US budget impact of increased payer adoption of the Flexitouch advanced pneumatic compression device in lymphedema patients with advanced chronic venous insufficiency and multiple infections

Abstract

Aims: To assess the budget impact to a US commercial health plan of providing access to the Flexitouch (FLX) advanced pneumatic compression device (Tactile Medical) to lymphedema (LE) patients with either comorbid chronic venous insufficiency (CVI) or frequent infections.

Methods: Budget impact was calculated over 2 years for a hypothetical US payer with 10-million commercial members. Model inputs were derived from published sources and from a case-matched analysis of Blue Health Intelligence (BHI) claims data for the years 2012–2016. To calculate the budget impact, the Status Quo budget (i.e. total cost for LE and sequelae-related medical treatment) was compared to the budget under each of three Alternate Payer Policy scenarios which assumed that a sub-set of patients was redistributed from their initial treatment groups to a group that received FLX. Model outputs included cumulative payer costs, net budget impact, and breakeven point. Sensitivity analyses were performed to assess the impact of model inputs on results.

Results: Increasing access to FLX yielded a favorable budget impact in every scenario. For LE patients with comorbid CVI, the three alternate scenarios resulted in cumulative 2-year budget impacts of –$52,841, –$173,317, and –$375,601, respectively. For LE patients with comorbid frequent infections, the three alternate scenarios resulted in cumulative 2-year budget impacts of –$192,729, –$259,339, and –$613,179, respectively.

Limitations: Use of claims data assumes accurate coding and does not allow one to control for disease severity or treatment adherence. Also, the distribution of patients between treatment arms was determined using claims data from a specific payer organization, and could differ for health plans with different coverage policies.

Conclusions: While previous studies have illustrated cost savings with adoption of FLX, US commercial health plans may also achieve tangible cost savings by expanding access to FLX for LE patients with comorbid CVI and multiple infections.

Keywords:

• Lymphedema
• budget impact
• pneumatic compression
• advanced pneumatic compression device (APCD)
• Flexitouch
• cellulitis
• chronic venous insufficiency (CVI)

JEL Classification Codes:

• I10
• I11

Introduction

Lymphedema (LE) is a costly chronic condition caused by lymphatic dysfunction that results in a high protein edema, typically in an extremity. LE can occur as a result of congenital factors (primary), but it is more often a result of an alteration of normal lymphatic function (secondary), which in high-income countries is most commonly due to chronic venous insufficiency (CVI)¹ and cancer treatments involving lymph node removal or irradiation². LE is associated with both physical and psychosocial problems for patients. The swelling and subsequent fibrosis, resulting from accumulation of fluid in the dermal and subcutaneous tissues, can cause pain and disfigurement, and impair mobility and function³. LE produces a diminished immune response, which can result in painful and costly infections⁴. Research into the costs of LE care to date has focused almost exclusively on cancer patients, most commonly post-breast cancer. These studies have demonstrated that medical costs for breast cancer patients who develop LE are 35–85% higher and over $20,000 greater than for those who do not^5,6. There remains a dearth of information on the costs associated with treating non-cancer related LE.

The clinical literature suggests that certain sub-populations have increased risk of LE morbidity⁷. Patients with phlebolymphedema, a vascular condition resulting from the combined effects of LE and chronic venous insufficiency (CVI), are known to be more difficult to manage than those with LE alone, due to the involvement of the two vascular systems⁸. These patients often experience significant edema, lipodermatosclerosis, and skin ulceration⁷. Infection, typically erysipelas, cellulitis, or lymphangitis, is also a common complication of LE⁹. Patients who experience one episode of cellulitis have a relatively high likelihood of recurrence¹⁰.

