Pulmonology

Cost-effectiveness of managing low-risk pulmonary embolism patients without hospitalization. The low-risk pulmonary embolism prospective management study

a b s t r a c t

Objective: Evaluate the cost-effectiveness and difference in length-of-stay when patients in the ED diagnosed with low-risk pulmonary embolism (PE) are managed with early discharge or observation.

Methods: Single cohort prospective management study from January 2013 to October 2016 of patients with PE diagnosed in the ED and evaluated for a primary composite endpoint of mortality, recurrent venous thromboem- bolism, and/or major bleeding event at 90 days. Low-risk patients had a PE Severity Index score < 86, no evidence of proximal deep vein thrombosis on venous compression ultrasonography of both lower extremities, and no ev- idence of right heart strain on echocardiography. Patients were managed either in the ED or in the hospital on Observation status. Primary outcomes were total length of stay, total encounter costs, and 30-day costs.

Results: 213 patients were enrolled. 13 were excluded per the study protocol. Of the remaining 200, 122 were managed with emergency department observation (EDO) and 78 with hospital observation (HO). One patient managed with EDO met the Composite outcome due to a major bleeding event on day 61. The mean length of stay for EDO was 793.4 min (SD -169.7, 95% CI:762-823) and for HO was 1170 (SD -211.4, 95% CI:1122-1218)

with a difference of 376.8 (95% CI: 430-323, p < 0.0001). Total encounter mean costs for EDO were $1982.95 and $2759.59 for HO, with a difference of $776.64 (95% CI: 972-480, p > 0.0001). 30-day total mean costs for EDO were $2864.14 and $3441.52 for HO, with a difference of $577.38 (95% CI: -1372-217, p = 0.15).

Conclusions: Patients with low-risk PE managed with ED-based observation have a shorter length of stay and lower total encounter costs than patients managed with Hospital-based observation.

Introduction

Background: venous thromboembolism , including Deep vein thrombosis and pulmonary embolism (PE), is a prevalent disease with significant associated mortality; as many as 900,000 patients per

Abbreviations: Hospital Observation, HO; ED Observation, EDO; Venous Thromboembolism, VTE; Pulmonary Embolism, PE; Low-risk Pulmonary Embolism, LoPE.

* Corresponding author.

E-mail address: [email protected] (J.R. Bledsoe).

year may be affected by venous thromboembolism annually in the US alone, and approximately 10-30% of PE patients die within one month of diagnosis [1-3] . However, VTE encompasses a spectrum of disease se- verity, not all of which is immediately life-threatening. In the past, PE has been managed with hospital admission likely on account of its sig- nificant associated mortality. The Pulmonary Embolism Severity Index (PESI) score is a prospectively validated tool which allows physicians to risk stratify patients with acute PE [4,5]. The American College of Chest physicians (CHEST) and European Society of Cardiology (ESC) guidelines for Venous thromboembolic disease (VTE) suggest that select

https://doi.org/10.1016/j.ajem.2020.12.042

low-risk patients with PE can be safely treated with outpatient man- agement [6-8]. Evidence for the safety of this approach has been lim- ited, with two prospective studies demonstrating noninferiority of home treatment with both enoxaparin/heparin and with rivaroxaban [9,10]. The recently published LoPE study further prospectively evalu- ated the safety of this approach in a cohort of patients with Low-risk Pulmonary Embolism (LoPE) [11]. This study found that treatment of select patients with LoPE is both safe and acceptable to patients; pa- tients with LoPE observed in the emergency department demonstrated 0% 90-day all-cause mortality, 0% recurrent DVT/PE at 90 days, a low rate of return to the emergency department (9.5%, 95% CI, 5%-14%), and only one patient suffered an adverse event, a major bleed on day 61 due to a traumatic injury.

The average PE Hospitalization costs approximately $8764 [12]. Ad- missions for PE represent more than $2 billion of healthcare expendi- ture annually in the United States [1]. The cost-effectiveness of an outpatient management strategy has not been evaluated. We hypothe- sized that outpatient management or early discharge would be cost ef- fective and result in significantly reduced Healthcare costs compared to a traditional and conservative approach including hospital admission. This study analyzes the same patient cohort as the parent LoPE study but investigates the cost effectiveness of the ED observation strategy that the LoPE study showed to be safe [11].

Methods
1. Study design and setting

This was an a priori planned cost and length of stay analysis of a single-arm prospective cohort management study performed at five centers in the Utah. Four centers are community suburban hospitals each with an annual ED census of 20,000 to 50,000 visits and one center is a tertiary referral center with an American College of Surgeons level 1 trauma center designation with an annual ED census of >90,000 visits.

