Use of CHAID Decision Trees to Formulate Pathways for the Early Detection of Metabolic Syndrome in Young Adults

Table 2

Decision rules for the prediction of the incidence risk of MetS from the CHAID algorithm.

Node number	Level 1	Level 2	Level 3	Level 4	MetS probability
29	94 < waist circumference ≤ 103	TG > 162	HDL ≤ 38	FPG > 94	94.4
11	Waist circumference > 103	HDL ≤ 38	*	*	82.1
17	86 < waist circumference ≤ 94	TG > 138	FPG > 92	*	52.2
12	Waist circumference > 103	38 < HDL ≤ 49	*	*	45.3
28	94 < waist circumference ≤ 103	TG > 162	HDL ≤ 38	FPG ≤ 94	40.0
21	94 < waist circumference ≤ 103	TG > 162	HDL > 38	*	31.8
27	Waist circumference > 103	HDL > 49	FPG > 103	*	26.7
19	94 < waist circumference ≤ 103	TG ≤ 162	FPG > 99	*	18.5
15	86 < waist circumference ≤ 94	TG ≤ 138	FPG > 103	*	16.7
5	Waist circumference ≤ 86	HDL ≤ 38	*	*	4.9
26	Waist circumference > 103	HDL > 49	FPG ≤ 103	*	2.1
6	Waist circumference ≤ 86	HDL > 38	*	*	0.0
14	86 < waist circumference ≤ 94	TG ≤ 138	FPG ≤ 103	*	0.0
16	86 < waist circumference ≤ 94	TG > 138	FPG ≤ 92	*	0.0
18	94 < waist circumference ≤ 103	TG ≤ 162	FPG ≤ 99	*	0.0

∗represents not significant. Growing method: exhaustive CHAID; dependent variable: MetS: metabolic syndrome, TG: triglyceride, HDL: high-density lipoprotein cholesterol, and FPG: fasting plasma glucose.