Development and validation of a predictive model for incident type 2 diabetes in middle-aged Mexican adults: the metabolic syndrome cohort

Table 4 Office-based and biochemical model for prediction of incident diabetes from Cox-proportional hazard regression models

Model parameters	Variable	β-coefficient	Wald	HR	95%CI	p-value	Points
Office-based model X² = 76.64 p < 0.0001 D_xy = 0.2915 c-statistic = 0.656	Age > 40 years	0.466	14.079	1.593	1.249–2.031	<0.001	2
	FDFH of T2D	0.242	4.560	1.273	1.020–1.590	0.033	1
	WHr > 0.5	0.725	5.809	2.065	1.145–3.725	0.016	3
	Arterial hypertension	0.503	18.471	1.654	1.315–2.080	<0.001	2
	BMI ≥30 kg/m2	0.388	11.500	1.474	1.178.1.845	0.001	2
Biochemical model X² = 446.815 p < 0.0001 D_xy = 0.4723 c-statistic = 0.752	Physical activity	−0.217	3.374	0.805	0.638–1.015	0.066	−1
	Age > 40 years	0.328	6.830	1.388	1.085–1.776	0.009	2
	TG > 150 mg/dL	0.517	14.932	1.677	1.287–2.185	<0.001	3
	Glucose 100–110 mg/dL	1.271	92.355	3.565	2.751–4.621	<0.001	4
	Glucose 111–125 mg/dL	2.097	176.167	8.138	5.971–11.091	<0.001	7
	Arterial hypertension	0.306	6.711	1.358	1.077–1.712	0.010	2
	Abdominal obesity (IDF)	0.422	5.576	1.525	1.074–2.165	0.018	2

Discrimination indexes from both regression models were obtained from k-fold cross-validation (k = 10) and were corrected for over-optimism

ISSN: 1472-6823