Although LE is a known cause of physical and psychosocial morbidity, treatment for chronic care patients can be challenging, and effective measures are seldom prescribed^3,11,12. In the absence of a permanent cure, LE patients must rely on long-term treatment strategies that improve symptoms and reduce complications. Initial conservative treatment of LE focuses on reduction of swelling. This phase of treatment may include manual lymphatic drainage (MLD), multilayer bandaging, compression garments, decongestive exercises, and/or elevation of the limb. For most patients, a subsequent maintenance phase of treatment is employed, usually in the home, which some studies suggest should include the use of a pneumatic compression device (PCD) to maintain edema reduction^13,14. There are two types of PCDs commonly employed to treat LE: segmented PCDs without calibrated gradient pressure (also known as simple PCDs or SPCDs, HCPCS code E0651), and segmented PCDs with calibrated gradient pressure (also known as programmable multi-chamber PCDs, advanced PCDs, or APCDs, HCPCS code E0652). In the US, Medicare and most commercial payers restrict access to any type of PCD to patients with severe, chronic LE whose condition is not responsive to some or all of the initial conservative therapy treatments listed above^15–20. Private payers typically enforce further restrictions on APCDs, requiring providers to either identify unique patient characteristics that would preclude the use of an SPCD or to demonstrate failure to respond to treatment with such a device^15–20.

APCDs have been found to provide significant clinical benefit, including reduced edema⁹ and lower incidence of cellulitis²¹. In addition, side-effects of PCD use are rare²² and newer APCDs may reduce the risk of complications, such as genital edema, observed with older devices²³. A specific APCD, Flexitouch (Tactile Medical), was chosen for evaluation in this study due to its robust efficacy data^9,21, including proof of its ability to stimulate the lymphatic system, which has not been proven for any other PCDs¹⁴. The efficacy of Flexitouch has been hypothesized to be related to its design, which includes up to 32 inflatable chambers (compared to eight chambers with other APCDs) and 18 treatment program options, as well as its unique compression therapy profile, which closely simulates the optimally-performed MLD technique (compared to the higher pressure “squeeze-and-hold” technique employed by other devices)²⁴. In this study, we sought to evaluate the budget impact of treating specific LE patient populations with Flexitouch, instead of with other APCDs, SPCDs, or conservative therapy alone in a representative, privately insured US population.

Study design and methods

Model overview

An economic model was developed to evaluate the plan-level budget impact of shifts in patient distributions between LE treatment modalities under alternate US payer coverage policy scenarios. Outputs included cumulative costs, net budget impact, and breakeven point (i.e. the time required for payers to recoup PCD acquisition costs). The model calculated net budget impact annually for each of 5 consecutive years following the initiation of LE treatment, but cumulative results were reported at 2 years following treatment initiation to represent the usual time frame of interest to a commercial health plan. All costs and outputs were reported in US dollars. The model perspective is that of a hypothetical US insurer with 10 million commercial members.

Patient population

The model evaluated LE- and sequelae-related costs to the payer for patients with non-filarial LE and either comorbid chronic venous insufficiency (CVI) or frequent infections (≥2 per year), both of which represent sub-populations at risk for complications and attendant high medical costs. The modeled patient populations reflect the demographics and characteristics of patients from an analysis of subsequently case-matched LE patients in the Blue Health Intelligence (BHI) Research Database, a de-identified HIPAA-compliant commercial administrative claims dataset, which contains data on over 165 million covered lives in the US. Claims from this database were analyzed for the complete years 2012–2016. About 81,000 patients with non-filarial LE were initially defined as having one inpatient or two outpatient claims that included LE diagnosis codes. From that sample, 27,000 patients who were continuously enrolled with medical benefits for at least 12 months prior to and 6 months following LE diagnosis were included in the analysis.

The distribution of patients across LE treatment modalities was also derived from the BHI claims analysis. Specifically, among diagnosed and treated LE patients, 86.6% received conservative care (CONS), 1.7% received CONS + SPCD (HCPCS E0651), and 11.6% received CONS + APCD (HCPCS E0652, includes Flexitouch and other APCDs). We applied the same distribution of patients across treatment options to the CVI and 2+ infection sub-populations.