1. Selection of participants

Consecutive patients diagnosed with acute PE by CT pulmonary arte- riography (CTPA) or high-probability Ventilation/perfusion scan and with a PESI score < 86 (corresponding to class I-II according to the PESI classification system [4,13]) as calculated by the treating emer- gency physician were eligible for enrollment. Enrollment screening was conducted as previously described [11]. Exclusion criteria included: children (< 18 years of age); prisoners; current pregnancy; high risk PE (patients with hypotension with sustained systolic BP < 95 mmHg or hypoxia with sustained room air oxygen saturation < 90%); hepatic dys- function; renal insufficiency; contraindications to enoxaparin, warfarin, rivaroxaban, or apixaban; present atrial or ventricular dysrhythmias; evidence of right heart strain on echocardiogram; or DVT proximal to the popliteal vein. Patients with uncertainty regarding health literacy, with additional conditions requiring hospital admission, or with a per- ceived social situation that would disallow reliable follow-up were also excluded.

1. Interventions

Qualified patients provided written informed consent and were ob- served for 12 to 24 h either in a standard ED B.D.(EDO) or in the hospital under outpatient observation status (HO) and underwent formal trans- thoracic echocardiography and CUS of both lower extremities and CUS of any symptomatic upper extremities (standard practices at the partic- ipating facilities). Patients were allocated to the EDO or HO group based on patient and clinician preference. During the observation period, pa- tients in either group who developed one of the exclusion criteria were excluded from further study procedures and analysis and were treated with routine care.

Qualified patients were treated with therapeutic anticoagulation using a Food and Drug Administration-approved medication (enoxaparin, enoxaparin transitioned to warfarin, rivaroxaban, or apixaban dosed according to VTE treatment standards) with the regi- men chosen by the treating emergency physician. All patients had con- sultation with a physician specializing in thrombosis and had outpatient follow-up arranged prior to discharge as previously described [11].

1. Outcomes

The primary cost outcomes for this study were defined as total en- counter costs, 30-day costs, and total length of stay. Median direct inci- dent encounter costs per patient were obtained through the Intermountain Healthcare cost master accounting system. Time- derived activity-based costing (TDABC) is a model that contains variable and Total costs for a standardized set of codes [14]. Hospitals conduct an- nual or biannual costing studies to update the variable and total cost for these codes. The cost data was validated for the present cohort by the authors. No patients were excluded based upon cost anomalies.

Incident costs vary by hospital due to many factors including differ- ent operations overhead and staffing acuity. In order to control for these variations, a standardized cost for each charge code in the TDABC sys- tem was derived by summing the cost and count of each charge code and then dividing the total cost by the count. All costs were pegged to 2015 US dollars (USD). No consumer-facing inflationary indices were used for either adjustment or standardization.

The primary clinical outcome of the LoPE study was defined as a composite of symptomatic recurrent VTE, major bleeding, or mortality within 90 days of enrollment. Outcomes and mortality data were evalu- ated and reported as previously described [11].

1. Analysis

All tests were computed using Student’s t-test and corresponding methods were used to create the confidence intervals on the means and differences of means. Both sample sizes were large enough that the results will be approximately correct even in the cases where the data indicates a non-normal population. We did not correct for multiple comparisons since each outcome was of specific interest.

This study was registered with clinicaltrials.gov and was approved by the Intermountain Healthcare institutional review board. ClinicalTrials.gov Identifier: NCT02355548

Results
1. Patients

1003 consecutive patients with a new diagnosis of acute pulmonary embolism were evaluated and assessed for eligibility. 790 were ex- cluded for PESI score > 85, 419 for another exclusion criterion, 123

Table 1

Demographic characteristics of ED Observation and Hospital observation groups

	ED Observation	Hospital Observation
	N (%)	N (%)
Female	71 (58.2)	36 (47.5)
Male	51 (41.8)	41 (52.5)
African American	1 (0.8)	1 (1.2)
Asian	0 (0.0)	0 (0.0)
Caucasian	111 (90.9)	71 (91.0)
Hispanic	7 (5.7)	5 (6.4)
Native American	0 (0.0)	0 (0.0)
Other	3 (2.4)	1 (1.2)
Tobacco use	14 (11.5)	9 (11.5)
Alcohol Use	16 (13.1)	14 (17.9)

declined to provide informed consent, and 30 potentially eligible pa- tients could not be screened. 213 patients met initial inclusion criteria and were enrolled in the study. 13 patients were subsequently excluded per the study protocol based upon echocardiography, compression ul- trasound, or for developing an exclusion criterion during the period of observation. Of the remaining 200 evaluable patients, 122 were treated with EDO and 78 with HO. One patient treated with EDO met the com- posite adverse clinical outcome due to a major bleeding event on day 61. Demographic characteristics of the EDO and HO groups were not sig- nificantly different (Table 1). The clinical characteristics of the EDO and HO groups also lacked statistical differences aside from height in centi- meters (Table 2); the HO group exhibited a mean height difference of

+2.7 cm in comparison to the EDO group. The mean PESI scores for EDO (53.9, 95% CI [51.0-56.7]) and HO (54.8, 95% CI [51.4-58.3]) did

not differ significantly.

