Translational Psychiatry

Table 4 Positive and negative predictive values of the multi-domain model in different simulated settings with different prevalence.

From: Precision psychiatry with immunological and cognitive biomarkers: a multi-domain prediction for the diagnosis of bipolar disorder or schizophrenia using machine learning

Pair of classifier in	Prevalence (%)
Different scenarios	10	20	30	40	50	60	70	80	90
Balanced Se and Sp
BD vs. SZ—PPV	23	40	53	64	72	80	86	91	96
BD vs. SZ—NPV	96	91	85	65	72	63	52	39	22
BD vs. HC—PPV	25	43	57	67	75	82	88	92	93
BD vs. HC—NPV	98	96	93	90	86	80	72	60	60
SZ vs. HC—PPV	32	53	65	75	82	87	91	95	93
SZ vs. HC—NPV	96	95	92	88	84	77	61	56	36
High Se—90%	10	20	30	40	50	60	70	80	90
BD vs. SZ—PPV	17	34	44	54	64	73	81	88	94
BD vs. SZ—NPV	98	92	92	88	84	77	68	56	36
BD vs. HC—PPV	19	34	47	58	68	76	83	89	95
BD vs. HC—NPV	98	96	93	90	85	79	71	59	39
SZ vs. HC—PPV	29	47	61	71	79	84	89	94	97
SZ vs. HC—NPV	99	98	95	92	88	83	76	65	45
High Sp—90%	10	20	30	40	50	60	70	80	90
BD vs. SZ—PPV	39	59	71	79	85	90	93	94	98
BD vs. SZ—NPV	95	90	83	76	68	59	48	35	18
BD vs. HC—PPV	40	60	72	80	86	90	93	96	98
BD vs. HC—NPV	95	90	84	77	69	59	48	35	20
SZ vs. HC—PPV	46	66	77	84	88	92	95	97	99
SZ vs. HC—NPV	97	94	90	85	79	71	62	48	29

Back to article page

Search

Advanced search

Quick links