post-SYMP COVID 70.3% ± 1.8% 69.6% ± 1.8% 78.0% ± 8.1%
PACS not suspected PACS suspected 58.7% ± 3.7% 59.4% ± 4.2% 52.5% ± 9.5%

Averages and standard errors are reported for model accuracy, sensitivity, and specificity.