Cost of Treating Venous Thromboembolism With Heparin and Warfarin Versus Home Treatment With Rivaroxaban

Transcription

1 ORIGINAL CONTRIBUTION Cost of Treating Venous Thromboembolism With Heparin and Warfarin Versus Home Treatment With Rivaroxaban Zachary P. Kahler, MD, Daren M. Beam, MD, MS, and Jeffrey A. Kline, MD Abstract Background: Target-specific anticoagulants such as rivaroxaban facilitate immediate discharge of lowrisk venous thromboembolism (VTE; including deep vein thrombosis [DVT] and pulmonary embolism [PE]) allowing treatment at home instead of hospitalization. Objectives: The objective was to compare costs accrued over 6 months by patients diagnosed with lowrisk VTE and treated at home with rivaroxaban versus usual care with heparin-warfarin. Methods: This case-control study calculated costs using the established charge-to-cost ratio from UB-04 billing claims of patients diagnosed at two metropolitan hospitals. Patients were defined as low risk by the Hestia criteria. All patients were anticoagulated for 6 months. Control patients were treated with usual care using low-molecular-weight heparin (LMWH) and then warfarin. Case patients were treated with an initial dose of rivaroxaban in the ED followed by same-day discharge home with rivaroxaban. Medians were compared by Mann-Whitney U-test. Results: Fifty cases and 47 controls were identified. Groups were well matched according to mean age, Charlson comorbidity score, and proportions by sex and location of thrombus. For all VTEs, median hospital charges for 6 months after diagnosis were $11,128 (interquartile range [IQR] = $8,110 to $23,390) for controls, compared with $4,787 (IQR = $3,042 to $7,596) for cases (Mann-Whitney U-test p < 0.001). Subgroup analyses of the first week of therapy, PE, DVT, and inpatient pharmacy costs retained significance, with costs for rivaroxaban-treated PE patients 57% lower than control PE patients (p < 0.001) and 56% lower for DVT patients (p = 0.003). Conclusions: Cost of medical care was lower for low-risk VTE patients discharged immediately from the ED with rivaroxaban therapy compared with patients treated with LMWH-warfarin. ACADEMIC EMERGENCY MEDICINE 2015;22: by the Society for Academic Emergency Medicine This article concerns the economic cost of medical care for low-risk ED patients with deep vein thrombosis (DVT) and pulmonary embolism (PE), known collectively venous thromboembolism (VTE). The context of this topic includes both the recent approval of target-specific anticoagulants (TSAs) for the treatment of VTE, as well as the unclear benefits of hospitalizing all patients diagnosed with PE. While DVT has been treated as an outpatient for many years, recently developed risk stratification tools have been validated to select patients with PE who have a low risk of shortterm complications. 1 7 In particular, the Hestia criteria identifies a subgroup of PE patients who have a 1.0% mortality rate as outpatients over 3 months and a low From the Department of Emergency Medicine (ZPK, DMB, JAK) and the Department of Cellular and Integrative Physiology (DMB, JAK), Indiana University School of Medicine, Indianapolis, IN. Received December 9, 2014; revisions received January 12 and January 21, 2015; accepted January 23, Presented at the Society for Academic Emergency Medicine Annual Meeting, Dallas, TX, May 2014; and the American College of Emergency Physicians Research Forum, Chicago, IL, October This study was funded in part by the Lilly Physician Scientist Award. JAK has received research grant funding from AHRQ, NIH, Genentech, Pfizer, and Octapharma within the past 3 years; owns stock in a medical device company, CP Diagnostics LLC; and is an inventor on a patent pending involving the subject matter. He is also a consultant to Sanofi-Aventis and a speaker for Sanofi-Aventis and received research funding from Sanofi-Aventis and from GlaxoSmithKline (current work). No other authors have any potential conflicts of interest to disclose. Supervising Editor: Brian Hiestand, MD, MPH. Address for correspondence and reprints: Jeffrey A. Kline, MD; jefkline@iupui.edu. 796 ISSN by the Society for Academic Emergency Medicine 796 PII ISSN doi: /acem.12713

