Low molecular weight heparins (LMWHs), which are fragments of commercial grade UFH and include enoxaparin, may offer an alternative to UFH for the treatment of DVT. LMWH is advantageous in that its use does not require laboratory monitoring and it has a more predictable anticoagulant effect than UFH, a longer plasma half-life requiring less frequent dosing, and better bioavailability when administered subcutaneously.² Furthermore, although LMWH and UFH are similar with regard to the frequency of bleeding complications, the incidence of heparin-induced thrombocytopenia (HIT) appears to be much lower with LMWH.³ These characteristics and current research^4-6 suggest that LMWH, including enoxaparin, may offer a viable alternative therapy for DVT and a candidate for home (outpatient) administration that would obviate the need for hospital admission. This could result in lower costs and a less disruptive treatment for the patient.

Patients residing in long-term care facilities (LTCFs) represent a group that could potentially benefit from the advantages of LMWH for the treatment of DVT. Current treatment of DVT requires patients to be removed from familiar surroundings and be hospitalized. This can be a disruptive and disorienting experience, especially for this population. Frequent monitoring (e.g., blood draws) can add to the discomfort, not to mention the added risk of anemia.⁷ Hospitalization also puts the patient at risk for hospital-acquired infections.⁸ Use of LMWHs may allow patients to remain in the skilled ward of the LTCF for DVT treatment.

A cost-effectiveness decision tree was created to compare enoxaparin in the skilled ward of a LTCF with UFH in the hospital for the treatment of proximal DVT. Given adverse event probabilities based on previous studies, this model identified the resources used to support treatment and estimated the overall treatment costs based on prior research,^9,10 the current rates for the skilled ward of an LTCF, and the average wholesale price for enoxaparin. The cost-effectiveness decision tree was rolled back in order to demonstrate the most cost-effective treatment pathway (Figure 1). Sensitivity analyses were performed to test for uncertainties in estimates in the model in order to test the robustness of the comparison. Threshold values were calculated to demonstrate at which point both treatment alternatives had the same expected value, i.e., the point at which both options became equally cost-effective. Finally, conclusions were drawn concerning the impact of enoxaparin used in the skilled ward of a LTCF for the treatment of proximal DVT.

The purpose of this article is to consider from the perspective of a managed care organization the cost-effectiveness of using UFH in the hospital versus enoxaparin in the skilled ward of a LTCF for the treatment of proximal DVT. Though one study has compared the use of LMWH at home to UFH in the hospital for the treatment of proximal DVT,⁶ to date there has been no comprehensive assessment of the cost-effectiveness of the use of these alternative agents in the treatment of proximal DVT that compares the hospital versus the LTCF setting.

Methods

Table 1. Drug Regimens Used in Clinical Trials
Study	Study Design	Medication	Regimen
Huet et al.4	Single group design; 12 days	Enoxaparin (n = 36)	1 mg/kg every 12 hours
Simonneau et al.5	Multicenter randomized clinical trial; 10 days	Enoxaparin (n = 67) Unfractionated Heparin (n = 67)	1 mg/kg every 12 hours   500 units/kg per 24 hours continuous infusion, then adjusted dose
Levine et al.6	Randomized clinical trial; 3 months	Enoxaparin (n = 253) Unfractionated Heparin (n = 247)	1 mg/kg every 12 hours   5,000 units initial dose, then adjusted dose continuous infusion

Probabilities of Clinical Events
The probabilities of recurrent or new DVT, pulmonary embolism (PE), major hemorrhage, minor hemorrhage, and heparin-induced thrombocytopenia (HIT) were calculated based on the adverse events reported in clinical trials that focused on treatment of proximal DVT with enoxaparin or UFH^4-6 (Table 2). Costs for treating these adverse events (except for HIT, which responds to discontinuation of therapy11) were based on prior research^9,10 and adjusted to 1996 U.S. dollars. It was assumed that patients experiencing any of these adverse events (except HIT) would be treated in the hospital. The total cost for a patient experiencing an adverse event in the cost-effectiveness decision tree included the cost for the five days of DVT therapy (i.e., with either enoxaparin in the skilled ward of a LTCF or UFH in the hospital, respectively) added to the cost to treat that adverse event in the hospital. By adding the cost of adverse events to the cost of five full days of DVT therapy, the most costly possible outcome was explored for each treatment alternative.

