Commercial Insurance Rates and Coding for Lymphedema Procedures: The Current State of Confusion and Need for Consensus

Abstract

Introduction

Surgical treatment of lymphedema has outpaced coding paradigms. In the setting of ambiguity regarding coding for physiologic procedures [(lymphovenous bypass (LVB) and vascularized lymph node transplant (VLNT)], we hypothesized that there would be variation in commercial reimbursement based on coding pattern.

Methods

We performed a cross-sectional analysis of 2021 nationwide hospital pricing data for 21 Current Procedural Terminology (CPT) codes encompassing excisional (direct excision, liposuction), physiologic (LVB, VLNT), and ancillary (lymphangiography) procedures. Within-hospital ratios (WHRs) and across-hospital ratios (AHRs) for adjusted commercial rates per CPT code quantified price variation. Mixed effects linear regression modeled associations of commercial rate with public payer (Medicare and Medicaid), self-pay, and chargemaster rates.

Results

270,254 commercial rates, including 95,774 rates for physiologic procedures, were extracted from 2,863 hospitals. Lymphangiography codes varied most in commercial price (WHR 1.76 – 3.89, AHR 8.12 – 44.38). For physiologic codes, WHRs ranged from 1.01 (VLNT; free omental flap) to 3.03 (LVB; unlisted lymphatic procedure), and AHRs ranged from 5.23 (LVB; lymphatic channel incision) to 10.36 (LVB; unlisted lymphatic procedure). Median adjusted commercial rates for excisional procedures ($3,635.84) were higher than for physiologic procedures ($2,560.40; p<0.001). Commercial rate positively correlated with Medicare rate for all physiologic codes combined, though regression coefficients varied by code.

Conclusions

Commercial payer-negotiated rates for physiologic procedures were highly variable both within and across hospitals, reflective of variation in CPT codes. Physiologic procedures may be undervalued relative to excisional procedures. Consistent coding nomenclature should be developed for physiologic and ancillary procedures.

Introduction

Over the past century, the surgical treatment of lymphedema has evolved from excisional operations to microsurgical restoration of lymphatic circulation. The debulking procedures that Charles and Sistrunk pioneered in the early 20^th century are now reserved for advanced chronic lymphedema in which there is profound excess skin and subcutaneous tissue.^1–3 Liposuction, popularized in the 1990s for the treatment of upper extremity lymphedema,⁴ is now similarly utilized in cases of advanced chronic lymphedema.³ While these techniques are effective in select patients with severe fat hypertrophy, they are not indicated in patients with earlier stages of lymphedema with a fluid-dominant limb. The majority of patients with mild to moderate lymphedema require microsurgery to provide an exit for lymph. These techniques, termed “physiologic procedures” because they restore lymphatic flow by circumventing lymphatic obstruction and/or recruiting functional lymph nodes, include lymphovenous bypass (LVB) and vascularized lymph node transplant (VLNT). Current evidence supports use of LVB for early stage lymphedema, VLNT for more advanced lymphedema, and combination approaches including LVB, VLNT, and/or liposuction in cases with large soft tissue excess.^3,5 Immediate lymphatic reconstruction is another emerging variant of physiologic procedures with preventive rather than therapeutic intent in breast cancer patients undergoing axillary lymph node dissection.⁶

Coding for lymphedema procedures has not kept pace with advances in surgical treatment. Current Procedural Terminology (CPT) codes, which are predominantly used for reimbursement claims, are intended to provide a “precise and unambiguous” standard nomenclature for physician services.⁷ However, for lymphedema surgery, surgeons are forced to fit new physiologic procedures within an unchanged coding framework, resulting in nonspecific and variable coding patterns. There is scant guidance on this topic, including from the American Society of Plastic Surgeons (ASPS), whose recommendations are limited to using a free tissue transfer CPT code for VLNT, and an unlisted CPT code for LVB with comparison to vascular repair codes for pricing information.⁸ National lymphedema initiatives have acknowledged the payment barriers linked to these coding problems, and identified improving reimbursement for lymphedema as a priority issue.⁹

While there has been much informal discussion among surgeons and finance departments about inadequacies in current lymphedema CPT codes, the implications of this coding ambiguity have not been systematically evaluated. The purpose of this study was to examine variation in commercial payer-negotiated rates for the spectrum of CPT codes that may be used for lymphedema physiologic procedures, both alone and in relation to excisional procedures. The secondary objective was to characterize this variation vis-à-vis lymphedema cost-effectiveness analyses based on Medicare rates, which have been instrumental in the justification and popularization of lymphedema physiologic procedures.

