Mr. Glenn McGuirk
Acting Director, Hospital and Ambulatory Policy Group Centers for Medicare & Medicaid Services
Center for Medicare
7500 Security Boulevard
Mail Stop C4-01-26
Baltimore, Maryland 21244
Re: Centers for Medicare and Medicaid Services Calendar Year 2014 New and Reconsidered Clinical Laboratory Fee Schedule Test Codes and Preliminary Payment Determinations
Dear Mr. McGuirk:
On behalf of the American Clinical Laboratory Association (“ACLA”), we are writing to express our disappointment in the Centers for Medicare and Medicaid Services’ (“CMS’s” or “the agency’s”) continued position, as stated in the preliminary payment decisions for new and reconsidered Clinical Laboratory Fee Schedule (“CLFS”) test codes for CY 2014, that it would not recognize the multianalyte assays with algorithmic analyses (“MAAA”) codes for payment purposes under the CLFS.1
ACLA is an association representing clinical laboratories throughout the country, including local, regional, and national laboratories that provide millions of clinical diagnostic laboratory services for Medicare beneficiaries each year. As a result, ACLA members will be affected directly by CMS’s decisions to continue not to recognize the MAAA class of diagnostic tests.
As we stated in our comments to the CY2013 CLFS preliminary payment determination, we believe that it is wrong for CMS to establish a blanket payment policy for all MAAA tests, and strongly disagree with the agency’s decision not to price the MAAA tests. Before setting a payment policy, it is important for the agency to recognize the function and purpose of the algorithms that are essential to the MAAA tests and why it is not possible to dissociate the algorithms from the so-called “underlying clinical laboratory tests.” ACLA disagrees that the Medicare program is precluded from paying for the MAAA tests, and it urges CMS to reconsider its position, and to permit contractors to continue to price and pay for these clinical laboratory tests, as they have for years. Until now, Medicare contractors and private payers alike have recognized the value of MAAA tests, including their potential to save money by avoiding trial and error on selecting best therapy for patients, and have reimbursed providers for them according to their value. CMS’s continued blanket position on the MAAA test codes could have a disastrous impact on personalized medicine.
I. Multi-analyte Assays with Algorithmic Analyses
ACLA is deeply concerned about CMS’s continued position that the codes describing MAAAs would not be priced.2
In last year’s CLFS preliminary payment determinations, CMS essentially proposed that the codes would be “inactive” and that only the “underlying clinical laboratory tests on which the MAAA is done” would be paid.3 The agency’s stated rationale at that time was: “Medicare does not recognize a calculated or algorithmically-derived rate or results as a clinical laboratory test since the calculated or algorithmically-derived rate or result alone does not indicate the presence or absence of a substance or organism in the body. Medicare uses other codes for payment of the underlying clinical laboratory tests on which the MAAA is done and we continue to recommend not separately pricing the codes.”4 The agency stated that “CMS uses other codes for payment of the underlying clinical laboratory tests on which the MAAA is done,” but the physician orders the MAAA test – not the “underlying clinical laboratory tests” – and in most instances, the underlying clinical laboratory tests are not separable from the MAAA algorithm.
In this year’s CLFS preliminary payment determinations, CMS reiterated its position that “[an] algorithm is not a clinical diagnostic test” and because “[the] CLFS only pays for clinical diagnostic laboratory tests,” the MAAA codes “are not payable under CLFS.”5 Interestingly, however, CMS stated in the preliminary determination and during the CLFS Public Meeting in July, that they will “continue to consider each individual test that comprises a MAAA code on its own merits.”6
However, the preliminary determination published by CMS fails to honor this test-by-test evaluation by rejecting the entire code of both the 2013 reconsidered MAAA codes as well as the new 2014 MAAA codes without providing any test-specific rationales for not recognizing the individual tests. We request, therefore, that CMS provide a specific rationale for each individual MAAA code which the agency intends not to recognize for Medicare payment.
