A novel model for predicting prolonged stay of patients with type-2 diabetes mellitus: a 13-year (2010–2022) multicenter retrospective case–control study

Table 3 Detailed performance metrics of the four models

Models	AUC	Sensitivity	Specificity	PPV	NPV
Models	(95% CI)	(95% CI)	(95% CI)	(95% CI)	(95% CI)
Training set	0.803	0.773	0.730	0.483	0.908
Training set	0.799–0.808	0.766–0.781	0.726–0.735	0.476–0.490	0.905–0.912
Internal validation set	0.794	0.833	0.655	0.451	0.920
Internal validation set	0.788–0.800	0.823–0.843	0.648–0.662	0.441–0.461	0.915–0.925
External validation set1	0.754	0.815	0.625	0.270	0.952
External validation set1	0.739–0.770	0.791–0.838	0.613–0.638	0.254–0.285	0.945–0.959
External validation set2	0.743	0.690	0.718	0.279	0.936
External validation set2	0.722–0.763	0.654–0.725	0.704–0.732	0.258–0.301	0.927–0.944

AUC: area under the curve; PPV: positive predictive value; NPV: negative predictive value; CI: Confidence Interval

ISSN: 1479-5876