Regular article| Volume 20, ISSUE 2, P177-183, March 01, 2018

# The History and Impact of Molecular Coding Changes on Coverage and Reimbursement of Molecular Diagnostic Tests

Transition from Stacking Codes to the Current Molecular Code Set Including Genomic Sequencing Procedures
Open ArchivePublished:December 18, 2017
Changes in coding and coverage generate an uncertain reimbursement environment for molecular pathology laboratories. We analyzed our experience with two representative molecular oncology tests: a T-cell receptor (TCR) β rearrangement test and a large (467-gene) cancer next-generation sequencing panel, the Columbia Combined Cancer Panel (CCCP). Before 2013, the TCR β test was coded using stacked current procedural terminology codes and subsequently transitioned to a tier 1 code. CCCP was coded using a combination of tier 1 and 2 codes until 2015, when a new Genomic Sequencing Procedure code was adopted. A decrease in reimbursement of 61% was observed for the TCR β test on moving from stacking to tier 1 codes. No initial increase in total rejection rate was observed, but a subsequent increase in rejection rates in 2015 and 2016 was noted. The CCCP test showed a similar decrease (48%) in reimbursement after adoption of the new Genomic Sequencing Procedure code and was accompanied by a sharp increase in rejection rates both on implementation of the new code and over time. Changes in coding can result in substantial decreases in reimbursement. This may be a barrier to patient access because of the high cost of molecular diagnostics. Revisions to the molecular code set will continue. These findings help laboratories and manufacturers prepare for the financial impact and advocate appropriately.
The promise of precision medicine—of optimizing and individualizing the treatment or prevention of disease on the basis of a genomic test result—is of potential benefit to many. Molecular diagnostic testing may help in the diagnosis and treatment of malignancies and plays a central role in identifying therapeutic targets or biomarkers that predict response to therapy. Ensuring broad access of patients to molecular diagnostic testing is key to delivering on the potential of precision medicine. However, although there is high demand from patients and clinicians, there are challenges to achieving widespread access to molecular diagnostics. Molecular diagnostic laboratories are faced with significant costs and infrastructure investments when implementing new technologies and complying with regulatory requirements.
• Sabatini L.M.
• Mathews C.
• Ptak D.
• Doshi S.
• Tynan K.
• Hegde M.R.
• Burke T.L.
• Bossler A.D.
Genomic sequencing procedure microcosting analysis and health economic cost-impact analysis: a report of the Association for Molecular Pathology.
This is a reality across the clinical laboratory industry. However, what sets molecular diagnostics apart from other laboratory testing is the uncertainty and lack of consistency in reimbursements for this rapidly evolving and quickly adopted testing, resulting from changes in molecular pathology coding.
• Dervan A.P.
• Deverka P.A.
• Trosman J.R.
• Weldon C.B.
• Douglas M.P.
• Phillips K.A.
Payer decision making for next-generation sequencing-based genetic tests: insights from cell-free DNA prenatal screening.
• Joseph L.
• Cankovic M.
• Caughron S.
• Chandra P.
• Hagenkord J.
• Hallam S.
• Jewell K.E.
• Klein R.D.
• Pratt V.M.
• Rothberg P.G.
• Temple-Smolkin R.L.
• Lyon E.
The spectrum of clinical utilities in molecular pathology testing procedures for inherited conditions and cancer: a report of the Association for Molecular Pathology.
• Messner D.A.
• Al Naber J.
• Koay P.
• Cook-Deegan R.
• Majumder M.
• Javitt G.
• Deverka P.
• Dvoskin R.
• Bollinger J.
• Curnutte M.
• Chandrasekharan S.
• McGuire A.
Barriers to clinical adoption of next generation sequencing: perspectives of a policy Delphi panel.
Molecular diagnostic laboratories use Current Procedural Terminology (CPT) codes to bill payers (Medicare, Medicaid, or private payers) for molecular diagnostic services. CPT codes are developed, published, and annually updated by the American Medical Association and are used to identify and describe medical, surgical, and diagnostic services, facilitating accurate communication with physicians, patients, and payers.
• Hirsch J.A.
• Leslie-Mazwi T.M.
• Nicola G.N.
• Barr R.M.
• Bello J.A.
• Donovan W.D.
• Tu R.
• Alson M.D.
• Manchikanti L.
Current procedural terminology: a primer.
Since the first molecular diagnostic CPT codes were introduced, the code set has undergone a series of changes aimed at clarifying diagnostic services (Figure 1).
The first molecular diagnostic CPT codes were introduced in 1993.
• Logue L.J.
Genetic testing coverage and reimbursement: a provider's dilemma.
These codes described common methods or steps involved in molecular diagnostic testing, such as DNA extraction (CPT code 83890) or enzymatic digestion (CPT code 83892). Because multiple technical steps are commonly used in one molecular test, multiple codes were used to bill one test, a practice known as code stacking, and the codes were collectively called stacking codes. The stacking codes were agnostic of gene or disease. This provided flexibility in test design and development, but was problematic for payers because such a coding system made it difficult to discern which test was being performed, and for what purpose [Association for Molecular Pathology (AMP) CPT Reform Proposal and Foley Hoag white paper, https://www.amp.org/AMP/assets/File/position-statements/2009/AMPCPTReformProposal_Final.pdf and http://www.foleyhoag.com/publications/ebooks-and-white-papers/2011/november/tempest-in-the-melting-pot-genomics-reimbursement-in-2012, respectively, last accessed April 12, 2017].
• Klein R.D.
Reimbursement in molecular pathology: bringing genomic medicine to patients.
In addition, there was concern from payers that laboratories were not incentivized to use the simplest protocols because the more stacking codes used, the more a laboratory may potentially collect.
• Logue L.J.
Genetic testing coverage and reimbursement: a provider's dilemma.
• Klein R.D.
Reimbursement in molecular pathology: bringing genomic medicine to patients.
