Rates of hospitalization among patients with deep vein thrombosis before and after the introduction of rivaroxaban

Figures & data

Figure 1. Illustration showing patients’ dispositions.

Table 1. Patients’ demographics and clinical characteristics – unmatched and matched rivaroxaban and historical LMWH/warfarin cohorts.

	Unmatched cohorts			Matched cohorts^a
	Current rivaroxaban	Historical LMWH/warfarin	Standardized difference (%)^b,c	Current rivaroxaban	Historical LMWH/warfarin	Standardized difference (%)^b,c
	(n = 134)	(n = 6347)		(n = 134)	(n = 536)
Matching factors
Demographics^d
Age, mean [median] (SD)	61.9 [66.0] (19.1)	62.7 [64.0] (17.8)	4.2	61.9 [66.0] (19.1)	62.4 [64.0] (17.9)	2.7
Gender, female, n (%)	67 (50.0)	3320 (52.3)	4.6	67 (50.0)	287 (53.5)	7.1
Primary payer type^d, n (%)
Medicare	73 (54.5)	2999 (47.3)	14.5	73 (54.5)	293 (54.7)	0.4
Medicaid	11 (8.2)	629 (9.9)	5.9	11 (8.2)	40 (7.5)	2.8
Blue Cross	16 (11.9)	689 (10.9)	3.4	16 (11.9)	68 (12.7)	2.3
HMO	7 (5.2)	553 (8.7)	13.7	7 (5.2)	31 (5.8)	2.5
Other/unknown	27 (20.1)	1477 (23.3)	7.6	27 (20.1)	104 (19.4)	1.9
Other insurance companies	15 (11.2)	956 (15.1)	11.5	15 (11.2)	51 (9.5)	5.5
Region of hospital^d, n (%)
Northeast	5 (3.7)	1167 (18.4)	46.7	5 (3.7)	19 (3.5)	1.0
South	108 (80.6)	3525 (55.5)	53.7	108 (80.6)	428 (79.9)	1.9
Midwest	6 (4.5)	881 (13.9)	32.6	6 (4.5)	29 (5.4)	4.3
West	15 (11.2)	774 (12.2)	3.1	15 (11.2)	60 (11.2)	0.0
Hospital characteristics^d, n (%)
Rural	11 (8.2)	654 (10.3)	7.2	11 (8.2)	45 (8.4)	0.7
Urban	123 (91.8)	5693 (89.7)	7.2	123 (91.8)	491 (91.6)	0.7
Large (≥500 beds)	33 (24.6)	1589 (25.0)	0.9	33 (24.6)	132 (24.6)	0.0
Teaching facility	20 (14.9)	1720 (27.1)	29.9	20 (14.9)	73 (13.6)	3.7
Admission source^d, n (%)
Physician referral	93 (69.4)	2901 (45.7)	47.9	93 (69.4)	382 (71.3)	4.1
Clinical referral	20 (14.9)	587 (9.2)	17.4	20 (14.9)	80 (14.9)	0.0
Transfer from hospital	1 (0.7)	45 (0.7)	0.4	1 (0.7)	4 (0.7)	0.0
Other source	20 (14.9)	2814 (44.3)	64.4	20 (14.9)	70 (13.1)	5.4
VTE risk factors^d,e, n (%)
Hypertension	71 (53.0)	3620 (57.0)	8.1	71 (53.0)	276 (51.5)	3.0
Age ≥60 years	77 (57.5)	3535 (55.7)	3.6	77 (57.5)	300 (56.0)	3.0
Hyperlipidemia	31 (23.1)	1879 (29.6)	14.7	31 (23.1)	117 (21.8)	3.1
Tobacco	31 (23.1)	1486 (23.4)	0.7	31 (23.1)	102 (19.0)	10.1
Diabetes	26 (19.4)	1357 (21.4)	4.9	26 (19.4)	100 (18.7)	1.9
Obesity	21 (15.7)	802 (12.6)	8.7	21 (15.7)	85 (15.9)	0.5
COPD	16 (11.9)	606 (9.5)	7.7	16 (11.9)	55 (10.3)	5.3
Central venous catheter	12 (9.0)	527 (8.3)	2.3	12 (9.0)	48 (9.0)	0.0
Active malignancy	10 (7.5)	479 (7.5)	0.3	10 (7.5)	48 (9.0)	5.4
Other serious infection	8 (6.0)	460 (7.2)	5.1	8 (6.0)	45 (8.4)	9.4
Major surgery	14 (10.4)	449 (7.1)	11.9	14 (10.4)	53 (9.9)	1.9
Congestive heart failure	12 (9.0)	379 (6.0)	11.4	12 (9.0)	40 (7.5)	5.4
Number of VTE risk factors^d, n (%)
0	12 (9.0)	628 (9.9)	3.2	12 (9.0)	42 (7.8)	4.0
1	31 (23.1)	1130 (17.8)	13.2	31 (23.1)	140 (26.1)	6.9
2	21 (15.7)	1473 (23.2)	19.0	21 (15.7)	90 (16.8)	3.0
≥3	70 (52.2)	3116 (49.1)	6.3	70 (52.2)	264 (49.3)	6.0
Nonmatching factors
Year of index hospitalization^d, n (%)
2011	0 (0.0)	3616 (57.0)	126.2	0 (0.0)	332 (61.9)	134.0
2012	17 (12.7)	2731 (43.0)	67.7	17 (12.7)	204 (38.1)	58.3
2013	117 (87.3)	0 (0.0)	176.0	117 (87.3)	0 (0.0)	176.0

^aRivaroxaban patients were matched exactly with four historical LMWH/warfarin patients using propensity-score calipers of 5%.

