TABLE 2.

Difference in Adjusted Average Length of Stay Between Dabigatran and Warfarin Treatment Groups

Explanatory Variables	Coefficient	Exp PEa	Standard Error	P Valueb
Intercept	1.0325	2.81
Treatment Group (reference: warfarin)
Dabigatran etexilate	-0.1358	0.87	0.0369	< 0.001
Use of bridging agent (LMWH and/or UFH; reference: no)	0.3816	1.46	0.0391	< 0.001
Patient age (in years)	0.0022	1.00	0.0022	0.306
Female (reference: male)	0.0036	1.00	0.0380	0.924
Payer type (reference: commercial)
Medicare	0.0973	1.10	0.0444	0.027
Medicaid	0.1468	1.16	0.0843	0.082
Other/unspecified	0.1725	1.19	0.2167	0.426
AF was primary diagnosis (reference: secondary diagnosis)	-0.5547	0.57	0.0420	< 0.001
Comorbidities (excluding those used in stroke risk calculation)
Cancer	-0.0659	0.94	0.0486	0.175
Chronic pulmonary disease	0.1919	1.21	0.0354	< 0.001
Cerebral vascular disease	0.0513	1.05	0.0535	0.338
Dementia	0.1939	1.21	0.0943	0.040
Metastatic carcinoma	0.2189	1.24	0.1033	0.034
Liver disease (mild through severe)	0.3476	1.42	0.1090	0.001
Paraplegia and hemiplegia	0.2441	1.28	0.0873	0.005
Peptic ulcer disease	0.4098	1.51	0.1174	< 0.001
Peripheral vascular disease	0.0314	1.03	0.0495	0.526
Renal disease	0.1404	1.15	0.0461	0.002
Rheumatologic disease	0.0463	1.05	0.0741	0.532
Left ventricular systolic dysfunction	0.0869	1.09	0.0560	0.120
Thromboembolism	0.1287	1.14	0.0642	0.045
Myocardial infarction	0.0317	1.03	0.0592	0.592
Coronary artery disease	-0.0389	0.96	0.0390	0.319
Peripheral artery disease	-0.1227	0.88	0.0668	0.066
Left ventricular systolic dysfunction	-0.0659	0.94	0.0486	0.176
CHADS2 scorec (reference: 2 [moderate risk])
0-1 (low risk)	-0.0255	0.97	0.0547	0.641
3-6 (high risk)	0.1341	1.14	0.0387	0.005
HAS-BLED scored (reference: 2 [moderate risk)]
0-1 (low risk)	0.0651	1.07	0.0765	0.397
3-9 (high risk)	0.0099	1.01	0.0409	0.809
Number of prior hospitalizations	0.0562	1.06	0.0223	0.012
Hospitalization within matched study sample (reference: 1st)
2nd	0.4376	1.55	0.0889	< 0.001
3rd or greater	0.1201	1.13	0.2915	0.680
Stay include time in the ICU/CCU (reference: no)	0.2812	1.32	0.0597	< 0.001
Hospital U.S. Census divisions (reference: Pacific)
New England	-0.2156	0.81	0.1257	0.086
Middle Atlantic	0.0921	1.10	0.0641	0.151
East North Central	0.0137	1.01	0.0693	0.844
West North Central	-0.1916	0.83	0.0786	0.015
South Atlantic	-0.0127	0.99	0.0557	0.820
East South Central	-0.0975	0.91	0.1008	0.333
West South Central	-0.0778	0.93	0.0638	0.223
Mountain	-0.3743	0.69	0.0563	< 0.001
Hospital bed size (reference: 300-499 beds)
1-199 beds	-0.1069	0.90	0.0617	0.083
200-299 beds	0.1354	1.14	0.0562	0.016
500+ beds	0.0728	1.08	0.0506	0.150
Hospital location (reference: rural)
Urban	0.0143	1.01	0.0656	0.828
Unknown	0.1082	1.11	0.0915	0.237
Hospital type (reference: community)
Academic	0.1142	1.12	0.0498	0.022

^aThe multiplicative effect of this level versus the reference level (e.g., 1.46 is interpreted as 46% longer length of stay; 0.87 as 13% shorter length of stay).

^bGeneralized linear model with gamma distribution (log link), fitted by generalized estimating equations performed using The Comprehensive R Archive Network geepack package.

^cOne point each for congestive heart failure, hypertension, age ≥ 75 years, and diabetes mellitus; 2 points for transient ischemic attack or stroke.

^dOne point each for current or prior hypertension; impaired kidney function (i.e., kidney disease); impaired liver function (i.e., liver disease); stroke, major bleeding event, or having condition that predisposes (i.e., bleeding diathesis); medication use that predisposes to risk of bleeding; abnormal INR (low detection rates in these data); alcohol abuse (low detection rates in these data); and age > 65 years.

AF = atrial fibrillation; Exp PE = exponentiated parameter estimate; ICU/CCU = intensive care unit/critical care unit; INR = international normalized ratio; LMWH/PS = low molecular weight heparin/pentasaccharide; UFH = unfractionated heparin.