Skip to main content
. Author manuscript; available in PMC: 2010 May 1.
Published in final edited form as: Am J Med. 2009 May;122(5):464–471. doi: 10.1016/j.amjmed.2008.10.034

Table 2.

Beta-Coefficients, Bayes Information Criteria (BIC), Calibration and Discrimination for Hypertension Risk Prediction Models, Development Cohort (N= 9,427)

Inclusive Model Simplified Model With Lipids Simplified Model Blood Pressure Only Model
Predictors Beta (95% CI) P value Beta (95% CI) P value Beta (95% CI) P value Beta (95% CI) P value
Systolic Blood Pressure
 <110 mm Hg Reference - Reference - Reference - Reference -
 110 – 119 mm Hg 0.51 (0.33,0.7) <0.001 0.5 (0.32,0.69) <0.001 0.51 (0.32,0.69) <0.001 0.58 (0.4,0.76) <0.001
 120 – 129 mm Hg 1.2 (1.01,1.4) <0.001 1.19 (0.99,1.38) <0.001 1.2 (1,1.39) <0.001 1.35 (1.16,1.55) <0.001
Diastolic Blood Pressure
 <65 mm Hg Reference - Reference - Reference - Reference -
 65–74 mm Hg 0.29 (0.07,0.52) 0.011 0.3 (0.08,0.53) 0.008 0.3 (0.08,0.53) 0.009 0.34 (0.12,0.57) 0.003
 75–84 mm Hg 0.66 (0.42,0.9) <0.001 0.67 (0.44,0.91) <0.001 0.67 (0.44,0.91) <0.001 0.76 (0.53,0.99) <0.001
Race/Ethnicity
 White/Asian/Other Reference - Reference - Reference - Reference -
 Black or Hispanic 0.59 (0.26,0.92) <0.001 0.58 (0.25,0.9) <0.001 0.58 (0.25,0.9) <0.001 - -
Total/HDL Cholesterol - - 0.09 (0.05,0.13) <0.001 - - - -
ApoB (mg\dL) 0.005 (0.003,0.007) <0.001 - - - - - -
Natural logarithm of
 Age (years) 0.92 (0.45,1.38) 0.002 1.22 (0.76,1.67) <0.001 1.28 (0.83,1.73) <0.001 - -
 BMI (kg/m2) 1.4 (1.03,1.77) <0.001 1.67 (1.31,2.03) <0.001 1.92 (1.59,2.26) <0.001 - -
 hsCRP (mg\dL) 0.13 (0.07,0.18) <0.001 - - - - - -
 Lp(a) (mg\dL) −0.06 (−0.1,−0.02) 0.002 - - - - - -
 Total Grains (servings/day) −0.17 (−0.27,−0.07) <0.001 - - - - - -
BIC 8788 8813 8820 8957
Harrell’s C-Index* (SE) 0.714 (0.013) 0.708 (0.013) 0.707 (0.013) 0.682 (0.013)
Hosmer-Lemeshow χ2 (p-value) 2.9 (0.94) 9.4 (0.31) 6.0 (0.64) 4.3 (0.83)
*

Harrell’s C-Index is analogous to the area under the ROC curve or AUC

Hosmer-Lemeshow χ2 measures the calibration of the model; significant p-values suggest lack of calibration.