Additionally, and uniquely, the BHI dataset enabled reliable identification of claims for a specific APCD, Flexitouch (FLX, Tactile Medical). This was principally due to direct distribution of the device by the manufacturer and submission of claims linked to the company’s unique NPI number. In contrast, other APCDs are generally distributed through third-party durable medical equipment (DME) vendors, which typically sell devices from multiple manufacturers; those claims are, therefore, linked to the NPI of the DME provider. As a result, we were able to divide patients receiving an APCD (11.6%) into two mutually exclusive treatment groups. Among patients in the BHI dataset, 57% of those receiving an APCD received CONS + FLX and 43% received CONS + Other APCDs. The calculated distribution of patients across the four treatment groups in the Status Quo scenario is shown in Table 1.

Table 1. Status quo and alternate payer policy patient treatment distributions^a.

Treatment groups	CONS	CONS + SPCD	CONS + other APCD	CONS + FLX
Status quo patient distribution	86.6%	1.7%	5.0%	6.6%
Alternate payer policy patient distribution
SPCD failure removal scenario	86.6%	0.0%	5.0%	8.3%^b
Exclusive APCD scenario	86.6%	1.7%	0.0%	11.6%^c
Expanded access scenario	77.9%	1.7%	5.0%	15.3%^d

^a Patient distributions between treatment groups were derived from all patients who qualified for the BHI analysis (i.e. not restricted to patients included in case-matched cohorts).

^b In the SPCD failure removal scenario, the 1.7% of patients using CONS + SPCD are shifted to the CONS + FLX treatment group. These patients are added to the 6.6% using CONS + FLX in the status quo scenario.

^c In the exclusive APCD scenario, the 5.0% of patients using CONS + Other APCD are shifted to the CONS + FLX treatment group. These patients are added to the 6.6% using CONS + FLX in the status quo scenario.

^d In the expanded access scenario, 10% of the 86.6% of patients using CONS only are shifted to the CONS + FLX treatment group. These patients are added to the 6.6% using CONS + FLX in the status quo scenario.

Modeled strategies

To evaluate the budget impact of payer adoption of FLX under various conditions, we modeled the impact of changes to current commercial coverage for FLX (Status Quo) on the proportion of patients likely to receive FLX. The study assumes that the Status Quo coverage policy requires failure of an SPCD prior to allowing access to APCDs, as is seen in many plans^15,17,18. We then designed three Alternate Payer Policy scenarios (Table 2): (1) SPCD Failure Removal Scenario; (2) Exclusive APCD Scenario; and (3) Expanded Access Scenario.

Table 2. Description of status quo and alternate payer policy scenarios.

Status quo scenario	Criteria requiring failure of an SPCD prior to initiation of APCD therapy
Alternate payer policy scenarios
SPCD failure removal scenario	Criteria requiring failure of an SPCD prior to initiation of APCD therapy are removed, allowing all patients prescribed a PCD access to Flexitouch
Exclusive APCD scenario	Flexitouch is selected as the exclusive APCD offered to plan members
Expanded access scenario	Criteria restricting access to PCDs are relaxed, allowing patients treated conservatively access to Flexitouch

We modeled Status Quo and Alternate Payer Policy patient distributions across treatment groups as follows: The SPCD Failure Removal Scenario shifted all patients from the CONS + SPCD treatment group in the Status Quo scenario to CONS + FLX in the Alternate Payer Policy scenario; the Exclusive APCD Scenario shifted all patients from the CONS + Other APCD treatment group in the Status Quo scenario to CONS + FLX in the Alternate Payer Policy scenario; and the Expanded Access Scenario shifted 10% of patients from the CONS treatment group in the Status Quo scenario to CONS + FLX in the Alternate Payer Policy scenario. Because it may be unlikely that a majority of conservatively managed patients would require maintenance therapy with a PCD, we selected 10% as a conservative estimate. The distributions of patients between treatment groups in each Alternate scenario (i.e. after shifts in treatment) are shown in Table 1.