Methods

Study Design and Data Sources

We performed a cross-sectional analysis of 2021 hospital pricing data for CPT codes that are used for direct excision (CPT 15832–33, 15836–7) and liposuction (CPT 15858–9), and may be used for LVB (CPT 25306, 34530, 35226, 35236, 38305, 38308, 38999), VLNT (15757, 15758, 49906), and lymphangiography (CPT 15860, 37799, 38790, 38792, 38900, 75801; See Supplemental Tables 1 and 2 for descriptions of CPT codes). CPT codes for physiologic and ancillary (i.e. lymphangiography) procedures were selected based on existing ASPS guidance,⁸ publications that specify codes employed in study methods,^10,11 and the coding practices of contemporary lymphedema surgeons, including the authors. Individual institutional review board approval was not required for this study, as it does no involve human subjects research.

Pricing data were obtained from Turquoise Health, a data broker that mines hospital websites for price disclosures that are required following implementation of the Hospital Price Transparency Rule on January 1, 2021.^12,13 Turquoise Health data are drawn from 5,700 US hospitals (94% of 6,090 total hospitals), including both academic institutions and community hospitals, and have been previously used for price variation research.^14–19 Pricing data included plan-specific commercial payer-negotiated rates, public payer (Medicare and Medicaid) rates, self-pay rates (i.e. out-of-pocket prices paid by the uninsured or patients paying outside of insurance), and list prices (i.e. the amount billed, commonly called “chargemaster rates”). All rates were adjusted using the Geographic Adjustment Factor (GAF) to account for city-level geographic variation in input costs, with the exception of Medicare rates that are intrinsically price-adjusted.²⁰

Statistical Analysis

Commercial price variation was quantified using within-hospital ratios (WHRs) and across-hospital ratios (AHRs).^21–23 For each CPT code, the WHR was calculated as the median of the maximum payer-negotiated rate divided by the minimum payer-negotiated rate per hospital, averaged over all hospitals. For each CPT code, the AHR was calculated as the 90^th percentile median GAF-normalized rate divided by the 10^th percentile median GAF-normalized rate across all hospitals. The Wilcoxon rank-sum test was used to compare GAF-adjusted commercial rates between excisional and physiologic procedure groups given the non-Gaussian distribution of the data.

Mixed effects linear regression with a random intercept by hospital was used to model the relationship between GAF-adjusted commercial rate and Medicare, GAF-adjusted Medicaid, GAF-adjusted self-pay, and GAF-adjusted list rates for physiologic CPT codes. This model was selected to account for inter-hospital variation in factors that can affect the relationship between rates at a given hospital (e.g. contract experience, number of patients per payer). Individual mixed effects linear regressions per physiologic CPT code were additionally performed to model the association between GAF-adjusted commercial and Medicare rates specific to each code, also with a random intercept by hospital. The likelihood ratio (LR) test evaluated the fitness of these models in relation to simple linear regression. Analyses were conducted using Stata/SE Version 15.0 (StataCorp LLC, College Station, TX). P-values of less than 0.05 were considered statistically significant.

Results

Hospital pricing data were extracted from 2,863 unique US hospitals with at least one commercial rate (84.9%, N=2,431), Medicare rate (61.6%, N=1,763), Medicaid rate (44.0%, N=1,261), self-pay rate (78.8%, N=2,257), and/or list price (84.5%, N=2,418). For the 21 queried CPT codes, there were a total of 270,254 unique commercial rates; 95,774 of these rates were for physiologic procedures.

