Table 2.
The performance of prediction model.
| Study | Best performing model | Area under the curve (95%CI)/C-value (95%CI) |
Incidence of PTS in the development cohort |
|---|---|---|---|
| Tao Yu, 2022 [12] | Extreme gradient boosting (XGBOOST) | 0.77 (0.74,0.80) | 58.90 % |
| Gradient boosting decision tree (GBDT) | 0.77 (0.74,0.80) | ||
| Lijun Zhu, 2022 [13] | Random forest | 0.722 | 21.81 % |
| Hao Huang, 2018 [14] | Cox regression analysis | 0.825 (0.747,0.903) | 47.70 % |
| Jiantao Zhang, 2022 [15] | Logistic regression | 0.773 (0.699–0.848) | 14.07 % |
| Peng Qiu, 2021 [16] | Logistic regression | 0.724 | 42.00 % |
| Anat Rabinovich, 2020 [17] | Logistic regression | 0.63 (0.59, 0.67) | 47.00 % |
| Anat Rabinovich, 2018 [18] | Logistic regression | 0.65 (0.64, 0.67) | 12.53 % |
| Marie Méan, 2018 [19] | Logistic regression | 0.77 (0.71–0.82) | 58.30 % |
| Elham E. Amin, 2018 [20] | Baseline model: Logistic regression |
0.67 (0.61, 0.73) | 45.70 % |
| Secondary model: Logistic regression | |||
| Tian'an Huang, 2022 [21] | Logistic regression | 0.825 (0.759, 0.892) | 31.90 % |