Model inputs

The treatment rates and sub-population prevalence inputs for the model were obtained from the analysis of BHI claims (Table 3). Specifically, patients with at least one claim for active LE treatment (including manual lymphatic drainage, LE education, and/or LE-related physical therapy/occupational therapy [PT/OT]) were identified to determine a representative treatment rate for diagnosed patients. Within the treated population, we identified the diagnosed prevalence of comorbid CVI or two or more LE-related infections within 12 months of the index LE-related claim.

Table 3. Status quo inputs.

Epidemiology and treatment					Sources
LE prevalence	0.1%				Moffatt et al.^27 Williams et al.^28
LE treatment rate^a	63.5%				2012–2016 BHI Claims Analysis
CVI prevalence^b	6.0%
Frequent infections prevalence^c	10.2%
Lympedema and sequelae-related costs by treatment group
	CONS	CONS + SPCD	CONS + other APCD	CONS + FLX
One-time PCD acquisition^d	N/A	$946	$5,459		2017 Medicare DMEPOS Fee Schedule^26
Medical cost inputs (mean per patient per year)
CVI sub-population
Home health	$370	$522	$511	$334	2012–2016 BHI Claims Analysis
Emergency	$11	$7	$45	$33
Inpatient	$8,715	$4,215	$4,186	$1,560
Physical therapy	$81	$34	$37	$39
Outpatient	$2,453	$2,313	$3,026	$1,090
Office	$584	$397	$705	$652
Lab	$1	$2	$3	$5
Other service location	$39	$0	$8	$126
Frequent infections sub-population
Home health	$649	$2,214	$921	$827	2012–2016 BHI Claims Analysis
Emergency	$71	$83	$116	$173
Inpatient	$16,382	$14,584	$13,308	$8,810
Physical therapy	$92	$92	$44	$87
Outpatient	$3,709	$5,329	$3,294	$3,026
Office	$826	$872	$1,039	$542
Lab	$2	$22	$4	$1
Other service location	$40	$31	$6	$121

^a Within the BHI claims database, patients diagnosed with non-filarial LE were identified by >1 inpatient claim or ≥2 outpatient claims within a 30 day period with the first or second ICD-9: 457.0, 457.1, 757.0, ICD-10: I97.2, I89.0, Q82.0. Treatment was identified by >1 claim for active LE treatment, including MLD [97124, 97140] AND/OR Education [97535] AND/OR PT/OT [97001, 97002, 97003, 97004, 97110, 97112, 97161, 97162, 97163, 97165, 97166, 97167, 97530]. Costs were identified from claims processed in 2012–2016. Costs associated with each qualifying patient were annualized.

^b CVI was identified by ≥1 inpatient claim or ≥2 outpatient claims within a 30 day period with the first or second claim for CVI [ICD-9: 459.3x, 459.8x, ICD-10: I87.2].

^c Infections were identified by the following codes: Cellulitis [ICD-9: 681.xx, 682.xx; ICD-10: L03.01x, L03.03x, L03.11x, L03.211, L03.213, L03.221, L03.31x, L03.81x, L03.90], Septic shock [ICD-9: 785.52, ICD-10: R65.21], Lymphangitis [ICD-9: 457.2, ICD-10: L03.02x, L03.04x, L03.12x, L03.212, L03.222, L03.32x, L03.89x, L03.91], Erysipelas [ICD-9: 035, ICD-10: A46], other local infections of skin and subcutaneous tissue [ICD-9: 686.xx, ICD-10: L08.xx].

^d PCD acquisition costs estimated as the midpoint of the published floor and ceiling rates in 2017 Medicare DMEPOS fee schedule. SPCD (HCPCS E0651) floor = $869, ceiling = $1,022; APCD (FLX and other APCDs, HCPCS E0652) floor = $5,017, ceiling = $5,902.