Table 1 enumerates the median GAF-adjusted commercial payer-negotiated rate per CPT code, in addition to within- and across- hospital ratios. WHRs ranged from a minimum of 1.01 for free omental flap (CPT 49906, VLNT) to a maximum of 3.89 for unlisted vascular procedure (CPT 37799, lymphangiography). AHRs ranged from a minimum of 4.60 for liposuction of the upper extremity (CPT 15879) to a maximum of 44.38 for intraoperative mapping of lymph nodes (CPT 38900, lymphangiography; Figure 1). Summary statistics and ratios for GAF-adjusted commercial rates are reproduced in Table 2 at the procedure group level. Commercial rates for excisional procedures (median $3,635.84) were higher than for physiologic procedures (median $2,560.40; p<0.001). Figure 2 illustrates the distribution of median commercial rates across all lymphedema procedure groups, and of physiologic procedures in comparison to token breast reconstructive procedures.²⁴

Table 1.

Descriptive statistics for commercial payer-negotiated rates for lymphedema operations

			GAF-adjusted commercial rate, $b			Within-hospital ratio (IQR)c	Across-hospital ratiod
Procedure (CPT code)a	No. rates	No. hospitals	Median (IQR)	p10	p90
Direct Excision
Thigh (15832)	6,990	740	3,489.29 (2,405.28 – 6,016.37)	1,485.36	9,194.84	1.86 (1.00 – 4.18)	5.34
Leg (15833)	6,047	692	3,634.97 (2,392.90 – 6,197.43)	1,454.99	9,146.69	1.72 (1.00 – 4.09)	5.36
Arm (15836)	7,194	753	3,489.29 (2,326.26 – 6,028.44)	1,339.04	9,683.98	2.07 (1.00 – 4.43)	5.32
Forearm or hand (15837)	5,563	669	3,196.61 (2,230.04 – 5,285.15)	1,161.16	8,998.27	1.69 (1.00 – 4.00)	4.95
Liposuction
Upper extremity (15878)	6,205	691	3,517.63 (2,133.89 – 5,378.44)	1,588.12	7,724.32	1.65 (1.00 – 3.91)	4.60
Lower extremity (15879)	6,443	716	4,426.60 (3,034.66 – 7,427.18)	1,833.66	11,848.44	1.63 (1.00 – 3.48)	6.29
LVB
Saph-pop vein anastomosis (34530)	5,104	625	4,781.38 (2,980.47 – 8,943.26)	1,968.31	14,083.59	1.58 (1.00 – 3.58)	5.66
Repair UE blood vessel (35206)	17,781	1,210	3,756.10 (2,375.66 – 6,299.60)	1,183.20	10,221.79	2.82 (1.00 – 6.04)	5.91
Repair LE blood vessel (35226)	19,567	1,260	1,910.32 (1,002.21 – 3,242.78)	620.33	7,190.83	2.99 (1.00 – 8.05)	7.70
Repair vessel with vein graft (35236)	6,118	691	5,534.09 (3,044.39 – 9,526.33)	1,546.32	14,861.32	1.59 (1.00 – 4.47)	8.62
Lymphatic channel incision (38305)	7,369	734	3,015.93 (1,681.70 – 5,262.50)	734.99	7,793.74	2.05 (1.00 – 4.85)	5.23
Lymphangiotomy (38308)	6,293	688	3,728.37 (2,027.44 – 6,819.05)	916.29	10,677.63	1.86 (1.00 – 4.77)	6.06
Unlisted lymphatic procedure (38999)	23,749	1,003	1,068.63 (643.66 – 2,368.32)	390.18	4,781.49	3.03 (1.00 – 10.38)	10.36
VLNT
Other flaps and grafts (15757)	3,625	489	4,307.49 (2,520.90 – 7,527.96)	1,692.51	10,539.07	1.10 (1.00 – 3.39)	5.34
Free fascia flap (15758)	3,551	450	4,298.62 (2,460.58 – 6,453.62)	1,654.97	10,106.96	1.33 (1.00 – 3.57)	6.53
Free omental flap (49906)	2,617	408	4,231.15 (2,159.98 – 7,512.56)	1,372.47	10,986.48	1.01 (1.00 – 3.53)	7.40
Lymphangiography
IV injection to test blood flow (15860)	6,409	730	632.51 (288.56 – 2,099.67)	160.05	4,316.50	1.76 (1.00 – 8.20)	21.60
Unlisted vascular procedure (37799)	50,264	1,507	2,000.59 (791.75 – 4,002.06)	245.05	8,425.45	3.89 (1.30 – 13.74)	11.60
Lymphangiography (38790)	16,275	878	687.16 (300.74 – 1,673.26)	105.26	3,260.82	3.00 (1.03 – 9.81)	22.99
Lymphoscintigraphy (38792)	40,209	1,946	625.56 (351.72 – 1,183.83)	177.20	2,340.36	3.38 (1.60 – 10.56)	11.38
Intraop mapping of LNs (38900)	9,756	846	1,249.59 (368.78 – 9,367.37)	196.55	9,367.37	2.47 (1.00 – 5.80)	44.38
Interpretation, unilateral (75801)	13,090	723	1,023.76 (554.49 – 1,740.06)	250.19	2,892.46	3.07 (1.19 – 7.85)	8.12