A two-digit alphanumeric modifier was developed by a College of American Pathologists working group to address this problem. The first digit indicated the disease type (eg, 0, neoplasia and solid tumors/sarcomas; 3, nonneoplastic hematology/coagulation; 8, metabolic and transport), and the second digit indicated a disease or gene (eg, 0A, BRCA1; 3A, factor V; 8A, CFTR). Because these modifiers described specific disease states, they ostensibly would allow accurate identification of molecular tests and allow tracking of test use (AMP CPT Reform Proposal and Foley Hoag white paper, https://www.amp.org/AMP/assets/File/position-statements/2009/AMPCPTReformProposal_Final.pdf and http://www.foleyhoag.com/publications/ebooks-and-white-papers/2011/november/tempest-in-the-melting-pot-genomics-reimbursement-in-2012, respectively).
• Klein R.D.
Reimbursement in molecular pathology: bringing genomic medicine to patients.
The modifiers were approved by the American Medical Association CPT Editorial Panel and were introduced in 2005.
Unfortunately, the modifiers were not widely adopted by either laboratories or payers, and by 2009, the Economic Affairs Committee of the AMP developed a proposal for a new CPT coding framework (AMP CPT Reform Proposal, https://www.amp.org/AMP/assets/File/position-statements/2009/AMPCPTReformProposal_Final.pdf). Recommendations were submitted to the American Medical Association CPT editorial board, which assembled a work group to further develop the codes.
The adopted codes were a mix of analyte-specific and more general method agnostic codes organized into a two-tiered system on the basis of the volume/use of each test; 120 analytes were coded in tier 1 and made up 80% of the total volume of molecular services offered, whereas tier 2 codes were organized into nine levels of increasing technical and interpretative work. Identification of a single germline variant by a restriction enzyme digestion–based analysis, for example, would be a level 1 tier 2 code (81400), whereas analysis of two to five exons by DNA sequence analysis would be a level 5 tier 2 code (81404). These molecular CPT codes fulfilled two important needs: they provided service-specific CPT codes (tier 1) inclusive of all of the analytic steps involved in the test, and they allowed for growth and development by coding the remaining less common and non–analyte-specific services by complexity level (tier 2) (AMP CPT Reform Proposal, https://www.acmg.net/docs/ACMG_2013_Molecular_Pathology_Rate-Setting_Guide.pdf, last accessed April 12, 2017).
• Klein R.D.
Reimbursement in molecular pathology: bringing genomic medicine to patients.
This coding scheme was adopted by the American Medical Association, and the first codes were published in the January 1, 2012, edition of CPT.
Centers for Medicare and Medicaid Services determinations of reimbursement rates for Medicare, one of the largest payers in the country, typically are used to benchmark the rates for other government payers (Medicaid) and for commercial (private) payers; thus, valuation of codes has wide-reaching effects. If the codes were placed on the Physician Fee Schedule, the codes would be valued by the Relative Value Scale Update Committee. If placed on the Clinical Lab Fee Schedule (CLFS), codes may be initially valued by crosswalk (in which payments are benchmarked to an existing code) or by gap fill [in which individual Medicare administrative contractors (MACs) set pricing on the basis of cost, submitted charges, and payment rates from other payers]. In November 2012, Centers for Medicare and Medicaid Services determined that the molecular CPT codes would be placed on the CLFS and would be initially valued by the gap fill process.
Centers for Medicare & Medicaid ServicesDepartment of Health and Human Services
Medicare program: revisions to payment policies under the physician fee schedule, DME face-to-face encounters, elimination of the requirement for termination of non-random prepayment complex medical review and other revisions to part B for CY 2013.
Additional codes have been introduced since 2012 to account for developments in the field of genetics and genomics. Multianalyte assays with algorithmic analyses codes describe tests in which multiple results from assays, such as DNA/RNA analyses, non–nucleic acid analyses (proteins, polypeptides, and carbohydrates), and/or fluorescent in situ hybridization, are analyzed using proprietary algorithms to derive a single result, such as a predictive score. These codes are typically unique to a single laboratory or the producer of such a test.
In 2015, codes for DNA or RNA sequence analyses in which simultaneous interrogation of multiple genes or genetic regions was performed [typically by next-generation sequencing (NGS)] were added under Genomic Sequencing Procedures (GSPs; http://mptrms.mckesson.com/rs/MckessonPT/images/2015CPTChangesPATH.pdf, last accessed April 12, 2017).
These codes encompassed targeted NGS-based panels, whole-exome sequencing, and whole-genome sequencing. Before the introduction of GSPs and multianalyte assays with algorithmic analyses, laboratories were stacking tier 1 and 2 codes to describe the panels they were performing. GSPs and multianalyte assays with algorithmic analyses were an attempt to make more transparent the actual practice of multiplexed molecular tests.
With each new iteration of coding for molecular procedures to better identify the testing performed, everyone using CPT, including payers and providers, obtains a clearer idea of the service being provided. However, with the introduction of any new codes, a new round of valuation is required and potentially results in changes in the valuation of that procedure and possibly a decrease in revenue for laboratories. We hypothesized that the resulting changes to molecular procedures from the original stacking codes to more analyte-specific tier 1 and/or tier 2 molecular pathology procedure codes and from tier 1 and 2 codes to GSP codes have led to a significant decrease in reimbursement for at least some tests performed before and after the new code introduction. The mechanism of this decrease is likely attributable to a lower valuation for the new code when compared with the sum of the individual stacked codes. However, we also hypothesize an increased rate of complete denial of a claim compared with stacking codes, where one or more codes of many may be denied, but the remainder paid. Given the continued need to develop more procedure-specific codes as new tests are developed in molecular pathology, we quantified the effect of changes in coding from stacking to tier 1 and 2 by using our experience with the T-cell receptor (TCR) β rearrangement test and tier 1 and 2 to GSP by using our experience with a >50 gene cancer panel on reimbursement for tests in a large academic medical center's molecular oncology practice.