1. Primary outcomes

The mean length of stay for EDO was 793.4 min (SD-169.7, 95% CI [762-823]) and for HO was 1170 (SD- 211.4, 95% C [1122-1218]),

with a mean difference of 376.8 min (95% CI [430-323], p < 0.0001).

The total encounter mean costs for EDO were $1982.95 and for HO were $2759.59, with a mean difference of $776.64 (95% CI [$972-

$480], p < 0.0001). The 30-day total mean costs for EDO were

$2864.14 and for HO were $3441.52, with a difference of $-577.38 (95% CI [-$1372-$217], p = 0.15).

Discussion

This multi-center prospective trial of patients with low-risk PE sup- ports the hypothesis that ED observation rather than hospital-based ob- servation results in significantly lower total encounter costs. This finding compares favorably with existing data supporting the cost- effectiveness of outpatient management of deep vein thrombosis with- out hospital admission [15,16]. Our findings support the idea that early outpatient treatment of LoPE is not only safe and satisfying to patients, but also significantly reduces the cost burden that this disease entity poses to the healthcare system. Given the robust evaluation undertaken on both the EDO and HO groups, it was conceivable that these evalua- tions would be more cost-effective if performed in hospital. These find- ings support the cost-effectiveness of an ED observation strategy, a departure from historically typical care.

The prospective cohort methodology utilized in this study is similar to other robust studies of cost-effectiveness related to the treatment of VTE [17], though many other such analyses have relied on modeling rather than direct study [17-19]. To the authors’ knowledge, this is the first study to prospectively evaluate the cost effectiveness of different observation strategies in low-risk PE.

As a prospective cohort study, this study lacks a control arm of “usual care” against which to compare length of stay and cost results. A

Table 2

Clinical characteristics of ED observation and Hospital observation groups

ED Observation Hospital Observation

	Mean	95% CI	Mean	95% CI
Age (yrs)	44.0	[41.4-46.5]	45.0	[41.8-48.2]
Height (cm)	172.7	[170.9-174.6]	175.4	[172.8-178.1]
Weight (kg)	88.1	[84.4-91.8]	91.6	[86.1-97.1]
RBC (K/mcL)	4.7	[4.7-4.8]	4.7	[4.6-4.8]
WBC (K/mcL)	9.2	[8.7-9.7]	8.9	[8.4-9.4]
Hbg (g/dL)	14.2	[13.9-14.5]	14.2	[13.9-14.5]
D-Dimer (ng/mL)	1292.8	[508.3-2077.3]	1646.0	[210.1-3502.1]
PT-INR	1.1	[0.9-1.4]	1.1	[NA]
Rev. Geneva Score	4.9	[4.3-5.5]	4.3	[3.6-5.0]
PESI Score	53.9	[51.0-56.7]	54.8	[51.4-58.3]

retrospective cohort control arm was planned at the time that this study was conducted, however equipoise was lost due to approval of di- rect oral anticoagulants for VTE in October 2010 during the study pe- riod. Due to protocol mandated bilateral lower extremity venous compression ultrasonography and echocardiography, this study’s Visit durations and costs associated with observation may be increased com- pared to usual care, but as this was mandated for both groups, this should not affect intergroup comparison.

The applicability of this study to the usual care of emergency depart- ment patients is enhanced by the study’s setting, and by its use of widely-available diagnostic studies to risk-stratify patients. Patients were recruited from four community centers and a tertiary care center in a large metropolitan area, and adverse outcomes in this study group were similar to previously reported rates [10,11]. We used the Pulmo- nary Embolism Severity Index Score (PESI) to identify low-risk patients, which is a validated clinical tool easily applied by emergency depart- ment clinicians [4,5]. Additionally, diagnosis and risk stratification of PE patients in this study was conducted according to standard practice at participating facilities, aligning this study with usual care. Exclusion criteria for this study are likewise readily applicable using typical emer- gency department laboratory evaluations and history from the patient. The Revised Geneva Score has been shown to identify patients who are not low-risk by PESI score but who may be safely discharged to out- patient management [20]. This study utilized the PESI score as it was the convention at treating facilities during the study period. It may be that use of the Revised Geneva Score could result in the identification of more patients safely eligible for an EDO strategy as compared to the

PESI score, but more research is needed to confirm this hypothesis.