Table 2. Adverse Event Frequencies in Clinical Trials
Study	Recurrent or New DVT	Pulmonary Embolism	Major Hemorrhage	Minor Hemorrhage	Heparin-Induced Thrombocytopenia
Adverse events associated with enoxaparin
Huet et al.4	0/30 (0.0%)	0/34 (0.0%)	2/36 (5.5%)	0/36 (0.0%)	0/36 (0.0%)
Simonneau et al.5	1/67 (1.4%)	0/67 (0.0%)	0/67 (0.0%)	4/67 (5.9%)	0/67 (0.0%)
Levine et al.6	12/247 (4.8%)	2/247 (0.8%)	5/247 (2.0%)	6/247 (2.4%)	5/247 (2.0%)
TOTAL	13/348 (3.7%)	2/348 (0.5%)	7/350 (2.0%)	10/350 (2.8%)	5/350 (1.4%)
POOLED %a	3.7%	0.6%	2.0%	2.8%	1.4%
Adverse events associated with unfractionated heparin
Simonneau et al.5	5/67 (7.4%)	2/67 (2.9%)	0/67 (0.0%)	0/67 (0.0%)	1/67 (1.4%)
Levine et al.6	15/253 (5.9%)	2/253 (0.7%)	3/253 (1.1%)	6/253 (2.3%)	3/253 (1.1%)
TOTAL	20/320 (6.2%)	4/320 (1.2%)	3/320 (0.9%)	6/320 (1.8%)	4/320 (1.2%)
POOLED %a	6.2%	1.2%	0.9%	1.8%	1.2%
aPooled % = ¸(ni x pi)/N; (n = total number of subjects experiencing adverse events in clinical trial i; p = probability of adverse events in clinical trial i; N = total number of patients in all clinical trials) DVT = deep vein thrombosis NA = not applicable

Resource Utilization and Unit Cost
For both alternatives, enoxaparin in the LTCF and UFH in the hospital, the costs considered were those associated only with the treatment of proximal DVT. Costs that were not considered included the cost of clinical tests to diagnose DVT, the costs of drug administration by nurses and ancillary supplies, and the cost of anticoagulation therapy (e.g., warfarin) following the discontinuation of UFH or enoxaparin therapy. These expenses were assumed to be equal for both treatments. Since compression ultrasonography is the preferred diagnostic technique for DVT, the costs for this procedure would be equivalent for both alternatives in the decision model (Figure 1).¹² A resource utilization chart (Table 3) describes all the costs that were considered and total costs for each treatment, including the cost to treat adverse events.

Table 3. Resource Utilization Chart for the Management of Proximal DVT
Nature of Expense	Itemized Cost ($)	Total Cost ($)
Hospitala
Physician Monitoring, 3 visits	167.00
Hospitalization	2,245.00
Unfractionated heparin, 30,000 units I.V./day ($9.00*day)	45.00
PTT, 1 test/day ($9.00*5 tests)	45.00
Complete blood cell count with platelet count, 2 tests	14.00
TOTAL		2,616.00
Long-term care facilityb
Skilled nursing ward, 5 days ($189.41/day)	947.00
Enoxaparin, 1 mg/kg every 12 hours, 5 days ($78.40/day)	392.00
Complete blood cell count with platelet count, 2 tests	34.00
TOTAL		1,373.00
Recurrent or new DVT (5-day successful therapy)b
Hospital TOTAL		5,032.00
Long-term care facility TOTAL		2,746.00
Confirmed pulmonary embolism (treated in hospital)a
Physician assessment, 1 visit	833.00
Hospitalization (15 days)	6,733.00
Ventilation-perfusion scan, 1 test	432.00
Chest radiograph, 1 exposure	71.00
Unfractionated heparin, 42,000 units I.V./day (15 days)	188.00
Warfarin, 5 mg/day (95 days)	59.00
PT and PTT, 21 tests	190.00
Complete blood cell count with platelet count, 2 tests	14.00
TOTAL		8,532.00
Major hemorrhage (treated in hospital)c
Direct cost	2,449.00
Hospitalization 3.2 days	900.00
TOTAL		3,349.00/event
Minor hemorrhage (treated in hospital)c
TOTAL		227.00/event
a1993 U.S. dollars adjusted to 1996 b1996 U.S. dollars c1992 U.S. dollars adjusted to 1996 DVT = deep vein thrombosis, PPT = partial thromboplastin time, PT = prothrombin times

Results

Sensitivity Analyses
Because the three clinical trials used to generate data for the cost-effectiveness decision tree provided a range of probabilities, it was important to evaluate the influence of variations in key costs on the results. These variations were made in order to determine when UFH in the hospital would become more cost-effective than enoxaparin in the LTCF. Thus, a sensitivity analysis was undertaken by: 1) increasing the duration of therapy with enoxaparin in the LTCF to eight, nine, or 10 days (instead of five days); 2) increasing the daily cost for the skilled ward of the LTCF to $300, $400, or $500 (instead of $189.41); and 3) decreasing the daily cost in the hospital setting to $400, $300, or $200 (instead of $449). These changes were incorporated one at a time; all other factors were kept constant. The results of the sensitivity analyses are presented in Table 4.

