Skip to main content

Advertisement

Table 5 Summary of a ridge multiple regression model for clotting parameters explaining the variability of INR as a dependent variable

From: Multivariate relationships between international normalized ratio and vitamin K-dependent coagulation-derived parameters in normal healthy donors and oral anticoagulant therapy patients

independent variable R 2 β s p β overall p
Fibrinogen 0.095 0.315 ± 0.068 5.91*10-6 -0.002 ± 0.019 0.921
APTT ratio 0.396 0.631 ± 0.055 1.98*10-23 0.024 ± 0.023 0.301
Factor II 0.893 -0.944 ± 0.024 0 -0.277 ± 0.042 3.07*10-10
Factor VII 0.884 -0.919 ± 0.028 0 -0.310 ± 0.036 1.68*10-15
Factor X 0.874 -0.935 ± 0.025 0 -0.229 ± 0.040 3.98*10-8
Protein C 0.783 -0.886 ± 0.033 0 -0.083 ± 0.035 0.019
Prothrombin F1+2 0.671 -0.820 ± 0.041 0 -0.088 ± 0.030 0.004
  1. Model estimated for log-transformed data in all variables with the accepted value of λ = 0.1. The values in the columns 2 and 4 of the table refer to the standardized regression coefficients (b i ) in the individual regression models (βs ± SE [βs]) including the individual independent variables, or to the overall ridge multiple regression model (βoverall ± SE [βoverall]) comprising all examined independent variables. The values of the determination coefficients (R2) denote the individual contributions of any given variable (estimated in single variable regression models) in explaining the overall INR variability; for the overall multiple regression model the corrected overall R2 = 0.9289; both individual and overall R2 were significant with p at least 0.0001. For the model of a ridge multiple regression with INR as a dependent variable and comprising all the independent variables the partial correlation coefficients rp are given in fig. 2.