Payer coverage policies for multigene tests (2024)

To the Editor

New technologies are enabling genetic tests that measure multiple rather than single genes. The availability of ‘multigene tests’ (panels and whole exome/genome sequencing tests) is rapidly growing, which raises critical questions about whether and how payers will cover them. Here, we survey the policies of the five largest US private payers that include multigene tests (including 55 policies that cover 313 multigene tests; Table 1, Supplementary Methods and Supplementary Table 1). Our findings reveal that most multigene tests are not covered by payers and that there is a high degree of variability as to how test coverage is assessed. We believe our analysis is the largest systematic review of US coverage policies for multigene tests to date. Efforts to obtain such data and carry out systematic analyses will be essential to increase the transparency and understanding of insurance coverage policies, a fundamental need for informed policy making.

Multigene tests are an integral component of new policy initiatives such as the US National Institutes of Health (NIH) Precision Medicine Initiative and the US National Cancer Institute Cancer Moonshot. For example, experts for the Moonshot Initiative noted that although genome sequencing is rapidly transforming cancer research, only a tiny fraction of cancer patients are having their tumors sequenced. This is because most payers, including Medicare, refuse to pay for the procedure, citing the lack of both clear actionability for the results and evidence that health outcomes are improved as a result of testing¹. These issues are not limited to tumors, but reflect questions about how to best use—and pay for—new genetic tests, which are relatively expensive despite recent decreases in costs. These issues will only intensify as multigene tests are increasingly used and the NIH Precision Medicine Initiative moves forward with the goal of sequencing a large cohort of volunteers.

The rapid expansion in the availability of multigene tests presents challenges regarding how their costs will be covered. As with other new technologies, a key determinant of whether and how multigene tests will be used is insurance coverage². Many questions remain about these tests’ clinical utility, but many providers are ordering them and thus payers are developing relevant coverage policies. However, currently no systematic registries of payer coverage policies for multigene tests exist that examine coverage across a full range of tests and conditions. Previous studies of coverage for genetic tests predate the specific policy challenges presented by multigene tests^3,4. As a result, debates on the topic of multigene test coverage have not been informed by systematic analyses of the current state of insurance coverage of these tests.

With a team of collaborators from multiple institutions (University of California, San Francisco; Tufts Medical Center, Boston; American Institutes for Research, Washington, DC; and the Center for Business Models in Healthcare, Chicago) and funding from the US National Human Genome Research Institute, we have developed an in-house payer coverage policy registry called the University of California San Francisco Center (UCSF) for Translational and Policy Research on Personalized Medicine (TRANSPERS) Payer Coverage Policy Registry. Currently, this registry is not publicly available, but we welcome research collaborations. The registry structure was developed based on a comparative analysis with other registries and input from stakeholders (Supplementary Methods). In this context, we define multigene tests as tests that analyze multiple genes by next-generation sequencing or chromosomal microarray analysis, with the resulting test report providing multiple results, not an algorithmic score. The registry includes data on what multigene tests are included in policies, whether multigene tests are covered (“medically necessary”) or not covered (“experimental and/or investigational”), and the evidence cited and rationales for coverage decisions. We report here on coverage across payers and conditions; other analyses have examined coverage for panels including BRCA1/2 (ref. 5), coverage of non-invasive prenatal screening tests⁶, and the evidence cited in coverage decisions⁷.

Our initial version (Version 1.0) of the registry includes publicly available national policies from the five largest US private (commercial) payers, representing 112 million enrollees⁸, though these policies do not necessarily reflect regional polices pertaining to particular benefit designs of states, unions, or individual private companies. (To our knowledge, it would be unlikely that such policies would single out multigene tests for separate coverage.) Policies current as of June 2015 were systematically coded by two authors (M.P.D. and P.A.D., with review by a third author K.A.P.) (Supplementary Methods). As noted previously⁹, policies are context-specific and thus we classified tests into the following categories: tumor profiling; inherited disease testing for neurologic, cancer, cardiovascular, or biochemical disorders; drug metabolism testing; whole exome or whole genome sequencing; and prenatal testing or carrier testing (Table 2). Policies discuss multigene tests using either a general description for the type of panel (e.g., “cancer susceptibility testing”) or a specific brand name (e.g., “BreastNext”). Tests may be included (‘mentioned’) or not included in a policy. When included, tests can be determined to be covered or not covered. Thus, some tests may have unknown coverage because they are not included or mentioned in a policy. Tests were placed into one of three mutually exclusive categories.

