NEW YORK – After analyzing millions of genetic testing claims, health technology company Concert Genetics has produced hard data showing that that there is little standardization in the way labs bill commercial payors for tests, particularly multi-gene panels.
The variability, described in a report released this week, may not only contribute to overspending for insurers, but it also creates a restrictive reimbursement environment for labs. Perhaps most importantly, inefficiencies in the payment system hinders patient access and makes it almost impossible to discern the true cost of these tests at a time when patients are being asked to shoulder a greater share of their healthcare costs, Concert executives said.
"Coding consistency is an absolutely necessary precursor to price transparency and being able to tell patients how much stuff actually costs," said Gillian Hooker, VP of clinical development at Concert. "If you don't know how it's going to be coded, you can't tell them how much something is going to cost. More than that, you can't tell them whether it can be covered or not because it all depends on how it's going to be coded."
In an effort to bring more clarity and predictability to this process, the firm has also developed software, called the Concert Coding Engine, that relies on a logic set or guiding principles to standardize how tests are coded and how labs bill for them. A more predictable and streamlined process would not only reduce administrative burdens on labs and payors, but in some cases it will also result in reduced test pricing and revenues for labs –- a risk they will undoubtedly consider when deciding whether to adopt Concert's software.
The lack of transparency around how labs bill payors for genetic tests has been an enduring problem for the field. A few years ago, as payors saw their spending on genetic testing increase, they became increasingly concerned that labs were stacking older CPT codes to inappropriately bill for tests. In response, the American Medical Association, the developer of CPT codes, released a new system for coding genetic tests –- Tier 1 codes described the analyte gauged by a test and Tier 2 codes described the base pairs or exons analyzed.
This new system, however, didn't provide the level of granularity that payors needed to discern which tests they were paying for, because it still allowed labs to stack codes when billing for tests. In recent years, the problem has been exacerbated by the increased use of multi-gene testing panels, since labs can stack codes to describe the different analytes gauged and methods used.
The fact that labs bill for their genetic test panels differently and that this variability increases costs is not news to commercial payors. "When we get a claim from a provider or a lab, we'll have upwards of 20 to 50 codes on there, and some of these codes aren't specific to what gene they're looking at, particularly the Tier 2 codes," said Henry Garlich, director of Blue Shield of California's healthcare value solutions and enhanced clinical programs. "That creates a problem with the appropriate use of these tests because we don't know what we're paying for, we don't know what tests these individuals are having, and we don't know the clinical utility of these genes they're looking at."
Gene panels are a rapidly growing segment of the genetic testing industry. Concert last year estimated that there were 74,000 genetic tests on the market and 14 new tests launched daily. Around 15 percent of new test launches were panels, which entered the market at a rate of more than 15 tests per week. In total, Concert estimated there are nearly 9,500 panel tests on the market.
Blue Shield of California's internal data suggests it is spending roughly $30 million to $40 million for genetic test panels that are performed in clinical situations, which don't meet its coverage criteria and would be considered inappropriate use. According to Garlich, panel tests are currently overused in the context of hereditary cancer risk assessments, in gauging somatic cancer variants to predict treatment response, in non-invasive prenatal testing (NIPT), and for rare diseases.
Although the AMA more recently began issuing proprietary laboratory analysis codes that can uniquely identify tests performed by specific labs, making it easier for insurers to identify tests and track their utilization, Garlich suspects that many labs haven't applied for these codes precisely for this reason. "If you use a single code, then it's easier for a health plan to use the appropriate use criteria within a medical policy to deny claims or pay for it at a rate the health plan thinks is fair market value," Garlich said. "Of course, the labs always think their tests are worth a lot more than what they truly are."