## Materials and Methods

### Representative Test for Stacking to Tier 1 Transition: TCR β Rearrangement

TCR β gene-rearrangement analysis was selected as a representative molecular test in our laboratory for analysis of the transition from stacking codes to tier 1. The test is performed in our laboratory using a commonly used standard method, PCR with Biomed-2 primers (Invivoscribe, San Diego, CA), followed by fluorescence detection using capillary electrophoresis. From 2011 through 2012, the TCR β rearrangement test was coded using multiple units of the following CPT (stacking) codes: 83891x1, 83900x3, 83901x42, 83907x1, 83909x1, and 83912x1. After 2012, the tier 1 CPT code 81340 was used. The G0452 Health Care Financing Administrators Common Procedure Coding System code for professional interpretation of a molecular test, previously coded as 83912 as a stacking code, is not used by our laboratory.

### Representative Test for Tier 1 to GSP Transition: CCCP

The Columbia Combined Cancer Panel (CCCP) is a 467-gene targeted NGS-based test. This test was first clinically offered in July 2014 and was initially coded using a combination of individual tier 1 and tier 2 CPT codes (81201, 81210, 81235, 81245, 81275, 81292, 81295, 81298, 81310, 81321, 81403, 81407, and 81408). After introduction of the GSP codes in 2015, this test was coded using the CPT code 81455. Charges for CCCP were developed to be consistent and competitive with similar commercially available panel tests. As in the case of the TCR β test, G0452, the professional interpretation code, is not used to bill for this service in conjunction with 81455 by our laboratory.

### Analysis of Reimbursement

All claims submitted between July 1, 2011, and September 30, 2016, for TCR β testing and between July 1, 2013, and September 30, 2016, for CCCP were included in this analysis. Dates of service were within the same time period. Information on the insurance carrier, primary diagnosis [International Classification of Diseases (ICD), Ninth or Tenth Revision], total reimbursement per CPT billed, and total test volume was collected for both TCR β and CCCP tests. Reimbursement is reported as a percentage of charges in all cases in observance of contractual obligations with third-party payers. Carriers were coded into three categories for analysis: commercial plans, government plans (Medicare/Medicaid), and managed government plans (ie, Medicaid health maintenance organizations and managed Medicaid/Medicare plans). Denials are defined as payments of $0 for omnibus codes or totals of$0 payments for stacking codes. All initially denied codes were appealed reflexively. If successful, the value for reimbursement is automatically updated. Therefore, our averages reflect reimbursements after final appeal.