2 ACADEMIC EMERGENCY MEDICINE July 2015, Vol. 22, No rate of short-term complications requiring rehospitalization. 8 In this article, we assumed that many patients with DVT have undiagnosed PE and therefore must undergo the same risk stratification as patients with documented PE. 9 Prior to TSAs, vitamin K antagonists (e.g., warfarin) were the mainstay of VTE treatment. While warfarin itself costs pennies per tablet, its use incurs multiple indirect costs, including the cost of low-molecularweight heparin (LMWH) and prothrombin time checks. Even highly compliant patients who report for prothrombin time checks are therapeutically anticoagulated at only about 70% of visits. 10 This poor time in therapeutic range leads to VTE recurrence in about 5% of patients during therapy, requiring a second prescription of LMWH. Moreover, warfarin poses three obstacles to immediate discharge of VTE patients in the ED: 1) patients must be taught to self-administer LMWH or have another person administer it, 2) patients must have access to warfarin and LMWH, and 3) patients must have access to qualified medical care facilities to monitor their anticoagulation within 3 to 5 days. 11,12 In our anecdotal experience, all three requirements are met in fewer than one-half of low-risk VTE cases in urban emergency departments (EDs) in the United States. The authors have developed a protocol for home treatment of low-risk patients with VTE diagnosed in two urban EDs using a TSA. 8 We compare the costs of medical care accrued by patients treated with this protocol to a matched, contemporaneous group of patients treated with LMWH-warfarin (hereafter referred to as warfarin or control ). We hypothesized a lower cost of care for patients treated with TSAs over a period of 6 months. We further hypothesized that the primary determinate of cost would be the need for hospitalization at the time of VTE diagnosis. METHODS Study Design This was a case-control study. The protocol was approved as a research study by the Indiana University Institutional Review Board (IRB) and is registered (NCT ). This case-control cost analysis was separately approved and registered by the Indiana University IRB ( ). Study Setting and Population The case patients were drawn from a protocol designed to allow TSA-based, home treatment of low-risk patients with DVT and PE diagnosed in two EDs in Indianapolis, Indiana. 8 These patients treated with TSAs follow up at 3 weeks with our outpatient clinic to ensure compliance. The outpatient treatment protocol was presented and approved as standard medical care to the pharmacy and therapeutics committees at the two hospitals. As part of our TSA-based home treatment protocol, patients who were low risk by Hestia criteria were immediately discharged from the ED after a single dose of rivaroxaban and enoxaparin. They were given prescriptions for 2 months of rivaroxaban by the treating emergency physicians. As described separately, the physicians completed a REDCap survey to confirm patient eligibility, notify the clinic providers, and ensure follow-up and access to medication. 8 The case population was composed of the first 50 consecutive patients from this protocol, from March to December, Inclusion criteria for both case and control patients were as follows: 1) image-proven DVT or PE on compression ultrasonography of an extremity or computerized tomographic pulmonary angiography of the chest, 2) charted evidence to confirm low-risk by Hestia criteria, 3) no overlap between case and control patients, and 4) 6 months of hospital billing data. Control subjects were identified from an institutional radiology database shared by the two hospitals prior to initiation of the TSA-based home treatment protocol from January 1, 2013, and consecutively enrolled until the number of DVT and PE patients were matched with case patients. All control DVT and PE patients were treated with LMWH and warfarin sodium for systemic anticoagulation and were diagnosed in the ED. During the chart review, data abstraction was performed by one investigator (ZPK) to obtain Charlson comorbidity scores, VTE recurrence, and mortality data. Study Protocol For the control group, 319 patients were identified in the radiology database using structured wildcard search terms to identify all patients with VTE. Of these 319, a manual review by one author (ZPK) of the radiology reports revealed 70 patients who met explicit inclusion criteria (written evidence that a board-certified radiologist or vascular medicine specialist interpreted the presence of clot on computerized tomographic pulmonary angiography or venous ultrasound, respectively). The most common reason for failure to meet inclusion criteria was existence of no prior to the search term; for example, no evidence of pulmonary embolism would still result in a positive search term in the radiology database. The charts of the potential patients were then assessed for Hestia criteria inclusion until 50 control patients had been identified. Appropriate matching of patients by illness severity (the major driver of cost) was ensured a priori by the fact that all patients were selected to be Hestia negative, and controls were selected to have similar a distribution of clot locations as the cases. The Charlson Comorbidity Index score was the preplanned outcome measure to