LE and sequelae-related costs were identified as any claims with LE as the first or second diagnosis (codes above), inpatient or outpatient claims with cellulitis diagnosis in any position (codes above), claims with septic shock diagnosis in any position (codes above), claims with lymphagitis diagnosis in any position (codes above), claims with erysipelas diagnosis in any position (codes above), claims with other local infections of skin and subcutaneous tissue diagnosis in any position (codes above) and inpatient claims or recurrent (>1) non-inpatient claims with ulcer diagnosis in any position [ICD-9: 707.x; ICD-10: I70.23x, I70.24x, I70.25, I70.33x, I70.34x, I70.35, I70.43x, I70,44x, I70.45, I70.53x, I70.54x, I70.55, I70.63x, I70.64x, I70.65, I70.73x, I70.74x, I70.75, L89.x, L97.x, L98.4x].

All cost inputs for the Status Quo scenario, including device acquisition and annual LE- and sequelae-related medical costs, are described in Table 3. Because device acquisition costs for commercial payers are not publicly available, we used Medicare reimbursement rates as a proxy²⁵. We also assumed that PCD costs were incurred by the payer, in full, at the start of year 1. This reflects the design of the claims analysis in which a PCD claim triggered the enumeration of subsequent costs, but it also reflects payer accounting for device purchases, in which costs are incurred up-front rather than amortized over the useful life of the device.

Per patient costs of medical care components for each treatment group were estimated from the analysis of BHI claims (Table 3). The claims analysis included costs associated with LE and relevant sequelae, as identified by a claim with a diagnosis code for primary or secondary LE, cellulitis, ulcers, septic shock, erysipelas, lymphangitis, or other local skin infection. Only costs following index treatment were included. Index treatment was defined for patients in the CONS treatment group as the first active LE treatment (e.g. compression garment, MLD, physical therapy) following the patient identification, and for patients in other treatment groups as receipt of a PCD following identification. These costs were then annualized, and mean costs were reported for each setting in which costs were incurred: i.e. inpatient, outpatient, emergency, physician office, home health, physical therapy, laboratory, and other service locations. Mean per patient per year costs were applied to all patients in the treatment group.

In the model, PCD costs were incurred as one-time costs at the beginning of year 1, but LE- and sequelae-related medical costs were calculated for the 12-month period following the initiation of treatment with conservative or PCD therapy. Because LE is a chronic condition, we assumed that medical costs would continue to be incurred beyond the first year. In the absence of information on the long-term cost consequences of LE, we assumed that medical costs remained constant for each subsequent year in the model.

Three case-matched cohorts were developed from the BHI claims dataset to control for differences in patient characteristics while evaluating the cost of treating patients with CONS + FLX vs each of the other three treatment modalities (CONS, CONS + SPCD, CONS + Other APCD). A stepwise propensity score matching approach was used to account for differences in clinical and demographic characteristics of the study cohorts. Cohorts were matched on Elixhauser comorbidity index components, age, gender, region of country, insurance type, and dummy indicators for the following clinical conditions: breast cancer, melanoma, uterine cancer, ovarian cancer, prostate cancer, cervical cancer, vaginal cancer, vulvar cancer, lymphoma, soft tissue sarcoma, congestive heart failure, CVI, venous leg ulcers, diabetes, iliac vein disorders, pulmonary hypertension, and postphlebitic syndrome. These case-matched comparisons were performed for both the CVI and recurrent infections sub-populations.

Status Quo scenario medical cost inputs for CONS, CONS + SPCD, and CONS + APCD treatment groups were estimated from these case-matched comparisons for each group vs CONS + FLX. Model inputs for annual costs in the Status Quo scenario are shown in Table 3 and for each Alternate Payer Policy scenario in Table 4.

Table 4. Alternate payer policy medical cost inputs (mean per patient per year).