This study allocated patients to EDO and HO based on patient and cli- nician preference. Though measured variables including validated clin- ical predictors of VTE adverse event risk such as PESI score and Revised Geneva Score [20-22] did not differ significantly between the EDO and HO groups, unmeasured variables may have biased towards hospitaliza- tion of a more clinically concerning cohort due to clinician concern. This would bias towards greater initial and 30-day costs in the HO group.

Mean 30-day costs were lower in the EDO group compared to the HO group, but this difference did not reach statistical significance. The lower 30-day costs in the EDO group do align with the encounter costs noted for the index visit. Our analysis did not exclude patients on the basis of cost anomalies. As the parent study was powered to de- tect a difference in Major adverse events between the EDO and HO group, this study was likely underpowered to account for the variance of relatively strong effects introduced by these cost anomalies.

Patients enrolled in this study were also ensured consultation with a thrombosis physician during their period of observation and were scheduled for follow-up prior to discharge. This careful selection of pa- tients appropriate for EDO is in accord with ACP guidelines but limits the applicability of this study; consultation with a thrombosis physician may not be widely available. The safety and cost-effectiveness of the EDO group was likely improved by interventions directed by the throm- bosis physician, such as safer selection of anticoagulants and dosing these medications to maximize efficacy and safety.

Conclusions

ED observation of carefully selected patients diagnosed with low- risk PE results in significantly shorter length of stay and costs associ- ated with the initial hospital encounter. Further research is needed to demonstrate cost savings associated with this strategy beyond the index visit.

Grant

This study was funded by the Intermountain Research and Medical Foundation through Grant No. 641awarded to J.R.B.

Meeting

ACEP 2019: A Cost-Effectiveness and Length-of-Stay Analysis of Early Discharge of Emergency Department Patients with Low-Risk Pul- monary Embolism.

Author contribution

J.R.B. takes responsibility for the paper as a whole. J.R.B., T.M., T.A., S.M.S., and S.C.W. conceived the study, designed the trial, and obtained research funding. J.R.B. and V.A. supervised the conduct of the trial and data collection. J.R.B., V.A., S.M.S., S.C.W., and R.P. undertook recruitment of participating patients and managed the data, including quality con- trol. T.A., L.D., and J.L. provided additional data acquisition. B.D.H. pro- vided statistical advice on study design, and G.S. analyzed the data.

B.D.H. chaired the data safety monitoring board. P.F. drafted the manu- script, and all authors contributed substantially to its revision.

Funding

This study was funded by the Intermountain Research and Medical Foundation (Grant No. 641).

Conflicts of interest

J.R.B. has received honoraria from AMAG for participating on an ex- pert panel. J.R.B, S. M. S. and S. C. W. have grant support from Bristol- Meyer-Squibb paid to IHC. None declared (V.A., R.P., T.A., B.D.H., L.D.,

J.L., G.S., T.M., C.G.E., P.F.).

CRediT authorship contribution statement

Joseph R. Bledsoe: Conceptualization, Funding acquisition, Meth- odology, Supervision, Investigation, Writing – review & editing. Scott

C. Woller: Conceptualization, Funding acquisition, Methodology, Writ- ing – review & editing. Scott M. Stevens: Conceptualization, Funding acquisition, Methodology, Supervision. Valerie Aston: Supervision, In- vestigation. Rich Patten: Supervision. Todd Allen: Conceptualization, Funding acquisition, Methodology, Supervision. Benjamin D. Horne: Methodology. Lydia Dong: Investigation. James Lloyd: Investigation. Greg Snow: Formal analysis. Troy Madsen: Conceptualization, Funding acquisition, Methodology. Patrick Fink: Writing – original draft, Writing – review & editing. C. Gregory Elliott: Conceptualiza- tion, Formal analysis, Writing – review & editing.

References

Michele G, Beckman WCH, Critchley Sara E, Ortel Thomas L. Venous thromboembo- lism: a Public health concern. Am J Prev Med. 2010;38(S):495-501.
Goldhaber SZ, Bounameaux H. Pulmonary embolism and deep vein thrombosis. Lan-

cet. 2012;379(9828):1835-46.

AJEM

Cost-effectiveness of managing low-risk pulmonary embolism patients without hospitalization. The low-risk pulmonary embolism prospective management study

a b s t r a c t

Leave a Reply Cancel reply