Individual payors are trying to control their molecular diagnostics spending through other mechanisms, for example, by hiring lab benefit managers –- as UnitedHealthcare and Anthem have done –- to police inappropriate billing and utilization, or by specifying codes in contracts with labs. Blue Shield of California, for example, has been addressing this problem via a process called "case coding," which requires the insurer to come to a contractual agreement with individual labs as to how to code and price specific genetic tests, including multi-gene panel tests. The lab's ID and this agreed upon code are then linked within Blue Sheild's claims processing system, so when the lab bills this code, it will be paid the contracted price.
This process allows Blue Shield to pay for panel tests but set pricing based on the genes that meet its evidentiary criteria for coverage. But case coding with each lab around individual tests and genes is labor intensive, Garlich noted, and because the method is proprietary to Blue Shield, it may not be used with other payors.
Indeed, insurers' current approaches lack transparency and haven't resolved the coding conundrum in a standardized way for all stakeholders. "There's no single entity that is in the position of reviewing or standardizing those particular methodologies. Nobody, not a payor, not one lab, not any entity has perfect visibility into what tests are on the market, how those tests are changing, and how to consistently apply these codes," said Concert CEO Rob Metcalf.
In an effort to characterize the extent of the problem, Concert Genetics analyzed 2.2 million genetic testing claims representing 35 million lives from 2016 to 2018, and unpacked the code signature underlying these claims. The company then mapped the codes back to a specific genetic testing unit, a category, or a domain. "When a claim comes in, it can be extremely hard to know what that claim represents: the specific test or the category of test or the domain can be difficult to discern," said Metcalf.
Concert describes each orderable test as a "genetic testing unit." A category comprises tests with the same clinical indication (i.e. hereditary breast and ovarian cancer panels or Noonan syndrome panels). Domains group tests into larger descriptive buckets such as NIPT, hereditary cancer tests, oncology tumor panels, and pharmacogenetics.
Looking across more than a dozen domains, Concert reported an average of between 2 and 18.5 codes per claim. "What our paper is reporting is that ... most tests across domains are billed with more than one code," Hooker said.
Across these same domains, Concert tracked significant variation in the combinations of codes labs billed. For example, there were nearly 10,000 distinct code combinations for oncology tumor panels, around 4,300 code combinations for pharmacogenetic tests, around 2,200 code combinations for hereditary cancer test panels, and around 650 code combinations for NIPT.
Metcalf explained that from the perspective of a payor, Concert's findings illustrate that the stream of information they're receiving on test utilization for billed codes is constantly changing. This means that payors are spending a lot of time trying to figure out what test was rendered and how they should apply medical policy, he said.
Concert also estimated how this coding variability impacts reimbursement by looking at the range of payments insurance companies made to labs for the tests they billed. Assuming an average test price of around $1,000 across all claims, the standard deviation for payment of carrier status panels and oncology tumor panels were the highest, at around $2,000. Payment for oncology algorithmic assays varied by approximately $1,600, hereditary cancer panels by more than $1,200, PGx testing by around $1,100, and NIPT by around $900.
Although differences in test quality, service, and brand impact pricing, Concert wrote in the paper that its analysis suggests that "highly variable, multi-code claims likely contribute to variation in payments and higher costs." More specifically, the pricing variability for genetic test claims captured in this report points to an inefficient molecular diagnostics market, according to Metcalf.
"These are some pretty large standard deviations, which suggest that there's a lot of variation in pricing," he said, adding, for example, that customers buying a box of diapers wouldn't expect the price to vary significantly based on the brand or the store.
Concert is hoping to bring more predictability to the process via its Coding Engine, which is a software application that draws on a database of genetic tests and specific test attributes, and maps them to specific codes according to expert-curated guidance. The aim, according to Concert, is to ensure that there is one way to code each test on the market.
In the process of developing the Coding Engine, Concert reached out to numerous labs and payors to better understand this mutli-faceted problem. What became clear to Metcalf after speaking with stakeholders is that while labs may be billing using a variety of codes, leaders at different health plans also weren't on the same page about how tests should be coded.