## Discussion

In this study, we show that for two distinct multiplexed tests, there was a 61% and 48% reduction in payments for TCR β rearrangement and CCCP, respectively, when new molecular procedure codes were adopted in our laboratory. For TCR β, the transition was from historical stacking codes to a single tier 1 code and was not accompanied by an immediate increase in complete rejections (ie, $0 payments). The CCCP test, which went from stacked tier 1 and 2 codes to a GSP code, however, was accompanied by a large increase in complete denials (from 20% to 58%) after adoption of the new code. After adoption, both new codes experienced a decrease in reimbursement over time, with increasing total rejections observed consistently with CCCP and after quarter 3 in 2015 with TCR. Our experience with these two tests highlights potential effects resulting from coding changes—decreased valuation and increased denials of coverage. With regard to valuation, the introduction of a new code to the CLFS initiates the determination by Centers for Medicare and Medicaid Services of a new value by either the crosswalk or gap-fill process. • Klein R.D. Reimbursement in molecular pathology: bringing genomic medicine to patients. Payment is set at the lesser of the amount billed, the local fee (for a geographic area), or a national limit. For multiplexed tests moving from stacking codes to tier 1 codes, like the TCR β rearrangement test, the national limitation amount (NLA) for the multiple units of applicable stacking codes was substantially higher (NLA 2012,$1193.54) than the NLA set for the new tier 1 code (NLA 2014 for CPT 81340, $285.02). In the case of CCCP, the sum of the NLA for the stacked tier 1 codes with prices on the CLFS was$2806.28 on the basis of the 2014 CLFS (https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/ClinicalLabFeeSched/Clinical-Laboratory-Fee-Schedule-Files.html, last accessed July 5, 2017). The local Medicare administrative contractor for New York (National Government Services) opted to gap fill the new GSP but did not immediately set a value, which resulted in no payment value for code 81455.
• Sireci A.N.
• Aggarwal V.S.
• Turk A.T.
• Gindin T.
• Mansukhani M.M.
• Hsiao S.J.
Clinical genomic profiling of a diverse array of oncology specimens at a large academic cancer center: identification of targetable Variants and experience with reimbursement.
Although multiplexed tests, such as the TCR β rearrangement test, saw decreased NLAs, pricing for some single-gene tests increased when moving from stacking to tier 1 codes. For example, pricing on the CLFS for BRAF V600 testing, a single-target test, improved from an NLA of $40.78 in 2012 (the sum of stacking codes CPT 83890, 83896, 83898, and 83912) to$179.25, the NLA for the tier 1 codes in 2014 (CPT 81210; https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/ClinicalLabFeeSched/Clinical-Laboratory-Fee-Schedule-Files.html, last accessed July 5, 2017).
From the coverage side, coding changes can potentially lead to an increase in $0 payments and denials. By consolidating codes, payers have the opportunity to completely deny payment for a test with a single denial. When stacked codes are used, although individual codes may be denied, the sum of reimbursements across all codes is more likely to be greater than zero. This was the major driver of decreased reimbursements in CCCP both during the transition and over time for the GSP code. Decreases in reimbursement through periodic cuts to the CLFS value of a code or through decreased coverage can occur even in the absence of any coding changes. In fact, a major decrease was observed in reimbursement of the TCR β code occurring over 2.5 years after adoption of the tier 1 code. This observation highlights another consequence of adoption of these new more procedure-specific codes: these codes are more amenable to nonpayment resulting from local coverage determinations (LCDs). LCDs are the policies drafted by local MACs to define the clinical scenarios (ICD9 or ICD10 codes) under which a CPT code will be paid (https://www.cms.gov/Medicare/Coverage/DeterminationProcess/LCDs.html, last accessed July 5, 2017). In 2016, the local MAC of New York (National Government Services) issued a coverage determination restricting the list of allowable ICD10 codes under which the TCR β CPT would be paid (https://www.cms.gov/medicare-coverage-database/details/lcd-details.aspx?LCDId=35000&ver=49&Date=&DocID=L35000&bc=iAAAABAAAAAAAA%3d%3d&, last accessed July 5, 2017). This had a dramatic and large impact on denials for TCR jumping from 15% to 61% in just one quarter. Similarly, the