Table 4. Sensitivity Analyses of Enoxaparin in the Skilled Ward of LTCF versus UFH in Hospital (comparison of cost-effectiveness)
Alteration in Model	Cost-Effectiveness Ratios
	Enoxaparin in LTCF ($)	UFH in hospital ($)
8-day extended skilled LTCF care with enoxaparina	2,619	3,124
9-day extended skilled LTCF care with enoxaparina	2,926	3,124
10-day extended skilled LTCF care with enoxaparina	3,231	3,124
$300/day skilled LTCF care with enoxaparin	2,334	3,124
$400/day skilled LTCF care with enoxaparin	2,905	3,124
$500/day skilled LTCF care with enoxaparin	3,479	3,124
$400/day hospital care with UFH	1,703	2,835
$300/day hospital care with UFH	1,703	2,244
$200/day hospital care with UFH	1,703	1,653
LTCF = long term care facility, UFH = unfractionated heparin aHospital therapy with UFH left at 5-days

The model was robust to certain changes in the length of therapy and to certain changes in daily room cost. Enoxaparin in the LTCF setting remained the more cost-effective option for treatment of proximal DVT when the length of therapy was increased to eight and nine days, leaving all other factors the same. At 10 days, however, enoxaparin in the LTCF was no longer more cost-effective. In addition, leaving all other factors the same, enoxaparin in the LTCF setting remained the more cost-effective option when the cost per day to stay in the LTCF was increased to $300 and $400. When the cost per day was increased to $500, however, this option was no longer more cost-effective. The threshold value above which enoxaparin in the LTCF ceases to be more cost-effective occurs when the room cost in the LTCF is increased to $439 per day. Finally, leaving all other factors the same, UFH in the hospital became the more cost-effective option when the hospital room cost per day was decreased to $200 in the sensitivity analysis. The threshold value below which UFH in the hospital first becomes more cost-effective occurs when the room cost for the hospital is decreased to $208 per day. To summarize, the sensitivity analysis demonstrated that enoxaparin in the LTCF setting remained the more cost-effective option when certain alterations were made to the duration of therapy and to the cost to stay within either facility. This reveals that the model is robust with respect to certain changes in duration of therapy and room cost per day. The threshold values reveal when UFH in the hospital would become the more cost-effective alternative.

Discussion

In this study, sensitivity analyses were undertaken based on two major cost drivers: duration of therapy and cost per day. Enoxaparin in the LTCF remained the more cost-effective alternative until length of therapy was increased to 10 days (leaving therapy in the hospital with UFH at five days) and when a threshold value of $439 was exceeded for daily room cost in the LTCF. Finally, UFH in the hospital became more cost-effective when the daily room cost in the hospital was decreased below a threshold value of $208, leaving all other factors the same.

Though treating proximal DVTs with enoxaparin, even in the outpatient setting, has been explored, the use of enoxaparin to treat DVTs within a LTCF has, to our knowledge, not been explored. Recently, Dedden et al.¹³ described a pharmacy-managed program for home treatment of uncomplicated acute proximal DVT with enoxaparin that was developed and implemented at a health maintenance organization. The program included 55 patients who were taught to self-administer enoxaparin at home. The patients received concurrent enoxaparin and warfarin therapy for at least five days, then continued on warfarin alone. The authors report that there were no complications or hospital admissions due to bleeding. A logical extension of this program is applying the same idea to the LTCF setting. The major difference is that nurses would be administering the enoxaparin in a LTCF instead of the patients themselves. In an age where cost containment in the health care system is the bottom line, this study demonstrates that using enoxaparin in the LTCF as an alternative to UFH in the hospital for the treatment of proximal DVT presents a possibility for reducing health care cost and inconvenience.

The major limitation identified in this study was the potential for variability in adverse events and costs. Data on the incidence of adverse events with either enoxaparin or UFH were based on three clinical trials. There is a possibility that the overall incidence of these adverse events could vary from those reported in this study if more clinical trials were performed. With regard to costs, it is highly likely that costs will vary from one institution to another. The sensitivity analyses performed in this study were done in order to incorporate a degree of this variability and to demonstrate the influence of altering costs on the model.

Conclusion

Copyright © 2000, American Society of Consultant Pharmacists, Inc. All rights reserved.