NEW YORK (GenomeWeb) – Academic institutions and molecular diagnostic companies are increasingly developing tests that run on next-generation sequencing platforms, but reimbursement for such tests can often seem arbitrary — varying depending on the specific insurer, by the test developer, the indication for the test, and even on an individual case-by-case basis.
At last week's Clinical Genome Conference in San Francisco, John Pfeifer, vice chair of clinical affairs and pathology at the Genomics and Pathology Services Laboratory at Washington University School of Medicine, discussed the NGS-based cancer test his lab has developed and how it is able to obtain reimbursement for around 90 percent of tests that it runs.
The key, Pfeifer said during a presentation at the conference, is to include only genes that impact prognosis, diagnosis, or therapy — the "bite-sized group of genes that can impact patient care."
Wash U's cancer panel comprises 40 genes, including those relevant for solid tumors, hematological tumors, as well as three pharmacogenomics genes. However, Pfeifer said that the lab also sequences 108 additional genes for research purposes. The lab only reports back and bills for 40 clinically relevant genes and "the cost of those 108 genes is absorbed by us," he said. Such an approach is OK with the Centers for Medicare and Medicaid Services, he added, as long as the "cost of the test is based on the 40 genes."
GPS charges $7,500 for its test and Pfeifer said mean reimbursement is between $4,000 and $5,000, but there is a large variation.
GPS is now working on a third version of its cancer panel, but unlike most laboratories that continuously expand the number of genes in their panels, GPS is moving in the opposite direction and creating smaller panels.
"Some payors are already saying that 40 genes is too much," Pfeifer said.
In the next iteration, which Pfeifer said will launch in the third quarter of this year, GPS will split genes related to solid tumors from those related to hematological cancers. The lab will still use the same set of capture probes, but oncologists will no longer have the option to order the 40-gene set as a single test.
One key factor that Pfeifer said sets GPS apart from some other academic clinical labs that design next-gen sequencing tests is that GPS does not rely on grants, institutional support, or philanthropy. It is completely self-sustaining. Therefore, it looks at next-gen sequencing as simply a tool to get a diagnostic or prognostic answer, rather than as a test in and of itself. As such, GPS will only use NGS if it makes both clinical and economic sense.
"We're not interested in the cost of whole-genome sequencing," he said. "We're interested in how much sequencing you have to do to make it cheaper than Sanger sequencing."
For GPS, that cut-off is around 2-2.5 kb of sequence. Then, he said, the lab figures out what clinical situations require more than 2 kb of sequence data when deciding what technology to use for its clinical tests.
Because it had to be self-sustaining from the beginning, Pfeifer said the lab looked at the key component for obtaining reimbursement and designed a test that would meet the criteria.
First, he said, the test had to demonstrate clinical utility, typically defined as improved outcomes for the patient. Second, the test had to have widespread adoption and could not be thought of as "boutique testing." And finally, the test has to have an immediate impact on the patient, not a future benefit. Cost savings is also important, he said, but less so than these other criteria.
The main factor to understand, he said, is that insurance will only pay for testing that is relevant to a patient's disease now. For instance, he said, for a metabolic test, doctors don't make the argument that lipid, liver, and pancreas enzymes should also be included in all samples just because that baseline information might be useful later. Similarly, he said, with next-generation sequencing, only genes that will offer insight into a patient's current disease should be tested.
In addition, he said, when GPS was developing its test, it changed the conversation with payors from whether the insurance company would pay for NGS testing to whether the company reimbursed for a specific result. The method doesn't matter as long as the test gives an accurate, clinically relevant result, he said.
GPS also developed its own white paper about NGS testing that it distributed to payors, and GPS is the one that contacts with the payors when it does the testing. Pfeifer said that originally, the ordering oncologist was responsible for billing the payor, but the lab found that oncologists ordered fewer tests and the lab did not get reimbursed as often.
Now, he said, nearly all of its tests are being reimbursed by both private companies and CMS. About one year ago, 95 percent of tests were being reimbursed, he said. That number has dropped to between 80 percent and 90 percent because GPS expanded the group of patients it tests. Even so, only around 5 percent are denied outright, Pfeifer said.
Insurance companies do not require pre-certification for around 55 percent of patients. In around 33 percent, GPS must make a phone call to obtain the reimbursement, and insurance companies request additional information in about 12 percent of cases, Pfeifer said.
"It's a done deal that we'll get paid," he said. "It's no longer a conversation about whether the test has utility. The conversation is now about how much we'll get paid."