A relatively small number of these codes were being considered in 2014, and the tests represented by the codes differ in their methodologies and approaches. We do not believe that it is prudent for CMS to propose a broad payment policy this year that would apply to all current and future MAAA tests. Rather, CMS should direct Medicare contractors to continue to price and pay for the few MAAA tests for 2013. It is important, however, for the agency to understand what MAAA tests are – and what they are not – when considering a payment approach for this burgeoning category of clinical laboratory tests.
A. Background on MAAA Tests
The American Medical Association’s (“AMA’s”) CPT Code Manual recognized distinct MAAA codes for the first time in 2013, although these clinical laboratory tests themselves are not new. The AMA has described MAAAs this way: “[MAAAs] are procedures that utilize multiple results derived from assays of various types, including molecular pathology assays, fluorescent in situ hybridization assays, and non-nucleic acid based assays (e.g., proteins, polypeptides, lipids, carbohydrates). Algorithmic analysis, using the results of these assays as well as other patient information (if used), is then performed and reported typically as a numeric score(s) or as a probability…[MAAA codes] encompass all analytical services required for the algorithmic analysis (e.g., cell lysis, nucleic acid stabilization, extraction, digestion, amplification, hybridization and detection), in addition to the algorithmic analysis itself.”7
As the AMA’s description demonstrates, MAAA tests vary greatly in their methodologies, in the types of underlying tests to which they are applied, and in the independent value to a physician of the results of any one of the “underlying tests.” An algorithm may be applied to the results of a genetic test performed on a tissue sample to determine which of a number of genes show mutations, or an algorithm might be applied to the results of a series of blood or chemistry tests. A physician may understand the implications of one or more of the test results to which a MAAA algorithm is applied, but most often, the results without the algorithm have no meaning to a physician. The MAAA tests are not monolithic, and therefore CMS should proceed cautiously when considering whether to establish broad payment policies for them.
The codes representing the MAAA tests are alike in that they encompass all analytical services performed by a clinical laboratory up to and including the algorithmic analysis. A physician orders a MAAA test itself; the physician does not order each individual service or expect to receive the results of each individual test. Rather, it is the single algorithmically derived final result that is useful to a physician and that a laboratory reports to a physician. For example, a physician may order a MAAA test that involves the analysis of multiple genes for mutations, and the application of that algorithm determines whether a treatment is likely to have therapeutic value. The individual results of each gene’s analysis would be of little value to the physician in the context in which the MAAA test is ordered – it is the probability score derived from the algorithm that drives the physician’s decision-making. In short, the MAAA tests are far more than the sum of their parts.
It is important to acknowledge that the so-called “underlying tests” to which the MAAA algorithms are applied are not the same as organ- or disease-oriented panels. Those panels are a collection of tests, ordered together. The individual results of each of the test are reported back to a physician, and generally, the physician is capable of making an independent judgment about a patient’s condition or prognosis based on the results of any or all of the panel’s tests. In contrast, a physician orders the MAAA test, not the “underlying tests.” The MAAA tests include a range of tests that may have little or no independent value with respect to the conditions being tested. With few exceptions, physicians cannot analyze the results of the MAAA’s “underlying tests” to reach accurate conclusions about their patients’ prognoses or susceptibility to therapeutic treatments. The clinically-validated results of the MAAA test are what provide the actionable intelligence about a patient’s prognosis or condition.
The MAAA tests represented by the new codes are vital to the development of personalized medicine, which allows health care providers to target care and treatment based on a person’s individual genetic makeup. In other words, personalized medicine helps physicians select “the right treatment for the right patient at the right time.” As FDA Commissioner Dr. Margaret Hamburg and Director of the National Institutes of Health Dr. Francis Collins wrote in the New England Journal of Medicine, “The success of personalized medicine depends on having accurate diagnostic tests that identify patients who can benefit from targeted therapies…Real progress will come when clinically beneficial new products and approaches are incorporated into clinical practice.”8 The development and clinical validation of a MAAA algorithm typically takes several years and a significant investment of resources on the part of the developer, sometimes tens of millions of dollars. Ultimately, however, these tests save money, because health care providers will select the best option first, reducing the time and money otherwise required by a trial and error process for selecting effective therapies for patients.
