Table 2

Diagnostic accuracy in development dataset 1 using each level of the MAINTAIN instrument as possible discrimination thresholds to classify dysfunctional patients (96 individuals were classified as dysfunctional by the original instrument in this sample)

Development dataset 1 (n = 249)
MS	n (DYS)	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)	Youden’s index
8	229	100.0	13.0	41.5	100.0	0.110
9	224	100.0	16.2	42.4	100.0	0.130
10	220	100.0	18.8	43.2	100.0	0.162
11	218	100.0	20.1	43.6	100.0	0.188
12	205	100.0	28.6	46.3	100.0	0.201
13	201	100.0	31.2	47.3	100.0	0.286
14	193	100.0	36.4	49.2	100.0	0.312
15	180	98.9	44.2	52.2	98.6	0.364
16	173	98.9	48.7	54.3	98.7	0.431
17	159	96.8	56.5	57.9	96.7	0.476
18	146	95.8	64.3	62.3	96.1	0.533
19	140	94.7	67.5	64.3	95.4	0.601
20	133	90.5	69.5	64.7	92.2	0.623
21	122	86.3	74.0	67.2	89.8	0.600
22	109	81.1	79.2	70.6	87.1	0.603
23	103	75.8	79.9	69.9	84.2	0.603
24	92	70.5	83.8	72.8	82.2	0.557
25	84	63.2	84.4	71.4	78.8	0.543
26	74	60.0	89.0	77.0	78.3	0.476
27	67	55.8	90.9	79.1	76.9	0.490

MS, MAINTAIN Score (emphasis represent recommended thresholds);
n(DYS), number of individuals classified as dysfunctional by the MAINTAIN instrument at that threshold;
PPV, Positive Predictive Value;
NPV, Negative Predictive Value