"Should the codes reflect the test that was actually rendered? Should the codes reflect what portions of those tests were covered under medical policy? Should they reflect what was in a lab-to-plan contract? Or was it entirely at the discretion of the laboratory or some other entity?" Metcalf posited.
In building its Coding Engine, Concert starts by accepting the basic premise that there is a single, repeatable way to code each test on the market, and that coding framework should promote simplicity over complexity whenever possible. In the case of multigene panels, for example, Concert's software would favor panel codes over stacking single-gene codes.
The company plans to launch the first version of the Coding Engine for molecular diagnostics. According to a report describing the solution, the Coding Engine has already streamlined codes for this segment of the testing market, paring down 36,000 coding combinations observed in one large genetic testing claims database to around 1,300 unique code combinations.
Concert's coding solution will be free for labs to use to code their own tests, while insurers can purchase annual licenses. Garlich is familiar with the solution Concert is trying to advance and believes it could be useful for all stakeholders, not just insurers. "It'll make it easier for labs, health plans, and doctors to get the efficiency they need to get tests covered and have more clarity around what tests are utilized," he said.
However, it remains to be seen whether labs will embrace Concert's solution, since it will likely reduce test pricing and labs' revenues, in some cases. "There is some degree of tradeoff," Metcalf said. "Transparency in coding will narrow variation in price. It will also streamline the reimbursement process, which has benefits for laboratories as well."
What Concert is proposing would undoubtedly be a disruptive and dramatic shift from the way tests are coded now. In order to accept such a solution, labs would have to take a long-term view of the whole market and accept that pricing variation will inevitably narrow as the use of genetic testing becomes more and more integrated in healthcare.
"[If] you can't easily translate between the test and the code and the reimbursement, you can't learn from the data systemwide in a way that we absolutely have to to ultimately provide better care for the patient," Metcalf said. "[Concert's coding solution] is likely to have an impact on certain laboratories and an impact on pricing but those are all things that are, in the long run, best for the ecosystem of personalized medicine."
Metcalf is hopeful that there will be a group of labs that see the overall benefits of a standardized approach to coding genetic tests and that will agree to use the coding solution Concert has created. Invitae is among lab industry players that would support the long-term, systemic gains of a more predictable coding system over short-term pricing gains. The company has adopted a transparent pricing policy and committed to lowering the cost of genetic testing to expand patient access.
"We make it very clear to payors what price we're going to bill them. We ask them what codes to use, we use those codes, and we stick to that price," Invitae CEO Sean Geroge said. But, he acknowledged that the genetic testing industry does have a history of stacking CPT codes to inappropriately bolster payment from insurers, which in turn has resulted in payors putting forth overly restrictive prior authorization policies on all genetic tests or employing lab benefit managers to rein in spending.
"We've seen $16,000 code stacks for carrier screening and $28,000 code stacks for exome sequencing from other companies," he said. Concerns over how certain labs were billing for hereditary cancer genetic test panels even grabbed the attention of the US Department of Health and Human Services' Office of Inspector General.
Labs' coding games have made insurers particularly unwilling to pay for genetic test panels, which George underscored ends up hurting patients most. For example, in the prenatal testing space, insurers will pay labs that stack codes for individual tests for cystic fibrosis, spinal muscular atrophy, and Fragile X syndrome. However, they will not pay for a panel that gauges 300 actional disease-linked genes in expecting mothers when it is billed using those same codes for CF, SMA, and Fragile X.
"You can look at this example and see how the confusion in all of this is preventing women today from getting expanded carrier screening, which would otherwise cost the exact same amount that payors are currently paying for tests for two or three different diseases," George said.
Invitae would support the type of coding solution Concert has developed that aims to standardize coding and pricing in genetic testing. "We are 100 percent behind it because that will help shine the light of day on a lot of practices that are hurting the industry and patients," George said. "I'm hopeful that through these efforts there might be a kind of a reawakening or maybe even just a calling to accountability."