National Government Services released a noncoverage LCD for code 81455 describing the service as experimental and noncovered for any ICD10 code [https://www.ngsmedicare.com (log-in required), last accessed July 5, 2017]. In our experience, MACs have increasingly taken an aggressive approach on issuing limited or noncoverage LCDs in response to changes in the molecular procedure code set. As the administrators of a limited pool of governmental resources faced with the rapid introduction of costly novel technologies, MACs are under increasing pressure to limit costs and maintain a balanced budget. However, the authors are concerned that these LCDs may reflect a limited understanding of the potential uses of these procedures and result in limited access of these tests for patients. There is an opportunity for organizations, such as the College of American Pathologists and the AMP, to engage government and private payers in collaborative projects for mutual education and to further define or refine the clinical scenarios in which molecular diagnostic testing may show utility. Dramatic and unpredictable changes in reimbursement can have a substantial impact on the finances of a molecular pathology laboratory, particularly when it involves a cost-intensive technology such as NGS. The development, implementation, and maintenance of molecular oncology tests using NGS require an upfront investment of capital and continued infusion of resources to cover specialized personnel, equipment, and expensive reagents. A recent analysis of costs of NGS panels estimated the cost of offering a panel similar to the CCCP test is$1948 per test.
• Sabatini L.M.
• Mathews C.
• Ptak D.
• Doshi S.
• Tynan K.
• Hegde M.R.
• Burke T.L.
• Bossler A.D.
Genomic sequencing procedure microcosting analysis and health economic cost-impact analysis: a report of the Association for Molecular Pathology.
The transition to the GSP code for this test, with the concomitant increase in complete denials (ie, \$0 payments), is a significant loss for our laboratory. Decreased reimbursement for CCCP in conjunction with increased use is unsustainable and presents an effective barrier to access for our patients.
Although coding changes may potentially threaten overall reimbursement for a test through decreased valuation and increased denials of payment, there are advantages to a transition to a single code for both payers and laboratories. Each of the aforementioned changes was driven by the desire to increase transparency in coding (in other words, to more clearly communicate what service is being provided by a code). This allows payers and providers an easier utilization tracking mechanism. In addition, these coding changes led to standardization of coding for a given test across laboratories. Particularly when using stacking codes, laboratory coding strategies varied widely by institution (AMP CPT Reform Proposal and Foley Hoag white paper). Generating more procedure/analyte-specific codes helps make coding less nebulous and more concrete. The resulting code consolidation, which allows for assignment of one code for a given analyte or procedure, may incentivize laboratories to use the most cost-effective method rather than choosing methods for which additional stacked codes might be used.
We present our institutional experience for two molecular diagnostic tests over time and show a significant decrease in reimbursement in the transition between codes and also throughout the lifetime of the new code. The average rate of reimbursement for TCR β has remained steady into 2017 after the abrupt decrease resulting from the LCD described earlier in this section. Reimbursements for CCCP have continued to decrease in 2017. It is likely that the rapid pace of technological development and clinical uptake in molecular pathology testing will lead to additional coding changes in the future as providers seek to most clearly communicate with payers through specific procedure codes. The molecular pathology community must be prepared to weigh the benefits of code clarification against potential financial impact on laboratories, particularly as more GSP codes are adopted to describe costly NGS-based testing. In addition, knowledge of the potential magnitude of impact from the coding changes presented herein may be helpful for molecular laboratories to model into the financial plans to better prepare for disruptions in revenue realization. Finally, device and test manufacturers should be aware of these reimbursement realities and factor them into the costs laboratories face in implementing advanced molecular diagnostic technologies.