	Patients shifted from CONS to CONS + FLX	Patients shifted from CONS + SPCD to CONS + FLX	Patients shifted from CONS + Other APCD to CONS + FLX
CVI sub-population
Home health	$334	$408	$380
Emergency	$33	$6	$36
Inpatient	$1,560	$297	$1,468
Physical therapy	$39	$16	$36
Outpatient	$1,090	$353	$1,283
Office	$652	$74	$600
Lab	$5	$0	$6
Other service location	$126	$0	$163
Frequent infections sub-population
Home health	$827	$845	$798
Emergency	$173	$200	$192
Inpatient	$8,810	$6,620	$9,355
Physical therapy	$87	$42	$88
Outpatient	$3,026	$4,130	$3,617
Office	$542	$389	$505
Lab	$1	$1	$1
Other service location	$121	$0	$175

To estimate the medical cost inputs for the sub-set of patients shifted to the CONS + FLX treatment group in each Alternate Payer Policy scenario, we performed case-matched comparisons of costs for CONS + FLX patients vs the treatment group from which the patients originated. For example, in the SPCD Failure Removal scenario, a group of patients were shifted from the CONS + SPCD treatment group to the CONS + FLX treatment group. Costs for these patients were derived from the case-matched analysis of patients treated with CONS + SPCD vs CONS + FLX. Each Alternate Payer Policy scenario is independent and, therefore, no individual patient is affected by more than one scenario. This assumption of independence allows us to evaluate the additive effects when multiple Alternate Payer Policies are enacted simultaneously.

Sensitivity analysis

One-way sensitivity analyses were performed for each model input to evaluate the impact of each parameter on budget impact and on model conclusions. Specifically, we varied estimates for LE and sub-population prevalence, treatment distributions, and the largest patient medical cost input buckets (inpatient and outpatient costs), by ±20%. PCD costs were varied across the range of the SPCD and APCD Medicare reimbursement rates.

Results

CVI sub-population

Budget impact results are shown in Table 5. In the Status Quo scenario, the total cumulative 2-year cost for LE- and sequelae-related care for all phlebolymphedema patients covered by the hypothetical payer is $8,957,911. Over the same 2-year time frame, the Alternate Payer Policy SPCD Failure Removal, Exclusive APCD, and Expanded Access scenarios cost the plan $8,905,070, $8,784,594, and $8,582,310, respectively. The difference between the two states result in a 2-year budget impact of –$52,841, –$173,317, and –$375,601, respectively, or net cost-savings.

Table 5. Cumulative total 2-year budget impact.

Status quo	Alternate payer policy		Budget impact^a	Breakeven point^b
CVI sub-population
$8,957,911	SPCD failure removal	$8,905,070	–$52,841	0.71 years
	Exclusive APCD	$8,784,594	–$173,317	N/A
	Expanded access	$8,582,310	–$375,601	0.65 years
Frequent infections sub-population
$27,785,492	SPCD failure removal	$27,592,761	–$192,729	0.41 years
	Exclusive APCD	$27,526,091	–$259,399	N/A
	Expanded access	$27,172,311	–$613,179	0.67 years

^a Budget impact is calculated as Alternate – Status quo; therefore, a negative budget impact reflects cost savings.

^b Breakeven point is the time required for payers to recoup PCD costs, which are incurred, in full, at the beginning of year 1.

In each of the three Alternate Payer Policy scenarios in which patients are granted access to FLX, the cost savings are realized primarily due to lower inpatient costs, with additional cost savings from lower outpatient costs. In the SPCD Failure Removal scenario, additional device acquisition costs for FLX are offset by healthcare cost savings at 0.71 years, while, in the Expanded Access scenario, breakeven occurs at 0.65 years.

Frequent infections sub-population

The total cumulative costs for LE and sequelae-related care incurred over 2 years by the hypothetical insurer for all patients within the sub-population of LE patients who contract two or more infections per year are $27,785,492 in the Status Quo scenario. Over the same 2-year time frame, the total cost of treating the same LE sub-population with frequent infections in the Alternate Payer Policy SPCD Failure Removal, Exclusive APCD, and Expanded Access scenarios was modeled to be $27,592,761, $27,526,091, and $27,172,311, respectively. The difference between the two states results in a 2-year budget impact of –$192,729, –$259,339, and –$613,179, respectively, or net cost-savings.