1. Covered when mentioned. Test is covered for at least one indication in that policy and not specifically excluded in any other policy. Multigene tests within the same policy that were covered for some clinical indications but not others were coded as “covered.”

2. Not covered when mentioned. Not covered in any policies that specifically mention the test.

3. Mixed coverage. Explicitly covered in at least one policy but not covered in at least one other policy.

We did not attempt to combine tests by, for example, combining brand name tests with the comparable general test descriptors. Rather, we listed test names as they are listed in the policies. Policies may list the general test name or the brand name or both (Table 2). Therefore, because test names listed in policies can be duplicative, we did not calculate the percentage of multigene tests covered.

Fifty-five policies include multigene tests (n = 313 tests; Table 1). Of these policies, 27% included tests across multiple test categories. The remaining policies focused on specific test categories, with the most common being policies focusing on inherited cancer risk assessment (20%), prenatal testing (13%), and inherited cardiovascular risk assessment (11%) (Supplementary Fig. 1). The number of tests included in each policy ranged from 1 to 27.

We found that 51% of policies covered none of the multigene tests specifically noted in the given policy, 22% covered all of the tests in the policy, and 27% covered some but not all of the tests in the policy (Table 1). Across policies, most test categories had multigene tests that were covered as well as tests that were not covered (Table 2 and Supplementary Table 2). The exceptions were multigene tests for drug metabolism, whole exome sequencing, and whole genome sequencing, which were not covered in any policies.

Analyses of coverage patterns (Table 2 and Supplementary Table 2) indicate that covered tests are typically those for specific conditions, genes, and populations where there is a base of synthesized evidence supporting genetic testing for that condition (e.g., guidelines), for example, tests for inherited neurological and cardiovascular conditions and prenatal tests are more often covered. For instance, targeted tumor profiling multigene tests for specific cancer types (myelodysplastic syndromes and non-small cell lung cancer tests with 5–50 genes) are covered by some payers; multigene tests specific to Lynch syndrome, hypertrophic cardiomyopathy, and long QT syndrome are often covered by payers for specific populations. Conversely, tests that are not covered are typically those for broad indications and/or tests that include large or undefined numbers of genes (e.g., multi-syndromic cancer risk testing panels).

We find particularly interesting those tests that are covered in at least one policy but not covered in at least one other policy (Table 2, last column). Testing categories that exhibited coverage variation were tumor profiling (n = 1 test), inherited disease testing for neurological disorders (n = 6 tests), inherited disease testing for cardiovascular disorders (n = 11 tests), inherited disease testing for biochemical disorders (n = 3 tests), and prenatal testing (n = 5 tests). As an example, multigene tests for inherited breast and ovarian cancer risk and tests for hypertrophic cardiomyopathy are covered by some payers but not by others.

Supplementary Table 2 illustrates variation in how similar multigene tests were listed and covered in policies. For example, in one policy a panel to determine whether a thyroid nodule is benign or malignant was covered as a brand name test (ThyGenX), whereas in another policy it was not covered when listed as a general test (“analysis of thyroid nodule fine needle aspiration using multigene tests”), although the brand name test can be assumed to fall into this general test category. Multigene tests for the general category of “hereditary hearing loss” are covered in all policies that mention them, but the more specifically listed tests (“non-syndromic hearing-loss testing multigene tests using multigene NGS [next-generation sequencing] sequencing tests”) are not covered in all policies.

The results in Supplementary Table 2 also suggest which test characteristics contribute to coverage decisions. One distinguishing characteristic is the number of genes included and the scope of the test (e.g., one payer covers targeted solid organ genomic sequencing multigene tests to test for non-small cell lung cancer when the panel includes 5–50 genes, but not a pan-cancer test that includes 315 genes (FoundationOne)). Another distinguishing characteristic is the population being tested. For example, although a panel, such as MaterniT21 (for non-invasive prenatal screening of pregnant women for fetal aneuploidies) is in the “covered” column, this coverage is for high-risk pregnancies and not necessarily for all indications (e.g., average risk pregnancies).