One example of a MAAA test included in those being priced for 2013 is the OVA1TM test, which helps a physician assess the likelihood that an ovarian mass is malignant and determine the course of treatment most likely to be successful for a patient. The OVA1TM test is a qualitative serum test that combines the results of five distinct immunoassays into a single numerical result. A higher score equates with a higher probability of malignancy and a greater need for a patient to be referred to a gynecologic oncologist for proper treatment. It greatly increases the chances that a non-gynecologic oncologist can detect malignancies, and it also is highly accurate for identifying women with no malignancies, resulting in fewer unnecessary referrals and complicated surgeries. (A fuller explanation of the OVA1TM test is attached as Exhibit A.)
Another example is the PreDX® Diabetes Risk Score (“DRS”), which is a multi-marker fasting blood test that assesses markers of inflammation, fat cell function, and glucose metabolism, and categorizes individuals as low, moderate, or high risk for conversion to diabetes within five years. The PreDx DRS has been shown to be more accurate than HbA1c or fasting glucose in predicting incident diabetes. The improved risk assessment provided by DRS is a function of changes in the seven markers (glucose, HbA1c, insulin, hs-CRP, adoponectin, ferritin, and interleukin 2-receptor alpha) that are representative of the multiple pathways that are dysregulated in the development of diabetes. It has been shown that PreDx DRS can be used to identify patients at risk for diabetes who are most likely to benefit from appropriate medical or lifestyle intervention, to reduce their risk of progressing to diabetes over the long term, and to monitor and potentially improve treatment outcomes.
Against this background of the MAAA tests, we address CMS’s assertion that the Medicare program cannot pay for the MAAA tests and the agency’s proposal that the MAAA codes would be inactive and that codes representing the “underlying tests” could be used, instead.
B. Medicare Is Permitted to Pay for MAAA Tests
The Social Security Act and implementing regulations do not contain CMS’s cited limitation on payment for algorithmically-derived results. CMS has not provided a citation for its assertion that a test that does not indicate “the presence or absence of a substance or organism” in the body is not recognized by the Medicare program and payable under the CLFS. ACLA has been unable to determine the source of the statement.9 There is no such definition of a clinical laboratory service in federal law. Further, Medicare does, in fact, pay for tests that do not indicate “the presence or absence of a substance or organism,” such as functionality tests, sensitivity tests, time measurements, and concentration measurements.
C. Laboratories May Not Simply Submit a Claim for a MAAA Test Using Codes for “Underlying Tests”
CMS’s proposal that laboratories seek payment for the underlying clinical laboratory tests on which a MAAA is performed does not comport with the way MAAA tests are used and performed, and a laboratory would violate Medicare law in doing so. A physician who orders a MAAA test typically does not order the “underlying tests” – he or she orders the MAAA test. Per Medicare billing rules, a clinical laboratory may submit a claim only for a test ordered by the beneficiary’s treating physician.10 Thus, if a physician does not order the “underlying tests,” a laboratory would not be permitted to bill the Medicare program for them. Furthermore, physicians rarely receive the underlying DNA, RNA, or protein measures to which a MAAA algorithm is applied to derive the score, because, generally, there is no value to the physician of those results alone in the context in which the MAAA test is ordered (e.g., deriving a probability of malignancy or recurrence). Performance of the “underlying tests” in the first instance is for the sake of applying the MAAA algorithm, not because of their independent value to the treating physician.