For the frequent infections population, cost-savings under the three policy scenarios are mostly due to lower inpatient costs. In the SPCD Failure Removal scenario, additional device acquisition costs for FLX are offset by healthcare cost savings at 0.41 years and in the Expanded Access scenario breakeven occurs at 0.67 years.

Sensitivity analysis

The results of the one-way sensitivity analyses demonstrate that the model was most influenced by diagnosed LE prevalence, treatment rate, and sub-population prevalence, followed by FLX acquisition costs. Because diagnosed LE prevalence, treatment rate, and sub-population prevalence are directly correlated with the size of the impacted population, the change in budget impact directly reflects the sensitivity multiplier. For instance, when each of these parameters is decreased by 20%, the budget impact is also decreased by 20%. Upward variation by 20% of FLX acquisition costs, inpatient costs, and outpatient costs result in less than a 6% increase in budget impact (i.e. reduction in cost savings). Despite variations in model input parameters, we found consistent cost savings across each Alternate Payer Policy scenario.

Discussion

Drawing on administrative claims data and other sources, we evaluated the 2-year budget impact of increased utilization of FLX by private US payers for specific LE sub-populations. Our model indicates that increasing access to FLX yielded a favorable budget impact in every modeled scenario. Thus, private health plans may achieve cost savings by granting or improving access to FLX for certain identifiable LE patient populations. Even though FLX adds one-time PCD acquisition costs of ∼$5,500 and ∼$4,500 for conservatively managed and SPCD patients (based on CMS reimbursement rates), respectively, net cost savings are achieved in less than 1 year due to lower costs of subsequent LE and sequelae-related medical care, primarily inpatient LE and sequelae-related expenditures. These findings cast doubt on the benefits of payer guidelines requiring both conservative therapy and documented failure of an SPCD prior to approving access to an APCD, specifically FLX, for patients in the studied sub-populations. While utilization of FLX following failure of an SPCD could not be evaluated in this study due to the small sample, other studies have demonstrated that early treatment reduces complications and halts progression of LE^26,27, suggesting effective treatment options should be employed as early as possible.

While several studies have evaluated the costs associated with LE treatment, to our knowledge this is the first study to model the costs of receiving different LE treatment options in case-matched cohorts. Existing literature has demonstrated the clinical and economic benefits of PCDs²⁸ and, specifically, FLX²¹, most commonly in cancer patients. Brayton et al.²⁸ identified reductions in medical costs of ∼ $12,000 for cancer-related LE patients for the 12 months after receiving a PCD compared to the previous 12 months. By comparing the medical costs for patients on CONS (Table 3) to the medical costs for patients shifted from CONs to CONS + FLX (Table 4), our study suggests that per-patient annual medical costs are reduced by > $8,000 for patients who undergo this payer policy shift, thus confirming the substantial savings achievable with PCD use. Similarly, drawing on a large administrative claims dataset, Karaca-Mandic et al.²¹ found that costs declined for patients receiving FLX by 37% and 36% for cancer-related and non-cancer-related LE patients, respectively. Of note, the study designs differ substantially: while Brayton et al.²⁸ and Karaca-Mandic et al.²¹ compare medical costs for the 12 months before and after initiation of PCD treatment, our study compared costs for case-matched cohorts treated with FLX vs other treatment modalities, including conservative treatment, SPCDs, and other APCDs.

Our results also demonstrate annual per patient medical cost savings for these other comparator arms, suggesting savings for patients shifted from CONS + SPCD to CONS + FLX ($6,336 and $11,000 for CVI and infections sub-populations, respectively) and from CONS + Other APCD to CONS + FLX ($4,549 and $4,001 for CVI and infections sub-populations, respectively). Lerman et al.²⁹ found similar results in patients with LE and comorbid CVI.

Finally, our study is the first to evaluate the budget impact of PCDs in LE patients with comorbid CVI or frequent infections from a US private payer perspective. Our model suggests that cost savings can be achieved in these sub-populations with greater use of FLX when compared to all other treatment modalities.