In summary, whereas some multigene tests are covered by payers, most are not covered, and how tests are addressed in policies varies. Our results are consistent with previous studies showing that genetic tests as a group are often not covered^3,4. However, the growing use of multigene tests appears to have prompted an increase in the number of policies covering such tests (e.g., a 2012 review found 41 policies on genetic tests as a whole from ten payers³, whereas we found 55 policies related specifically to multigene tests from only five payers).

The variation in how tests are addressed is unsurprising, given the complexities in multigene testing. Although payers evaluate multigene tests similarly to single-gene tests, using a framework based on analytical and clinical validity and clinical utility, we found in prior work that coverage policies for multigene tests are inherently more challenging for payers than policies for single-gene tests. Multigene tests do not fit the standard testing paradigm where a single (validated) marker provides a single (clinically useful) test result. For example, 80% of payers interviewed stated that the inherent characteristics of tumor sequencing, which provides multiple results and includes novel and often unvalidated markers, challenge the standard definition of ‘medical necessity’ required for coverage¹⁰. Although the intent of our study was exploratory research, and we did not test hypotheses, we found that several test characteristics—the number of genes included, the scope of the test, and the population being tested— appear to be associated with coverage and will thus be examined in future research.

These findings have important policy implications. First, ongoing, systematic analyses of coverage policies are needed to provide objective evidence to improve understanding of policies and payer decision-making. Creating a registry, such as USCSF TRANSPERS Payer Coverage Policy Registry, can provide an objective unbiased research tool for organizations (e.g., academics, test developers, payers) to examine coverage policies across conditions or tests, with an ultimate goal of facilitating greater transparency and predictability in payer coverage decision-making as new technologies emerge^10–13 (see Supplementary Table 3 for a list of other resources). Given the increase in multigene test use and the importance of coverage for successful adoption of the tests and equitable access, it will be important to continue to systematically review coverage policies as they evolve. Policies often change and our data only reflect policies as of June 2015. Since then, several changes in coverage have already occurred: a payer now covers Foundation Medicine’s (Cambridge, MA, USA) tumor profiling panel for patients with metastatic stage IV non-small-cell lung cancer¹⁴, and another payer has issued a policy covering whole exome sequencing when a patient meets all of a detailed list of criteria¹⁵.

A second policy implication is that registries and other types of structured reviews are needed that systematically summarize coverage policies, given that these policies are complex and thus difficult to assess without a structured database. Our findings illustrate how payers present information differently in their policies; for example, payers use general test names as well as brand names, which make it difficult to determine exactly what tests are covered. Similarly, because payers often add multigene test panels to preexisting coverage policies, policies often include a range of genetic tests in the same policy. Another complexity is that policies do not always specify the genes or variants included, further complicating the assessment of test coverage. Lastly, the test names listed in policies are not comprehensive, as policies list only a subset of the total number of available multigene tests. A registry can be used to address these complexities by identifying evidence gaps, illuminating variation in payer policies and serving as the baseline (with updating on an annual basis) for future analyses of coverage-related research questions^10–13. Of particular interest for future research will be examining the impact of the recent change in CPT (current procedural terminology) coding policies on the use of specific, unique test codes for multigene tests, rather than ‘code stacking’ (the use of multiple CPT codes for reimbursem*nt of a genetic test, which can result in wide variation in the charge for the same test in different laboratories and often prevents payers from understanding what exactly has been tested).

Our initial analyses have several limitations in terms of the size of our sample, lack of detail in reasons for coverage and on the specifics of individual multigene tests, the inherent limitations of published policies (which do not discuss all of the factors that determine coverage and which—due to test complexity—require judgment in interpreting the data) and information as to whether tests may be reimbursed even when not formally covered (coverage policies do not dictate payment). Going forward, we intend to add more private payers as well as regional and public payers to our data set; add variables that will enable detailed analyses of reasons for coverage decisions; conduct in-depth analyses of specific multigene tests; and update the results and analyzing policies over time.

In conclusion, coverage and reimbursem*nt of new genetic tests has often been cited as a key requirement for their adoption. Our study provides the first systematic review focusing on coverage of multigene tests to better inform discussions about this issue.

Supplementary Material

Acknowledgments

This work was supported by a grant to K.A.P. from the National Human Genome Research Institute (R01HG007063-S1). We are grateful to C. Liu (Albert Einstein School of Medicine) for coding help.

Footnotes

References