assess validity of matching. Three of these 50 patients billing records were unable to be located by the hospital billers, presumably because of a medical record number mismatch, and were thus excluded from the final cost analysis as they had no data. The billing department for each hospital provided UB-04 forms that contained all charges incurred by patients over 6 months, starting with the initial date of diagnosis. These charges were then imported into SPSS, Version 22.0, and analyzed. For the primary analysis, an attempt was made to analyze only charges associated with VTE diagnosis and treatment. Charges that two authors agreed were not associated with VTE diagnosis or treatment (such as follow-up visits to orthopedic surgery, lab draws for lipid checks, and spe-

3 798 Kahler et al. COST OF OUTPATIENT TREATMENT cialty visits such as ophthalmology referrals) were removed for the primary analysis. To ensure that this unblinded culling of charges did not affect our results, a sensitivity analysis was also performed for all charges for medical care. To obtain representative cash costs for medications, one author (ZPK) called seven pharmacies, located at the planned multicenter sites for a related grant, to determine the out-of-pocket cost of rivaroxaban, enoxaparin, and warfarin. This amount was intentionally not adjusted for insurance coverage, patient assistance programs, or financial hardship incentives, to obtain the unadjusted cost billed by the pharmacy. The cost of medical care can be estimated from charges using an adjustment factor, also known as the cost-to-charge ratio (CCR) method, using the Medicare impact file to correct for markup. 13 The Wishard operating CCR in 2013 was 0.417, while Methodist had an operating CCR of These ratios were used to estimate each patient s cost of medical care. Outcomes The primary outcome was the difference in total hospital system charges over 6 months between patients treated with TSAs (case group) and those treated with LMWH-warfarin (control group). charges were taken directly from each hospital s UB-04 forms. The UB-04 form contains revenue codes that the hospital submits to insurers and Medicare for reimbursement. To obtain this amount, all UB-04 charges were summed, starting from the date of diagnosis and ending at 6 months. Secondary analyses include: 1) subgroup analysis of the cost of initial diagnosis, determined from the UB-04 form coded to the date of initial diagnosis, which encompassed either isolated ED visit or an entire hospital bill; 2) subgroup analysis of pharmacy charges from the UB-04 forms, including cost of medications and administration; 3) 6-month charges for only DVT patients; 4) 6-month charges for only PE patients; 5) subgroup analysis of patients who were never hospitalized; and 6) a sensitivity analysis of all charges, regardless of relationship to VTE diagnosis, over 6 months. Data Analysis Study data were collected and managed using REDCap electronic data capture tools hosted at Indiana University. 14 REDCap (Research Electronic Data Capture) is a secure, Web-based application designed to support data capture for research studies, providing an intuitive interface for validated data entry, audit trails for tracking data manipulation and export procedures, automated export procedures for seamless data downloads to common statistical packages, and procedures for importing data from external sources. We estimated our sample size using previous charge data that were similarly obtained from billing data for 22 patients with VTE diagnosed in urban EDs in the United States. 15 The raw charge data were not normally distributed (p < 0.05, Shapiro-Wilk), but log 10 transformation established normality. The mean (SD) log 10 (- charge) was $4.31 ($0.485). We assumed that an important difference in charges in terms of health policy would require a 50% reduction in the mean of the log 10 of the cost and charge. Sample size was estimated with an a of 0.05 and b of 0.20, for a statistical power of 80%. This calculation yielded n = 42 needed per group. Statistical analysis was performed using SPSS (V.22). Analysis of charges was performed using a Mann-Whitney U-test for medians as data had nonnormal distributions. Categorical variables were assessed using a chisquare test. Descriptive statistics were determined for appropriate variables. Ninety-five percent confidence intervals (CIs) were determined for t-statistics. Data were calculated to a significance of a = 0.05 and b = 0.20 where appropriate. We did not adjust alpha because we report a single predetermined primary outcome (6- month costs) with several post hoc secondary outcomes. RESULTS We requested 6 months of hospital UB-04 charge data for 50 case (rivaroxaban) and 50 control (LMWH-warfarin) subjects. However, the billing data could not be found by the hospital billing department for three control subjects identified per our consecutive enrollment protocol, leading to 47 control subjects with charge data. Demographic data and clinical features of cases and controls are compared in Table 1. There were no significant differences between groups for key demographic data, location of clot, or prior VTE. As all patients were selected by the Hestia criteria, which require the absence of significant medical conditions, comorbid conditions were rare in both groups, consisting mainly of chronic lung disease, diabetes, and prior myocardial infarction. The main outcome of match validity, the mean (SD) Charlson Comorbidity Index, was 2.6 (3.57) for cases and 2.78 (2.57) for controls (p = 0.78), suggesting that the patients were well matched for illness severity. Of the 47 control patients, 18 (38%, all of which were DVT) were treated at home. All control patients were treated with bridging LMWH for 5 to 7 days and were then transitioned to warfarin (LMWH-warfarin). All 50 case (rivaroxaban) patients were discharged home from the ED on the same day of diagnosis. There were no deaths during the 6-month period in any of the patients. No case or control patients had VTE recurrence within the 6-month period. Table 2 presents initial charges and costs of medical care for the first week after VTE diagnosis. When taken in aggregate, case patients treated with rivaroxaban were found to have median initial hospital and pharmacy charges of $2,854 (interquartile range [IQR] = $2,497 to $4,722) compared to control patients treated with LMWH-warfarin, with median initial charges of $8,080 (IQR = $4,363 to $16,387; Mann-Whitney U p = 0.003). Table 3 presents charges and costs of medical care for 6 months after diagnosis. Over 6 months, the cases treated with rivaroxaban had median total charges of $4,787 (IQR = $3,041 to $7,596) compared to controls treated with LMWH-warfarin, who had median charges of $11,128 (IQR = $8,110 to $23,390; Mann- Whitney U p = 0.002). Analysis of the DVT and PE subgroups demonstrated lower charges and costs with rivaroxaban treatment. Patients with PE had initial median charges of $4,435

4 ACADEMIC EMERGENCY MEDICINE July 2015, Vol. 22, No Table 1 Demographic and Clinical Characteristics of the Patients Rivaroxaban Group (n = 50) LMWH-Warfarin Group (n = 47) Characteristic n or average % or SD n or average % or SD p-value* Mean age (yr) Male sex Race White Black Other Ethnicity Hispanic DVT Proximal Distal PE Lobar Segmental or multiple segmental Subsegmental Charlson comorbidity score Previous venous thromboembolism ized < DVT = deep venous thrombosis; LMWH = low-molecular-weight heparin; PE = pulmonary embolism; VTE = venous thromboembolism *Significance between means of normal distributions compared using Student s t-test, categorical variables with chi-square test, and skewed data with Mann-Whitney U-test. Table 2 Charges and Calculated Cost of Medical Care for the First Week After Diagnosis of Venous Thromboembolism Variable Rivaroxaban IQR1 25% IQR3 75% LMWH-Warfarin IQR1 IQR3 p-value Total costs* $1,126 $769 $1,404 $3,058 $1,671 $5,799 <0.001 Total charges $2,854 $2,497 $4,722 $8,080 $4,363 $16,387 <0.001 DVT charges $2,746 $2,409 $4,019 $6,662 $2,017 $12, PE charges $4,435 $2,553 $7,284 $11,375 $7,905 $20,010 <0.001 Mean outpatient pharmacy cash price $384 $626 DVT = deep venous thrombosis; IQR = interquartile range; LMWH = low-molecular-weight heparin; PE = pulmonary embolism; VTE = venous thromboembolism. *Calculated using the Medicare charge to cost ratio index for each hospital. Includes bridging LMWH for warfarin Table 3 Charges and Calculated Cost of Medical Care for Six Months After Diagnosis of Venous Thromboembolism Variable Rivaroxaban IQR1 IQR3 LMWH- Warfarin IQR1 IQR3 p value VTE related costs* $1,446 $1,143 $2,842 $4,006 $2,692 $8,476 <0.001 VTE related charges $4,787 $3,041 $7,596 $11,128 $8,110 $23,390 <0.001 All medical costs $6,443 $3,745 $9,511 $13,086 $8,369 $26,791 <0.001 Charges for patients with DVT $4,094 $3,028 $6,541 $9,366 $4,564 $21, Charges for patients with PE $7,008 $3,037 $10,809 $16,416 $9,635 $24,881 <0.001 Line-item inpatient pharmacy charges $215 $139 $510 $742 $340 $1,756 <0.001 Mean outpatient pharmacy cash price $1,742 $721 Calculated 6-month total hospital system charges $6,529 $4,783 $9,338 $11,849 $8,831 $24,111 <0.001 DVT = deep venous thrombosis; IQR = interquartile range; LMWH = low-molecular-weight heparin; PE = pulmonary embolism; VTE = venous thromboembolism. *Calculated using the Medicare charge to cost ratio index for each hospital. Includes bridging LMWH for warfarin. Total charges from UB-04 forms plus cash cost of medication.