Our study has several limitations. First, the use of claims data assumes accurate coding and does not allow for detailed understanding of the exact clinical circumstances of LE treatment. Although treatment arms were case-matched, we were not able to determine the LE severity through claims. Second, while it is not possible to determine the level of adherence to therapy based on claims (i.e. a claim for receipt of a PCD does not guarantee prescribed use), the data as analyzed reflect costs observed in real-world management of LE. Third, the distribution of patients between treatment arms in this study was determined using claims from the BHI database, comprised of members of Blue Cross and Blue Shield Association (BCBS) health plans. This distribution may differ for other health plans, particularly as some BCBS plans have less restrictive coverage policies regarding APCDs^16,19 than other known policies^15,17,18,20. Fourth, pharmaceutical costs for LE patients were excluded from the model. Pharmaceutical costs are likely to be covered by a patient’s pharmacy benefit, while the costs included in our model are medical and DME costs most likely to be covered within the plan’s medical benefit. As further support for the exclusion of pharmaceutical costs, Brayton et al.²⁸ also found that pharmaceutical costs related to LE care were negligible for both patients treated with PCDs and those not treated with PCDs. Finally, while there is known to be significant under-diagnosis and under-treatment of LE^30,31, it is unclear in which direction improved diagnosis of LE would impact our model results.

Despite these limitations, our study provides clear evidence of cost savings achieved through increased utilization of FLX in specific lymphedema populations. We recognize that models such as this one are only a representation of reality and that payers may not shift all patients at once to FLX. For example, it may be the case that, under some plans, when a step edit is removed, patients may get a variety of APCDs. However, our objective was to estimate the budget impact of moving specific LE sub-populations to FLX, and, because medical cost savings outweigh additional device costs for these patients, shifting an even greater share of patients to FLX over time should generate even greater cost savings. The observed lower LE and sequelae-related medical costs may also under-estimate the true impact of shifting patients to FLX, as unrelated healthcare costs as well as non-healthcare costs may be reduced with improvement in overall health.

Conclusion

This analysis provides important insights for individuals involved in the development of medical policies for private US payers. The primary finding of this study, that near-term cost savings can be achieved by shifting specific LE sub-populations to FLX, calls into question current US payer coverage policies that restrict access to PCDs in general, and to FLX specifically. The clinical benefits of FLX, including reduction in episodes of cellulitis²¹, reduced limb volume and edema^9,13,14, and improvement in skin fibrosis and function^9,13,14, have been demonstrated in several publications, but, until now, the budget impact and return on investment has remained uncertain. Our results suggest that US health plans may achieve near-term cost savings by removing existing requirements for failure of SPCDs and other restrictions to FLX and instead granting faster access to FLX for the identified LE patient populations.

Figure 1. One-way sensitivity analysis, sensitivity multiplier = ±20%. Sensitivities were tested for inpatient and outpatient costs for patients shifted to CONS + FLX from other treatment groups.

Transparency

Declaration of funding

This study was sponsored by Tactile Medical (Minneapolis, MN), the company that manufactures the Flexitouch pneumatic compression device for the treatment of lymphedema.

Declaration of financial/other interests

TO is a consultant to Tactile Medical, a company that manufactures a product to treat lymphedema, where he serves as Chief Medical Officer. He holds stock options as part of the compensation plan. AMC, JAG, JI, and LG received consultative reimbursement from Tactile Medical for their independent performance of the analysis. TN received consultative reimbursement from Health Advances for his independent performance of the statistical analysis. No other disclosures are reported. JME peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Acknowledgments

No assistance in the preparation of this article is to be declared.

Data availability statement

The claims data that support the findings of this study are available via purchase from Blue Health Intelligence. The methodology section contains description of the specific claims data utilized. Medicare fee schedule data is openly available from the Centers for Medicare and Medicaid Services at https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/DMEPOSFeeSched/DMEPOS-Fee-Schedule-Items/DME17-D.html.