Determining a diagnostic cut-off on the Teate Depression Inventory

Figures & data

Table 1 Descriptive statistics of the participant sample (N=125)

	Frequency	Percent (%)
Women	89	71.2
Age, M ± SD (range) in years	42.71±15.26 (18–79)
Educational attainment, M ± SD in years	11.43±4.43
Diagnostic groups*
Nondepressed	34	10
TDI, M ± SD	14.35±6.07
Mildly depressed	21	19
TDI, M ± SD	26.86±9.48
Moderately depressed	33	29
TDI, M ± SD	42.00±15.77
Severely depressed	37	42
TDI, M ± SD	60.82±9.93

Note:

* Diagnoses were conducted according to the patient version of the Structured Clinical Interview for the DSM-IV-TR Clinical Version (SCID-I; First et al).³⁶

Abbreviations: M, mean; SD, standard deviation; TDI, Teate Depression Inventory; DSM-IV-TR, Diagnostic and Statistical Manual of Mental Disorders Version IV, Text Revision.

Figure 1 Receiver operating characteristic (ROC) curves for the Teate Depression Inventory (TDI) (red curve).

Notes: The chance line is the green line. The closer the curve is to the upper left hand corner of the graph, the better the diagnostic performance of the TDI. ROC curves were constructed by comparing the nondepressed group versus the mildly depressed group (A); the nondepressed and mildly depressed groups versus the moderately depressed group (B); and the nondepressed, mildly depressed, and moderately depressed groups versus the severely depressed group (C).

Table 2 Coordinates of the three ROC curves

First cut-off			Second cut-off						Third cut-off
Cut-off	Sensitivity	1 – Specificity	Cut-off	Sensitivity	1 – Specificity	Cut-off	Sensitivity	1 – Specificity	Cut-off	Sensitivity	1 – Specificity	Cut-off	Sensitivity	1 – Specificity	Cut-off	Sensitivity	1 – Specificity
≥1	1	1	≥1	1	1	≥38.5	0.727	0.018	≥1	1	1	≥38.5	1	0.284	≥71.5	0.108	0.023
≥2	1	0.971	≥2	1	0.982	≥39.5	0.697	0.018	≥2	1	0.989	≥39.5	1	0.273	≥73.5	0.054	0.023
≥5	1	0.941	≥5	1	0.964	≥40.5	0.636	0.018	≥5	1	0.977	≥40.5	1	0.25	≥74.5	0.027	0.023
≥6.5	0.952	0.941	≥6.5	0.97	0.945	≥41.5	0.606	0.018	≥6.5	1	0.955	≥41.5	1	0.239	≥78	0.027	0.011
≥7.5	0.905	0.853	≥7.5	0.939	0.873	≥42.5	0.576	0.018	≥7.5	1	0.898	≥42.5	1	0.227	≥82	0	0
≥8.5	0.905	0.824	≥8.5	0.939	0.855	≥43.5	0.515	0.018	≥8.5	1	0.886	≥43.5	1	0.205
≥10	0.905	0.794	≥9.5	0.939	0.836	≥44.5	0.394	0.018	≥9.5	1	0.875	≥44.5	0.892	0.159
≥11.5	0.905	0.765	≥10.5	0.909	0.836	≥45.5	0.273	0.018	≥10.5	1	0.864	≥45.5	0.892	0.114
≥12.5	0.905	0.735	≥11.5	0.909	0.818	≥46.5	0.212	0.018	≥11.5	1	0.852	≥46.5	0.892	0.091
≥13.5	0.905	0.676	≥12.5	0.909	0.8	≥48	0.182	0.018	≥12.5	1	0.841	≥47.5	0.838	0.08
≥14.5	0.857	0.618	≥13.5	0.909	0.764	≥49.5	0.152	0.018	≥13.5	1	0.818	≥48.5	0.811	0.08
≥15.5	0.857	0.382	≥14.5	0.909	0.709	≥51	0.152	0	≥14.5	1	0.784	≥49.5	0.811	0.068
≥16.5	0.857	0.265	≥15.5	0.909	0.564	≥55	0.121	0	≥15.5	1	0.693	≥50.5*	0.811	0.057
≥17.5	0.857	0.147	≥16.5	0.909	0.491	≥62	0.091	0	≥16.5	1	0.648	≥51.5	0.757	0.057
≥18.5	0.857	0.088	≥17.5	0.909	0.418	≥70.5	0.061	0	≥17.5	1	0.602	≥53.5	0.73	0.045
≥21*	0.857	0.059	≥18.5	0.909	0.382	≥78	0.03	0	≥18.5	1	0.58	≥55.5	0.703	0.045
≥24.5	0.762	0.059	≥21	0.909	0.364	≥82	0	0	≥21	1	0.568	≥56.5	0.649	0.045
≥26.5	0.571	0.059	≥24.5	0.879	0.327				≥24.5	1	0.534	≥57.5	0.622	0.045
≥28	0.476	0.059	≥26.5	0.848	0.255				≥26.5	1	0.477	≥59.5	0.622	0.034
≥29.5	0.429	0.029	≥28	0.848	0.218				≥28	1	0.455	≥61.5	0.595	0.034
≥30.5	0.286	0.029	≥29.5	0.848	0.182				≥29.5	1	0.432	≥63.5	0.541	0.034
≥32	0.143	0.029	≥30.5	0.848	0.127				≥30.5	1	0.398	≥65.5	0.432	0.034
≥34	0.143	0	≥32	0.848	0.073				≥32	1	0.364	≥66.5	0.324	0.023
≥35.5	0.095	0	≥34	0.848	0.055				≥34	1	0.352	≥67.5	0.243	0.023
≥43	0.048	0	≥35.5*	0.848	0.036				≥35.5	1	0.341	≥68.5	0.216	0.023
≥51	0	0	≥37	0.818	0.018				≥37	1	0.318	≥69.5	0.135	0.023

Note:

* Cut-off value maximizes both sensitivity and specificity.

Abbreviation: ROC, receiver operating characteristic.

Table 3 Prediction parameters for total Teate Depression Inventory (TDI) score (N=125)

Cut-off score	Sensitivity	Specificity	Positive predictive power	Negative predictive power	Classification accuracy
21	0.86	0.94	0.94	0.87	0.90
36	0.82	0.98	0.98	0.84	0.90
50	0.81	0.94	0.93	0.83	0.88

Notes: Sensitivity = true positives/(true positives + false negatives); specificity = true negatives/(true negatives + false positives); positive predictive power = true positives/(true positives + false positives); negative predictive power = true negatives/(true negatives + false negatives); classification accuracy = (true positives + true negatives)/(true positives + true negatives + false positives + false negatives).

FirstMBSpitzerRLGibbonMWilliamsJBStructured Clinical Interview for DSM-IV Axis I Disorders, (SCID-I), Clinical VersionWashington, DCAmerican Psychiatric Association1997

 Google Scholar