Takeaway: ILMN is a long bench idea on our Position Monitor


We have been working to understand the current backdrop for genetic testing in recent weeks and have both been analyzing our claims data for genetic testing CPT codes. We have also been evaluating a number of data sets including ClinVar, GTR, NIH RePORTER, and others, to better describe and monitor trends in adoption.  We had the opportunity to discuss some of our data with an expert in the field who has spent time at a number of well known labs and companies in the space.  As we have become more optimistic about the advent of broader adoption from developing trends such as polygenic risk scores and liquid biopsy. However, our contact was significantly less optimistic.  We've included a table from our claims data showing the top 20 diagnosis codes used for genetic tests.  Its worth noting the top volume lab in our data set is Progenity a prenatal test company, but but our expert suggested that a large percentage of prenatal tests are paid out of pocket for several reasons.

Field Notes | ILMN | "50% of these labs will go under..." - genetop20

Field Notes

  • Test "price points are no longer tied to cogs"
  • In 2012-2013 the market took off when venture capital money got involved, lots of NGS tests were developed
  • Managed care reimbursement won't cover the cost of the test at many of the price points now offered and "it is extraordinary this is happening"
  • Venture capital money seems to be subsidizing the cost of goods to run these tests
  • "You can't be profitable without being dishonest"
  • Yes there are many approved single gene tests for rare disease, but they have "non existent" volume
  • A high volume lab for these single gene tests may see 200 tests per year, but it is 0 for most other labs
  • The exception is Cystic Fibrosis test
  • Large panels may consist of 35 genes where cogs run $100-$200 for the assay, but a quality interpretation can cost $1,000 to $6,000 for an accurate review
  • A liquid biopsy may require 1-2X the cost of goods and 3-4X the interpretation costs
  • There are off the shelf software interpretation solutions, which automate the process, but they have a lot of false positives and false negatives
  • Invitae, at $250 test price points, must be "taking short cuts"
  • Color is another low cost lab
  • MYGN has the best managed care team
  • Interpretation software vendors include Fabric Genomics, WuXi, Emperean, Qiagen, Strand Genomics
  • Once a variant is logged, the incremental cost is lower, but there are always new variants
  • Some vendors use ClinVAR, scrape internet, NOMAD, etc, but quality is low because "nobody cares, and nobody notices"
  • For example "if you don't get caught and say a pregnancy is terminated because of a false positive, there is no consequence to the business"
  • Some tests have a 30% false positive rate
  • The problem is that the data is relying on 3 decades of literature that is wrong
  • For BRCA, the public data is really good, but for many genes the data is not accurate
  • An inaccurate data point will cause a lot of false positives, but it takes time to review all of the information even for a highly qualified person
  • ClinGen is a great group offering unenforceable guidelines
  • There is an managed care effort to targeting highest quality tests from the highest quality labs
  • "50% of labs will go under" because they won't make the investments or don't have the expertise
  • A test review will ask "do all of the genes make sense" on this panel?
  • "Is there evidence that what the lab is selling has the evidence to support the claim"
  • In a paper last June, Brugada syndrome test was shown that only 1 gene out of the 21 tested was valid
  • Academic volume is so low it does not matter, and academic labs "never come up in conversation" with managed care
  • But academics have an outsized voice in the literature and in policy
  • Oncology claims are dealt with much more leniently by Managed Care
  • For cancer, is not hard not to to make a claim, low bar compared to single gene tests and rare disorders
  • Polygenic risk scores will take a significant amount of work and effort, but "should be entirely possible" but there is not enough patient data yet, odds are they validate on the training data, which is a problem, but it may not matter
  • Nutri-system weight loss test sells like "hot cakes"
  • Weight loss test, but the literature suggests the genes don't matter, you're just selecting for motivated patients and getting  a placebo effect
  • Whole genome is a super risky proposition because there will be 4-5 false positives per person, and you will never ever get that reversed or out of your medical record
  • Fears of parents make it more likely patients, when they can, pay out of pocket and do not connect test result to a patient record, especially for their kids
  • For Managed Care " the cost of the test does not matter, what matters is all of the downstream costs which is way more expensive than a $5000 test"
  • Germ line coverage and access is dropping, and reimbursement price is falling, Managed Care "seems sick of it"
  • For UNH their lab network announcement was not coordinated with Optum
  • Managed Care is offloading to third party review by LBNs
  • who are paid more for denials.
  • In some cases managed care has license out to an LBN to write policy, which leads to more denials, and more money for LBN
  • Edicore who are good, AIM, Beacon, Avalon
  • Germline testing is mainly pre-natal, but mostly paid for out of pocket
  • Oncology, like GH, is seeing out of control costs
  • code stacking is common
  • There are examples where labs charge $22,000 for a single test
  • Some labs "skate for a while" before payor knows, "its just a big game"
  • Labs will "probe the system" for highest reimbursement, Myriad won't break the law, but they are especially good at this
  • There "Lots of dumb labs out there"
  • Some labs "double bill" meaning they break up a claim into multiple submissions over several weeks.
  • "How long will this go on?" Venture capital is fueling the losses
  • Invitae must be making a data play at $250 per test, which "just doesn't make any sense"
  • Flatiron is "smart" solution
  • Fulgent, not a lot of volume, don't really see them come up, but they are at all the meetings
  • 50-80% of pre-natal genetic testing is out of pocket
  • There are only 20 meaningful labs and maybe 200 in total that matter
  • Many labs on GTR don't then remove themselves when they close, lots of zombie entries

Thomas Tobin
Managing Director