5 800 Kahler et al. COST OF OUTPATIENT TREATMENT and 6-month charges of $7,008 in the rivaroxaban group, compared with $11,375 and $16,416 respectively in the LMWH-warfarin group (initial p < 0.001, 6-month p < 0.001). Patients with DVT treated with rivaroxaban had initial and 6-month median charges of $2,746 and $4,094, respectively. Patients with DVT treated with LMWH-warfarin had significantly higher initial and 6-month charges of $6,662 and $9,365, respectively (initial p < 0.001, 6-month Mann-Whitney U p < 0.001). An analysis of only patients who were never hospitalized throughout their 6 months of treatment reveals a narrowed cost difference. Forty-five rivaroxaban patients were found to have 6-month CCR costs + cash pharmacy outpatient costs of $5,932 (IQR = $4,745 to $8,594) compared to the 15 control patients at $9,016 (IQR = $3,535 to $10,622). These results did not retain significance (p = 0.51), suggesting that hospitalization, not treatment regimen, is the prime driver of cost. Itemized pharmacy charges were determined by summing line-item pharmacy costs for each patient in all available UB-04 forms. Inpatient pharmacy median charges were found to be $215 (IQR = $139 to $510) in the rivaroxaban cases compared to $742 (IQR = $340 to $1,756) for the LMWH-warfarin controls (p < 0.001). The authors also called seven hospital outpatient pharmacies in November 2013 to ascertain cash-only costs of rivaroxaban, enoxaparin, and warfarin. These data are detailed in Table 4 and demonstrate a median cash cost of $1,856 for 6 months of rivaroxaban treatment and $724 for 6 months of enoxaparin/warfarin (p = 0.03). While the median monthly cash cost of warfarin was found to be only $15, the median enoxaparin cost was $688, which inflates the overall cost of warfarin therapy. To present a raw estimate, the total hospital system charges are calculated without any insurance vouchers. Finally, a sensitivity analysis was performed with all medical charges, regardless of source (Table 3). In this sensitivity analysis, all charges, even those incurred through routine preventative maintenance visits, specialty clinic visits, etc., were assessed. The results for all VTE (DVT and PE) were proportional to the primary analysis, with 6-month charges of $6,443 incurred by rivaroxaban patients (IQR = $3,745 to $9,511) compared with $13,086 for the LMWH-warfarin patients (IQR = $8,369 to $26,791; p < 0.001). DISCUSSION In this two-center study, low-risk VTE patients discharged directly from the ED were found to have significantly lower hospital charges, both initially and over 6 months. This finding was maintained when the full cash price of medicines was included, and in a sensitivity analysis that included all medical costs for any reason. The new TSAs allow for immediate anticoagulation without need for complicated bridging procedures and thus allow for discharge of low-risk patients. The primary driver of the reduced cost of medical care for patients treated with the TSA-based outpatient protocol was reduced cost of hospitalization. Our total cost data at 6 months are similar to that estimated by Seaman et al. 16 using a Markov analysis applied to data from the EINSTEIN trial, and the cost difference is similar to that estimated by Lefebvre et al. 17 We found 6-month total costs to be $1,446 for rivaroxaban versus $4,006 for LMWH-warfarin for a sample with a mean age of 50 years. The Markov model estimated costs of Seaman et al. 16 at $3,195 for rivaroxaban versus $6,188 for LMWH-warfarin for a male of 60 years of age treated for 3 to 12 months. To our knowledge, these are the first data to compare the costs of care for patients treated with TSAs to those of patients treated with LMWHwarfarin in a real-world setting. One major concern with TSAs is their cost to the patient, as warfarin is widely available on most $4 Table 4 Out-of-pocket Costs for Rivaroxaban Versus Low molecular weight Heparin and Warfarin at Seven U.S. s Rivaroxaban Standard Treatment Location 21 Days 15 mg BID 30 Days 20 mg qday 95 months Total 7 Days Enoxaparin 80 mg BID 30 Days Warfarin 5 mg qday 6 Months Warfarin 5 mg qday Standard Care Total Wishard $ $ $ $1, $ $6.19 $37.14 $ Memorial Methodist $ $ $1, $1, $ $6.00 $36.00 $ Carolinas $ $ $1, $1, $1, $15.20 $91.20 $1, Medical Center Northwestern $ $ $1, $1, $ $18.69 $ $1, Memorial Mayo Clinic $ $ $1, $1, $ $16.40 $98.40 $ Rochester Pennsylvania $ $ $1, $2, $98.00 $37.64 $ $ University of $ $ $ $1, $ $10.00 $60.00 $ Mississippi Median $ $ $1, $1, $ $15.20 $91.20 $ BID = taken twice a day; q = every.