D. Medicare Does Currently Recognize and Pay For Calculated and/or Algorithmically-Derived Results
CMS already pays for other such tests with algorithmically-derived results. For example, in 2005, CMS began reimbursing providers for the HIV bioinformatics code (CPT code 87900, Infectious agent drug susceptibility phenotype prediction using regularly updated genotypic information). This describes a method of determining individually specific and effective drug treatment regimens based on a patient’s specific viral load response by applying a predictive model of drug resistance or susceptibility. The AMA’s description of the code states, “The prediction of phenotypic behavior derives from comparison of the genotypic patterns of the patient with a large relational database of actual phenotypic and genotypic information that is continuously updated with recent clinical isolates representing the changing nature of the pandemic.”11 Without the HIV bioinformatics information represented by CPT code 87900, the “underlying test” to which the information is applied would have little value for a physician who is trying to determine which treatment regimen will work for a specific patient at a specific time. CMS has recognized – rightly – the value in the complex calculation represented by CPT code 87900.
As you may know, several Medicare contractors already have determined that payment for MAAA tests is appropriate. In setting payment levels, the contractors have looked at a variety of information, including payments received from private payors, the potential savings to the Medicare program due to proper test utilization, and what Medicare would pay for tests that require similar resources (independent of the algorithm). Also, several laboratories currently are in discussions with contractors about payment for different MAAA tests.
Like the HIV bioinformatics code for which the Medicare program already pays, MAAAs are complex calculations with independent predictive value. ACLA recognizes that the Medicare program does not pay health care providers separately for simple calculations, such as the calculation of a patient’s low density lipoprotein (“LDL”) derived from total cholesterol, high density lipoprotein (“HDL”), and triglycerides, which can be calculated in a physician’s office with a pocket calculator. No specialized training is required for such a calculation, and a health care provider does not need to make any investment of time, money, or other resources into developing the calculation. MAAAs, in contrast, are more like the tests with algorithms for which CMS already pays. They weigh numerous variables to arrive at a score, and the relationship between and among the variables are not widely known and must be validated through expensive clinical trials.
E. A Decision to Continue Not to Recognize MAAA Codes for Payment Purposes Would Negatively Impact Their Development
Dr. Hamburg, Dr. Collins, and many others have recognized the tremendous promise of personalized medicine. However, a broad payment policy not to pay for the MAAA codes could thin the ranks of potential developers of the tests. It would be far too risky for most laboratories to invest years of research and millions of dollars to develop a test that may not be paid by Medicare. Additionally, many commercial contracts instruct laboratories to submit claims in the same way claims are submitted to the Medicare program; the Medicare program’s failure to recognize the MAAA codes could eliminate payment by commercial payors. Until now, the Medicare contractors and private payors alike have recognized the value of MAAA tests and have reimbursed providers for them according to their value. CMS’s continued position on the MAAA test codes would be another step in the wrong direction and could have a disastrous impact on personalized medicine.
F. CMS Should Evaluate Each MAAA Test on its Own Merits, Not Establish a Broad Payment Policy for All MAAA Tests
CMS stated in the CY2014 CLFS preliminary determination and during the CLFS Public Meeting in July 2013 that they will “continue to consider each individual test that comprises a MAAA code on its own merits.”12
However, the preliminary determination published by CMS fails to honor this test-by-test evaluation by rejecting the entire code of both the 2013 reconsidered MAAA codes as well as the new 2014 MAAA codes without providing any test-specific rationales for not recognizing the individual tests. We request, therefore, that CMS provide a specific rationale for each individual MAAA code which the agency intends not to recognize for Medicare payment. ACLA agrees that tests should be reviewed based upon their individual merits and that CMS should reflect such review in its statement supporting each determination. The agency’s blanket statement applying to all MAAA codes does not provide this test-by-test individual assessment and rationale supporting the CMS’s determination.
For several years, Medicare contractors have priced the tests now described as MAAAs on a case-by-case basis and generally have developed fair prices for the tests. In general, they have looked at some of the same factors considered in a gapfilling process: charges by laboratories, rates paid by other payors, resource use, and the inherent value of a test to patient management. ACLA recommends that, for the purpose of payment determinations, CMS should continue to defer to Medicare contractors that have expertise with MAAA tests, and the tests should continue to be paid under the CLFS as laboratory tests.
Thank you very much for your consideration of ACLA’s comments. We look forward to discussing this with you further and to working with CMS on these important issues.
President, American Clinical Laboratory Association