CPT, Current Procedural Terminology. GAF, Geographic Adjustment Factor. IQR, interquartile range. IV, intravenous. LE, lower extremity. LN, lymph node. LVB, lymphovenous bypass. p90, 90^th percentile median rate. p10, 10^th percentile median rate. Saph-pop, saphenopopliteal. UE, upper extremity. VLNT, vascularized lymph node transplant.

^aCPT descriptions are abbreviated. For full descriptions, see Supplemental Table 1.

^bCalculated based on the median payer-negotiated rates at each hospital.

^cCalculated as the median of the median of the maximum payer-negotiated rate divided by the minimum payer-negotiated rate per hospital.

^dCalculated as the median of the 90^th percentile median GAF-normalized rate divided by the 10^h percentile median GAF-normalized rate across all hospitals.

Figure 1. Within- and across-hospital variation in GAF-adjusted commercial rate per CPT code for (A) excisional, (B) physiologic, and (C) ancillary procedures.

AHR, across-hospital ratio. LVB, lymphovenous bypass. VLNT, vascularized lymph node transplant. WHR, within-hospital ratio. WHR plots exclude outliers. Note difference in x-axis scale for lymphangiography plots. Red dashed lines separate procedure groups.

Table 2.

Descriptive statistics for commercial payer-negotiated rates for lymphedema operations, by procedure group

			GAF-adjusted commercial rate, $b			Within-hospital ratio (IQR)c	Across-hospital ratio (IQR)d
Procedure groupa	No. rates	No. hospitals	Median (IQR)	p10	p90
Excisional
Direct Excision	25,794	2,854	3,473.88 (2,339.83 – 5,868.39)	1,336.71	9,194.84	1.85 (1.00 – 4.13)	5.33 (4.32 – 5.34)
Liposuction	12,648	1,407	4,029.29 (2,455.78 – 6,233.10)	1,666.71	9,706.60	1.64 (1.00 – 3.71)	6.29 (4.60 – 6.29)
Total	38,442	4,261	3,635.84 (2,376.03 – 5,995.71)	1,468.00	9,194.84	1.78 (1.00 – 4.06)	5.34 (4.95 – 5.36)
Physiologic
LVB	56,114	5,208	2,536.70 (1,078.55 – 4,972.70)	596.50	9,914.84	2.38 (1.00 – 5.73)	6.06 (5.91 – 8.62)
VLNT	9,793	2,350	4,300.04 (2,400.04 – 7,511.86)	1,502.72	10,709.99	1.13 (1.00 – 3.55)	6.53 (5.34 – 7.40)
Total	89,656	7,558	2,560.40 (1,138.32 – 4,906.42)	608.42	9,117.76	2.16 (1.00 – 5.40)	6.53 (5.91 – 7.70)
Other/Ancillary
Lymphangiography	136,003	6,630	1,009.34 (416.46 – 2,510.33)	196.86	5,194.66	3.03 (1.11 – 9.35)	11.60 (11.38 – 22.99)

GAF, Geographic Adjustment Factor. IQR, interquartile range. LVB, lymphovenous bypass. p90, 90^th percentile median rate. p10, 10^th percentile median rate. VLNT, vascularized lymph node transplant.