## Acknowledgment

We thank Aaron D. Bossler for helpful comments and discussion regarding the manuscript.

## References

• Sabatini L.M.
• Mathews C.
• Ptak D.
• Doshi S.
• Tynan K.
• Hegde M.R.
• Burke T.L.
• Bossler A.D.
Genomic sequencing procedure microcosting analysis and health economic cost-impact analysis: a report of the Association for Molecular Pathology.
J Mol Diagn. 2016; 18: 319-328
• Dervan A.P.
• Deverka P.A.
• Trosman J.R.
• Weldon C.B.
• Douglas M.P.
• Phillips K.A.
Payer decision making for next-generation sequencing-based genetic tests: insights from cell-free DNA prenatal screening.
Genet Med. 2017; 19: 559-567
• Joseph L.
• Cankovic M.
• Caughron S.
• Chandra P.
• Hagenkord J.
• Hallam S.
• Jewell K.E.
• Klein R.D.
• Pratt V.M.
• Rothberg P.G.
• Temple-Smolkin R.L.
• Lyon E.
The spectrum of clinical utilities in molecular pathology testing procedures for inherited conditions and cancer: a report of the Association for Molecular Pathology.
J Mol Diagn. 2016; 18: 605-619
• Messner D.A.
• Al Naber J.
• Koay P.
• Cook-Deegan R.
• Majumder M.
• Javitt G.
• Deverka P.
• Dvoskin R.
• Bollinger J.
• Curnutte M.
• Chandrasekharan S.
• McGuire A.
Barriers to clinical adoption of next generation sequencing: perspectives of a policy Delphi panel.
Appl Transl Genom. 2016; 10: 19-24
1. CPT 2017 Professional Edition. American Medical Association, Chicago, IL2017
• Hirsch J.A.
• Leslie-Mazwi T.M.
• Nicola G.N.
• Barr R.M.
• Bello J.A.
• Donovan W.D.
• Tu R.
• Alson M.D.
• Manchikanti L.
Current procedural terminology: a primer.
J Neurointerv Surg. 2015; 7: 309-312
• Logue L.J.
Genetic testing coverage and reimbursement: a provider's dilemma.
Clin Leadersh Manag Rev. 2003; 17: 346-350
• Klein R.D.
Reimbursement in molecular pathology: bringing genomic medicine to patients.
Clin Chem. 2015; 61: 136-138
• Centers for Medicare & Medicaid Services
• Department of Health and Human Services
Medicare program: revisions to payment policies under the physician fee schedule, DME face-to-face encounters, elimination of the requirement for termination of non-random prepayment complex medical review and other revisions to part B for CY 2013.
Fed Regist. 2012; 77: 68891-69380
2. CPT 2013 Professional Edition. American Medical Association, Chicago, IL2013
3. CPT 2015 Professional Edition. American Medical Association, Chicago, IL2015
• Sireci A.N.
• Aggarwal V.S.
• Turk A.T.
• Gindin T.
• Mansukhani M.M.
• Hsiao S.J.
Clinical genomic profiling of a diverse array of oncology specimens at a large academic cancer center: identification of targetable Variants and experience with reimbursement.
J Mol Diagn. 2017; 19: 277-287