Predictive Models for Suicide Attempts in Major Depressive Disorder and the Contribution of EPHX2: A Pilot Integrative Machine Learning Study
Table 3
Classification performance of DSA recognition.
AUC (95% CI)
Sensitivity (95% CI)
Specificity (95% CI)
Accuracy (95% CI)
PPV (95% CI)
NPV (95% CI)
-score
S-Model A
0.702 (0.603, 0.802)
52.5% (36.1%, 68.5%)
73.9% (61.9%, 83.7%)
66.1% (56.4%, 74.9%)
53.8% (37.2%, 69.9%)
72.9% (60.9%, 82.8%)
0.660
S-Model B
0.752 (0.658, 0.847)
55.0% (38.5%, 70.7%)
82.6% (71.6%, 90.7%)
72.5% (63.1%, 80.6%)
64.7% (46.5%, 80.3%)
76.0% (64.7%, 85.1%)
0.719
S-Model C
0.801 (0.719, 0.884)
60.0% (43.3%, 75.1%)
81.2% (69.9%, 89.6%)
73.4% (64.1%, 81.4%)
64.9% (47.5%, 79.8%)
77.8% (66.4%, 86.7%)
0.732
AUC: area under the receiver operating characteristic curve; PPV: positive predictive value; NPV: negative predictive value; S-Model: model for recognising depressed suicide attempters.