^aCPT descriptions are abbreviated. For full descriptions, see Supplemental Table 1.

^bCalculated based on the median payer-negotiated rates at each hospital.

^cCalculated as the median of the median of the maximum payer-negotiated rate divided by the minimum payer-negotiated rate for each hospital.

^dCalculated as the median of the 90^th percentile median GAF-normalized rate divided by the 10^h percentile median GAF-normalized rate across all hospitals.

Figure 2. Comparative kernel density distributions of GAF-normalized commercial rates for physiologic lymphedema procedures in comparison to (A) excisional lymphedema procedures and (B) breast reconstructive procedures.

LVB, lymphovenous bypass. VLNT, vascularized lymph node transplant. Breast procedures correspond to the following CPT codes: insertion of tissue expander (CPT 19357), breast reconstruction with free flap (19364), and breast reduction (19318).

For all physiologic CPT codes combined, mixed effects regression at the hospital level yielded significant associations of commercial rate with Medicare rate (coefficient 0.47, 95% CI 0.39 – 0.55, p<0.001), Medicaid rate (coefficient 0.33, 95% CI 0.26 – 0.40, p<0.001), self-pay rate (coefficient 0.25, 95% CI 0.18 – 0.31, p<0.001), and list price (coefficient 0.12, 95% CI 0.10 – 0.15, p<0.001). The relationship between the five payer rates per CPT code is illustrated in Figure 3 for excisional and physiologic procedures.

Figure 3. Comparison among GAF-adjusted rates per CPT code for excisional and physiologic procedures.

LVB, lymphovenous bypass. VLNT, vascularized lymph node transplant. All rates except for Medicare are normalized based on the Geographic Adjustment Factor (GAF). Red dashed lines separate procedure groups.

Output from individual simple linear regressions of commercial rate as a function of Medicare rate for each physiologic CPT code is summarized in Table 3. For all CPT codes except free omental flap (CPT 49906), there was a significant positive association between commercial and Medicare rate, with variable magnitude of the regression coefficients. LR test p-values were <0.001 for all models, indicating superiority of mixed effects compared to simple linear models. Figure 4 shows the relationship between GAF-adjusted commercial rate and Medicare rate for all physiologic CPT codes combined.

Table 3.

Mixed effects linear regression per physiologic CPT code for GAF-normalized commercial rate as a function of Medicare rate, with random intercept by hospital

CPT Description (Code)	Coefficient	95% Confidence Interval	p-value
LVB
Saph-pop vein anastomosis (34530)	0.72	0.57 – 0.88	<0.001
Repair UE blood vessel (35206)	0.95	0.82 – 1.07	<0.001
Repair LE blood vessel (35226)	0.88	0.77 – 0.98	<0.001
Repair vessel with vein graft (35236)	0.65	0.46 – 0.84	<0.001
Lymphatic channel incision (38305)	0.94	0.72 – 1.16	<0.001
Lymphangiotomy (38308)	0.80	0.48 – 1.12	<0.001
Unlisted lymphatic procedure (38999)	1.51	1.17 – 1.85	<0.001
VLNT
Other flaps and grafts (15757)	0.20	0.08 – 0.33	0.001
Free fascia flap (15758)	0.30	0.18 – 0.43	<0.001
Free omental flap (49906)	0.19	−0.01 – 0.40	0.068

LE, lower extremity. LVB, lymphovenous bypass. UE, upper extremity. VLNT, vascularized lymph node transplant.

Figure 4. Relationship between commercial and Medicare rates for physiologic procedures.

Graphed values are between 5^th and 95^th percentiles for respective rates. Slope of line approximates coefficient from simple linear regression of commercial rate as a function of Medicare rate: 0.75 (95% CI 0.71 – 0.80, p<0.001).

Discussion

Our analysis of nearly 100,000 unique commercial rates for CPT codes utilized for lymphedema physiologic procedures demonstrates substantial variation in commercial reimbursement based on coding pattern. For instance, for LVB, CPT 38308 (lymphangiotomy) yielded a median reimbursement of $3,728.37, compared to $1,068.63 for CPT 38999 (unlisted lymphatic procedure). Though within- and across-hospital ratios also indicate price variation for excisional procedures, the magnitude of this variation was greater for physiologic procedures when considered based on individual CPT code (excisional WHR 1.63 – 2.07, AHR 4.60 – 6.29 vs. physiologic WHR 1.01 – 3.03, AHR 5.23 – 10.36) and collectively based on procedure group (excisional WHR 1.78, AHR 5.34 vs. physiologic WHR 2.16, AHR 6.53). Variation was most pronounced for lymphangiography (WHR 1.76 – 3.89, AHR 8.12 – 44.38). Lower median commercial rates for physiologic compared to excisional procedures suggest that physiologic procedures are undervalued relative to excisional procedures, particularly given the additional specialized training and postoperative care required for microsurgical reconstruction. Additionally LVB and VLNT are among the most technically challenging types of microsurgery performed compared, for instance, to a standard deep inferior epigastric perforator (DIEP) flap for breast reconstruction. Though DIEP flap reconstruction encompasses an aesthetic component that does not apply to lymphedema physiologic procedures, LVB involves the technical feat of a typically sub-0.5 mm anastomosis and VLNT often requires lymphatic mapping and reoperating on a heavily scarred, radiated axilla. One might expect that reimbursement would at least be within range of other microsurgical procedures, but lymphatic microsurgery falls far short of this mark.

Variation for lymphedema CPT codes is comparable to that of other CPT codes for plastic surgical procedures calculated via identical methodology. Our group has previously demonstrated that for breast reconstruction, WHR and AHR ranges were 1.6 – 2.5 and 4.5 – 18.3, respectively, with greater variation for implant-based (WHR 1.9 – 2.5, AHR 11.5 – 18.3) compared to autologous (WHR 1.6 – 2.0, AHR 4.5 – 11.2) reconstructive procedures.²⁴ Cleft lip and palate procedures, including primary lip and palate repair, secondary repair, and cleft rhinoplasty, had WHRs ranging from 2.0 – 2.9, and AHRs from 5.4 – 13.7.²⁵ However, unlike codes for breast reconstructive and cleft lip and palate surgery, which have well-defined CPT codes that were tailored to these procedures, codes for lymphedema physiologic procedures are unclear. In the setting of both substantial commercial rate variation and coding ambiguity, commercial reimbursement for lymphedema physiologic procedures will vary dramatically based on the selection and combination of codes that surgeons choose to bill for these procedures.

Coding ambiguity in the setting of significant price variation for physiologic CPT codes has consequences for providers’ financial sustainability and patient access to surgical treatment of lymphedema. If hospital reimbursement varies widely based on coding patterns, those operating at the lower end of the reimbursement range may be faced with challenges regarding the financial feasibility lymphedema practice. Hospitals may discourage surgeons from performing these procedures if not economically favorable. In addition, insurance coverage for physiologic procedures remains sporadic, as physiologic procedures are considered investigational by some insurers.^8,26 A patient’s access to physiologic procedures therefore depends on selection of approved CPT codes combined with strictly defined diagnosis codes, both of which require negotiation with insurers and careful planning by finance departments. The resources spent deciphering lymphedema billing introduce substantial waste into the healthcare system, as does price variation that is dependent upon code selection rather than quality metrics.

Though important for billing and reimbursement, consistent use of CPT codes for physiologic procedures is also needed to strengthen the case that these procedures should be considered standard of care. Over the past decade, as “pay for performance” initiatives have gained traction in US healthcare, CPT codes have become the predominant method utilized by the Centers for Medicare & Medicaid Services (CMS), third-party payers, and national organizations to track clinical performance measures in procedure-based care.⁷ There are currently 16 prospective studies on VLNT demonstrating significant improvement in limb volume, quality of life, and reduced incidence of cellulitis,^27–43 and similarly strong evidence for LVB.^5,36 However, ambiguity in CPT coding has precluded nationwide collation and analysis of data, and thus it has not yet been possible to perform outcomes studies on a national scale. Consistent CPT coding of LVB and VLNT is thus critical to assembling a cohesive body of evidence from national quality improvement and claims databases that can be cited as rationale for lifting the “investigational” label that currently challenges access to care.