6 ACADEMIC EMERGENCY MEDICINE July 2015, Vol. 22, No prescription lists. The full cost of the TSAs is rarely paid by patients; for example, out of the 183 patients treated to date in our outpatient clinic as of January 1, 2015, only two have paid full price. There is a possibility of selection bias, as patients who perceive that they may not be able to pay for a new medication may opt for usual care, but we have no evidence that this occurred. In practice, the patient assistance programs alleviated this concern. The majority of our TSA patients paid $5 per prescription, with almost all paying less than $20, attributable to several patient assistance programs. These foundation-based patient-assistance programs are available for apixaban, dabigatran, edoxaban, and rivaroxaban. The manufacturers typically cover any costs over $10 per month for all privately insured patients or provide a subsidy for medication costs for low-income patients after evaluation of a tax return or lack thereof (L. Usegi, personal communication, 2014). Medicaid and most Medicare plans cover these medicines, although some Medicare vendors require prior authorization. It remains possible that if all patients treated with LMWH-warfarin had been immediately discharged, cost differences would have been reduced. As discussed above, when only nonhospitalized patients were analyzed, there was no difference in cost between cases and controls. However, from a practical standpoint, this scenario is unlikely in many EDs. Ensuring patient compliance with LMWH injection and prothrombin time monitoring for warfarin often poses formidable obstacles, especially for low-income patients. The TSA outpatient protocol only requires a limited clinic at 3 weeks to ensure medication compliance, and in the experience of the authors, even this limited visit could have been handled for many patients with a telephone call. A patient assistance program provided a reliable method for our low-income patients to afford their medicines. Although the primary source of reduced costs with the TSA outpatient protocol was reduced hospitalization, the line-item pharmacy charges are also statistically significant when compared between groups. While rivaroxaban s unsubsidized outpatient pharmacy cost was $1,021 more on average over 6 months, the warfarin patients incurred $527 more in itemized hospital pharmacy charges. This is attributable to the high cost of inpatient medications. The major barrier to outpatient treatment with TSAs results from the need to qualify patients for the assistance programs. While most patients qualify for some type of assistance program, these are difficult to address in real time in a busy ED. Our two centers use case managers and social workers during business hours to ease this transition. Nonetheless, the authors receive several phone calls per week from patients who cannot afford their medications until a prior authorization has been approved or a patient assistance program has been initiated. As emergency clinicians become familiar with data supporting the safety of discharge of low-risk PE and DVT patients, we anticipate rapid adoption of similar protocols. Patients in our study have demonstrably lower hospital charges and have been previously shown to have improved outcomes over similar patients who are hospitalized. 5 LIMITATIONS We compared hospital charges taken directly from UB- 04 forms. While the CCR has been criticized as inaccurate for estimating costs on individual patients, it would have been prohibitively difficult to obtain each patient s negotiated payments for only the UB-04 claims incurred during the 6-month period. 