Table 1 Summary of comparing different methods on determining optimal cut-off in population based and simulation studies
Author | Year of publication | Underlying distribution | Sample size | Performance | |
|---|---|---|---|---|---|
D | ND | ||||
Hajian-Tilak [15] | 2018 | Normal Logistic | ✖ | ✖ | DOR had a poor performance in cut-point selection |
Unal [5] | 2017 | Normal Gamma | 50 100 200 | 50 100 200 | IU had the least MSE and relative bias compared with other methods |
50 50 50 | 100 150 200 | ||||
Habibzade et al. [11] | 2016 | ✖ | ✖ | ✖ | Considering the costs of FP and FN, the analytical methods had a better performance than others |
Perkins et al. [26] | 2006 | ✖ | 120 | 120 | The difference between the Youden and Euclidean was negligible in determining optimal cut-point |
Liu [7] | 2012 | Normal | 50 100 150 | 50 100 150 | Youden index had a higher MSE than Euclidean and product methods |
100 200 200 | 200 100 200 | ||||
Gerke et al. [27] | 2022 | Normal | ✖ | ✖ | ✖ |
Rota et al. [9] | 2014 | Normal Gamma | 50 100 200 | 50 100 200 | Eucliden and product methods had lower MSE and relative bias than Youden index |
50 50 50 | 100 150 200 | ||||