The significant correlation of commercial with Medicare rates for physiologic procedures demonstrated in this study suggests that prior cost-effectiveness analyses can be generalized to commercial payer perspectives, though price variation based on CPT code selection weakens these analyses’ economic modeling. All published national analyses demonstrating the cost-effectiveness of both preventive and therapeutic physiologic procedures utilize Medicare and/or Medicaid fee schedules in their cost forecasting, as these rates were historically the only publicly-available rates prior to the Hospital Price Transparency Rule.^10,44,45 Only one of these studies discloses the CPT codes utilized in their modeling.¹⁰ Though we demonstrate an overall correlation of commercial and Medicare rates for physiologic CPT codes, the magnitude of this correlation varied depending on the code. Thus, given the large heterogeneity in coding patterns, the strength and generalizability of the conclusions from prior cost-effectiveness studies depend on code selection. If coding ambiguity and price variation continue, future authors of cost-effectiveness studies should disclose the CPT codes utilized in their models to facilitate interpretation of their findings.

The overarching solution to the coding problems and consequences highlighted in this study is to establish CPT codes that are specific to LVB and VLNT, and not borrowed from preexisting surgical procedures. This would involve working with the American Medical Association (AMA), plastic surgery interest groups, and other surgical and medical specialty societies with common interest to add lymphedema-specific codes to the current CPT code repertoire. Given that it most recently took several years of working with the AMA CPT Panel to revise codes for breast reconstruction,⁴⁶ this is a long-term goal. In the short-term, we can nonetheless work to establish a common coding language among existing CPT codes. This will involve collaborative efforts not only among lymphedema surgeons nationwide, but also with current payers including CMS, the Veteran’s Administration, and commercial insurers.⁹

Limitations of this study include shortcomings of Turquoise Health data, specifically gaps in the data due to incomplete nationwide compliance with the Hospital Price Transparency Rule and lack of rigorous external validation.¹⁴ However, noncompliance has been shown to be relatively haphazard based on a sampling of 100 random and 100 high-revenue hospitals, suggesting that selection bias from differential disclosure based on revenue status is minimal.⁴⁷ It is important to note a few caveats related to price comparisons: multiple LVBs may be performed during a single operation, physiologic procedures are often performed in conjunction with lymphangiography, and surgeons may additionally bill for CPT 69990 (use of operating microscope) with LVB codes. Total reimbursement per case for physiologic procedures may therefore be greater than indicated by a single CPT code. However, due to the use of modifiers that reduce reimbursement for procedures performed in multiples, and the large variation in lymphangiography CPT code reimbursement, it is unclear whether these caveats compensate for the price inequity between excisional and physiologic procedures. A further limitation is that codes utilized in our analysis may be employed in alternative fashions due to coding ambiguity. For example, CPT 38999 (unlisted lymphatic procedure) may be used for VLNT instead of LVB. Since there limits to the number of times a single code may be used per case, a combination procedure with multiple physiologic procedures may see codes, particularly the unlisted code, used in a relatively random fashion. Furthermore, codes (e.g. free omental flaps, blood vessel repair with vein grafts) may be used for indications other than lymphatic restoration, and we cannot control for the bias that this introduces into the study. Regardless of these limitations, our data illustrate the problems associated with ambiguous CPT coding, and provide an impetus to reach a coding consensus that will allow the field of lymphedema surgery to continue to move forward.

Conclusion

Commercial payer-negotiated rates for physiologic procedures varied substantially both within and across hospitals based on coding pattern. Physiologic procedures may be undervalued relative to excisional procedures due to outdated coding and extensive variation. In order to streamline billing, preserve patient access, allocate appropriate value, and track clinical outcomes needed to support widespread acceptance, future efforts should focus on establishing a consistent coding language for lymphedema physiologic and ancillary procedures.

Supplementary Material

Supplemental Table 1

Supplemental Table 1. CPT codes for surgical treatment of lymphedema. ILR, immediate lymphatic reconstruction. LVB, lymphovenous bypass. VLNT, vascularized lymph node transplant.

Supplemental Table 2

Supplemental Table 2. CPT code descriptors for lymphedema surgical procedures. N/A, not available.