18 While these data do represent all available billing data for two hospital systems, it is possible that these patients sought care at additional facilities. However, these two hospital systems cover the majority of patients in their proximity. An additional limitation is that these findings may only apply to low-risk patients as determined by the Hestia criteria. CONCLUSIONS We present a case-control study evaluating the cost of care for patients diagnosed in the ED with low-risk venous thromboembolism. When comparing the patients treated with low-molecular-weight heparin and warfarin to those patients treated as outpatients with the target-specific anticoagulants, the costs of medical care were significantly lower for the target-specific anticoagulant patients. Costs were approximately 59% to 65% lower for the target-specific anticoagulant patients in the first week after diagnosis and 56% to 57% lower after 6 months. References 1. Squizzato A, Donadini MP, Galli L, Dentali F, Aujesky D, Ageno W. Prognostic clinical prediction rules to identify a low-risk pulmonary embolism: a systematic review and meta-analysis. J Thromb Haemost 2012;10: Jimenez D, Aujesky D, Moores L, et al. Simplification of the pulmonary embolism severity index for prognostication in patients with acute symptomatic pulmonary embolism. Arch Intern Med 2010;170: Moores L, Aujesky D, Jimenez D, et al. Pulmonary Embolism Severity Index and troponin testing for the selection of low-risk patients with acute symptomatic pulmonary embolism. J Thromb Haemost 2010;8: Kline JA, Roy PM, Than MP, et al. Derivation and validation of a multivariate model to predict mortality from pulmonary embolism with cancer: The POMPE-C tool. Thromb Res 2012;129:e Zondag W, Hiddinga BI, Crobach MJ, et al. Hestia criteria can discriminate high- from low-risk patients with pulmonary embolism. Eur Respir J 2013;41: den Exter PL, Gomez V, Jimenez D, et al. A clinical prognostic model for the identification of low-risk patients with acute symptomatic pulmonary embolism and active cancer. Chest 2013;143: Zondag W, Mos IC, Creemers-Schild D, et al. Outpatient treatment in patients with acute pulmonary embolism: the Hestia Study. J Thromb Haemost 2011;9:

7 802 Kahler et al. COST OF OUTPATIENT TREATMENT 8. Beam D, Kahler ZP, Kline JA. Immediate discharge and home treatment of low risk venous thromboembolism diagnosed in two U.S. emergency departments with rivaroxaban: a one-year preplanned analysis. Acad Emerg Med 2015;22: Garcia-Fuster MJ, Fabia MJ, Furio E, et al. Should we look for silent pulmonary embolism in patients with deep venous thrombosis? BMC Cardiovasc Disord 2014;14: DeSantis G, Hogan-Schlientz J, Liska G, et al. STA- BLE results: warfarin home monitoring achieves excellent INR control. Am J Manag Care 2014;20: Le Gall C, Jacques E, Medjebeur C, et al. Low molecular weight heparin self-injection training: assessment of feasibility, tolerance and economic analysis in emergency departments. Eur J Emerg Med 2006;13: Harrison L, McGinnis J, Crowther M, Ginsberg J, Hirsh J. Assessment of outpatient treatment of deep-vein thrombosis with low-molecular-weight heparin. Arch Intern Med 1998;158: Chen LM, Jha AK, Guterman S, Ridgway AB, Orav E, Epstein AM. cost of care, quality of care, and readmission rates: penny wise and pound foolish? Arch Intern Med 2010;170: Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (RED- Cap) a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform 2009;42: Kline JA, Shapiro NI, Jones AE, et al. Outcomes and radiation exposure of emergency department patients with chest pain and shortness of breath and ultralow pretest probability: a multicenter study. Ann Emerg Med 2014;63: Seaman CD, Smith KJ, Ragni MV. Cost-effectiveness of rivaroxaban versus warfarin anticoagulation for the prevention of recurrent venous thromboembolism: a U.S. perspective. Thromb Res 2013;132: Lefebvre P, Coleman CI, Bookhart BK, et al. Costeffectiveness of rivaroxaban compared with enoxaparin plus a vitamin K antagonist for the treatment of venous thromboembolism. J Med Econ 2014;17: Shwartz M, Young DW, Siegrist R. The ratio of costs to charges: how good a basis for estimating costs? Inquiry 1995;32: