Skip to main content
. Author manuscript; available in PMC: 2016 Oct 5.
Published in final edited form as: J Manag Care Spec Pharm. 2016 Mar;22(3):281–292. doi: 10.18553/jmcp.2016.22.3.281

TABLE 2.

Estimated Coefficients of Key Covariates from a Multinomial Logistic Regression

Variable NOAC Effect P Value EXP (Effect)
Intercept Dabigatran   −2.35 < 0.0001 0.10
Rivaroxaban   −2.95 < 0.0001 0.05
Time vs. July 4, 2012–December 31, 2012 October 1, 2010–March 3, 2011 Dabigatran   −0.69 < 0.0001 0.50
October 1, 2010–March 3, 2011 Rivaroxaban −16.21   0.930 0.00
March 4, 2011–July 3, 2011 Dabigatran     0.19   0.020 1.21
March 4, 2011–July 3, 2011 Rivaroxaban −16.14   0.930 0.00
July 4, 2011–November 3, 2011 Dabigatran     0.48 < 0.0001 1.62
July 4, 2011–November 3, 2011 Rivaroxaban   −5.61 < 0.0001 0.00
November 4, 2011–March 3, 2012 Dabigatran     0.41 < 0.0001 1.51
November 4, 2011–March 3, 2012 Rivaroxaban   −1.62 < 0.0001 0.20
March 4, 2012–July 3, 2012 Dabigatran     0.31 < 0.0001 1.37
March 4, 2012–July 3, 2012 Rivaroxaban   −0.44 < 0.0001 0.65
Region vs. Northeast Midwest Dabigatran   −0.28 < 0.0001 0.75
Midwest Rivaroxaban     0.11   0.490 1.11
South Dabigatran     0.19   0.000 1.21
South Rivaroxaban     0.83 < 0.0001 2.29
West Dabigatran   −0.02   0.830 0.98
West Rivaroxaban     0.07   0.700 1.07
Age vs. < 65 years 65–74 Dabigatran     0.62 < 0.0001 1.87
65–74 Rivaroxaban     0.36   0.100 1.44
75 or older Dabigatran     0.53 < 0.0001 1.69
75 or older Rivaroxaban     0.54   0.010 1.72
Sex Women Dabigatran   −0.08   0.090 0.92
Women Rivaroxaban     0.24   0.010 1.28
Race vs. White Black Dabigatran   −0.38   0.000 0.68
Black Rivaroxaban   −0.54   0.010 0.58
Hispanic Dabigatran   −0.18   0.110 0.83
Hispanic Rivaroxaban   −0.15   0.570 0.86
Asian Dabigatran     0.10   0.550 1.10
Asian Rivaroxaban     0.10   0.790 1.10
Other Dabigatran     0.22   0.210 1.25
Other Rivaroxaban   −0.21   0.650 0.81
Death Died Dabigatran   −0.58 < 0.0001 0.56
Died Rivaroxaban   −0.68   0.010 0.51
Patient’s insurance, cost-sharing status, and drug coverage benefit at the initiation of an anticoagulant vs. non-LIS in the initial phase Dual-eligibles Dabigatran   −0.35 < 0.0001 0.71
Dual-eligibles Rivaroxaban   −0.62 < 0.0001 0.54
Non-dual LIS Dabigatran   −0.26   0.010 0.77
Non-dual LIS Rivaroxaban   −0.39   0.080 0.67
Non-LIS in the coverage gap phase Dabigatran   −0.03   0.790 0.97
Non-LIS in the coverage gap phase Rivaroxaban   −0.22   0.320 0.80
Non-LIS in the catastrophic phase Dabigatran     0.35   0.130 1.42
Non-LIS in the catastrophic phase Rivaroxaban   −0.49   0.420 0.62
National PACE or employer-sponsored Dabigatran     0.31 < 0.0001 1.36
National PACE or employer-sponsored Rivaroxaban     0.24   0.090 1.28
Preexisting chronic conditions Acute myocardial infarction Dabigatran   −0.27   0.010 0.77
Acute myocardial infarction Rivaroxaban   0.07   0.710 1.07
Stroke/transient ischemic attack Dabigatran   −0.19   0.000 0.83
Stroke/transient ischemic attack Rivaroxaban   −0.53 < 0.0001 0.59
Chronic kidney disease Dabigatran   −0.40 < 0.0001 0.67
Chronic kidney disease Rivaroxaban   −0.34   0.000 0.71
Congestive heart failure Dabigatran   −0.29 < 0.0001 0.75
Congestive heart failure Rivaroxaban   −0.49 < 0.0001 0.61
Hypertension Dabigatran     0.40 < 0.0001 1.49
Hypertension Rivaroxaban     0.58   0.000 1.79
Previous use of other medications NSAID Dabigatran     0.48 < 0.0001 1.62
NSAID Rivaroxaban     0.72 < 0.0001 2.05
Antiplatelet Dabigatran     0.16   0.070 1.18
Antiplatelet Rivaroxaban     0.11   0.560 1.12

Note: Variables were selected by the stepwise variable selection method. Thirteen variables were selected among 37 variables.

LIS = low-income subsidy; NOAC = novel oral anticoagulant; NSAID = nonsteroidal anti-inflammatory drug; PACE = Programs of All-Inclusive Care for the Elderly.