^bFor continuous variables, the standardized difference is calculated by dividing the absolute difference in means of the LMWH/warfarin and the rivaroxaban cohorts by the pooled SD of both groups. The pooled SD is the square root of the average of the squared SDs.

^cFor categorical variables with two levels, the standardized difference is calculated using the equation below where P is the respective proportion of participants in each group: (P_{LMWH/warfarin}-P_rivaroxaban)/√[p(1-p)], where p = (P_{LMWH/warfarin} + P_rivaroxaban)/2.

^dEvaluated up to 12 months before and on the index hospital visit.

^eAdditional matching VTE risk factors (i.e., frequency <5%) not reported in this table include rheumatoid arthritis; multiple trauma; pneumonia; abdominal surgery; varicose veins; treatment with erythropoiesis-stimulating agents; coagulation defect; inflammatory bowel disease; treatment with aromatase inhibitors; hip, pelvis, or leg fracture; thrombophilia; treatment with selective estrogen receptor modulators; surgical resection of abdominal or pelvic cancer; contraceptive pill (use of oral); pregnancy; immobility; spinal cord injury; total hip replacement; and total knee replacement. All these additional matching VTE risk factors were well balanced with standardized differences <10%.

Abbreviations: COPD = Chronic obstructive pulmonary disease; HMO = Health Maintenance Organization; LMWH = Low-molecular-weight heparin; SD = Standard deviation; VTE = Venous thromboembolism.

Table 2. Comparison of the rate of hospital admission among rivaroxaban and LMWH/warfarin users.

	Number of hospital admissions		Rate of hospital admission (%)		RR
	Current cohort	Historical cohort	Current cohort	Historical cohort	RR (95% CI)^a	P value^b
Matched rivaroxaban and historical LMWH/warfarin, n	134^c	536^d
Patients not discharged within 24 hours	80	441	60	82	0.73 (0.62–0.84)	<0.001
Comparison to assess the time trend in hospital admissions among different populations
Time-trend matched cohort 1, n	536^e	536^d
Patients not discharged within 24 hours	426	441	79	82	0.97 (0.91–1.02)	0.242
Time-trend matched cohort 2, n	1781^f	3562^g
Patients not discharged within 24 hours	1390	2876	78	81	0.97 (0.94–0.99)	0.018
Time-trend matched cohort 3, n	1915^h	3830^i
Patients not discharged within 24 hours	1470	3093	77	81	0.95 (0.92–0.98)	<0.001

^a95% CIs were obtained using nonparametric bootstraps with 999 replications.

^bThe P values were estimated using the achieved significance level as reported in Efron B, Tibshirani RJ. An introduction to the bootstrap. First edition. New York: Chapman & Hall; 1993:436.

^cAll rivaroxaban patients.

^dHistorical warfarin patients were matched 4:1 to rivaroxaban patients.

^eCurrent LMWH/warfarin patients were matched 1:1 to the same historical LMWH/warfarin patients that were matched to all rivaroxaban patients.

^fAll current LMWH/warfarin patients.

^gHistorical LMWH/warfarin patients were matched 2:1 to all current LMWH/warfarin patients.

^hAll current patients (rivaroxaban and current LMWH/warfarin).

ⁱHistorical LMWH/warfarin patients were matched 2:1 to all current patients (rivaroxaban and current LMWH/warfarin).

Abbreviations: CI = Confidence interval; LMWH = Low-molecular-weight heparin; RR = Rate ratio.

Figure 2. Effect of rivaroxaban versus LMWH/warfarin on hospitalization avoided in DVT patients using rivaroxaban.

Table 3. Comparison of the rate of subsequent hospital visits among matched rivaroxaban and historical LMWH/warfarin users.^a

Subsequent hospital visits	Number of patients with subsequent hospital visit^b		Rate of subsequent hospital visits (%)		RR
	Current cohort	Historical cohort	Current cohort	Historical cohort	RR (95% CI)^c	P value^d
Within 1 month, n	134	536
Patients with a subsequent hospital visit	2	12	1.5	2.2	0.67 (0.00–2.29)	0.523
Within 2 months, n	133	532
Patients with a subsequent hospital visit	4	25	3.0	4.7	0.64 (0.15–1.39)	0.266
Within 3 months, n	125	500
Patients with a subsequent hospital visit	6	31	4.8	6.2	0.77 (0.24–1.64)	0.531
Within 6 months, n	90	360
Patients with a subsequent hospital visit	4	26	4.4	7.2	0.62 (0.13–1.60)	0.342

^aPatients with at least 1 month, 2 months, 3 months, and 6 months from their index hospital visit until the end of data (December 2013) were included.

^bAny hospital visit after the index hospital visit was considered a subsequent hospital visit, regardless of whether the new hospital visit was followed by a discharge within 24 hours or an admission.

^c95% CIs were obtained using nonparametric bootstraps with 999 replications.

^dThe P values were estimated using the achieved significance level as reported in Efron B, Tibshirani RJ. An introduction to the bootstrap. First edition. New York: Chapman & Hall; 1993:436.

Abbreviation: CI = Confidence interval; LMWH = Low-molecular-weight heparin; RR = Rate ratio.