More evidence needed before genetic tests become mainstream

August 1, 2013

Physicians have yet to master the implications of genetic test and their relationship to treatment, while plans await outcome evidence

Lack of evidence in personalized medicine makes the practice especially difficult to manage for clinicians and payers. Generally, the concept involves tailoring medical treatment to an individual’s unique genetic profile.

A study published in the October 2012 issue of the Journal of Personalized Medicine (“Insurance Coverage Policies for Personalized Medicine”) cites the relatively slow adoption rate of personalized medicine as partly attributable to insufficient evidence of clinical benefits. There is also a hesitancy by insurers to cover personalized medicine without such evidence-based proof of effectiveness.

The study explored 41 coverage policies for 49 unique tests, including 22 for disease diagnosis, prognosis and risk, and 27 for pharmacogenetics. Insurers covered 50% or fewer tests on the selected list. The lack of broader coverage was primarily attributed to lack of evidence.

However, when treatment called for drugs for which pharmacogenetic information was included in drug package inserts, more tests were covered. In other words, if the prescribing information can link diagnostic tests to health outcomes, insurers are more likely to cover the tests. Insurers appear more confident about covering tests when the inserts indicate the approval of the FDA.

Herceptin (trastuzumab), a breast cancer drug, is one of the best known examples of a drug that is prescribed through the practice of personalized medicine. The drug targets breast cancer patients who express the HER2 protein, and testing for the protein is recommended so clinicians can determine in advance if the patient will respond to the treatment.

The field is progressing, but the search for evidence is slowing its pace.

“Encouraged by health information technology, individualized medicine is able to leverage the robustness of real-world evidence and help identify the right clinical pathways and treatment protocols,” says Terry Hisey, vice chairman, U.S. Life Sciences leader for Deloitte in Philadelphia.

On the other end of the spectrum, Hisey says, the science and necessary research funding are not yet available. Not to mention that while testing costs have come down, the interpreting the results is still costly and complex.

In 2010, Vanderbilt University Medical Center in Nashville launched personalized medicine initiatives for cancer and commonly used drugs to treat cardiovascular conditions, immunological diseases and transplant rejection.

Dan Roden, MD, assistant vice chancellor for personalized medicine at the medical center, says one of the drivers behind the programs was his vision of being able to one day genotype patients and embed the genetic information into their electronic medical records. Such information could improve treatment now and in the future.

However, his initial vision was impractical 10 years ago because most prescribers would not know what to do with the information. Even today, physicians are unfamiliar with the more than 2,700 inherited genetic disorders for which tests are available and the 3,000 or more available genetic tests for treatment of diseases.

“In the Vanderbilt program, cancer patients’ EMRs house genetic information at the point-of-care as a result of routine genotyping at the DNA level to find appropriate therapies linked to mutations that are increasing a cancer’s growth,” Dr. Roden says. “The initial emphasis was on mutations in melanoma and lung cancer that might be directly related to existing and new therapies, and the program is now expanding to breast and colon cancers.”

Once data are put into an EMR, Dr. Roden says, decision support can assist the physician. If a drug is prescribed for someone with a genetic variant, the EMR will inform prescribers, for example, that another medication could be more effective.

The program has also included testing for more than 13,000 heart patients, and researchers found about 25% have a mutation that predicts a poor response to Plavix (clopidogrel). The information can help physicians determine a better treatment path.

Dr. Roden says the medical center underwrites the genotyping and is expecting to gain buy-in from payers to share the cost for gathering information that could improve outcomes and reduce exacerbations.

The medical center also has developed a biobank, holding 170,000 samples of DNA linked to de-identified medical records. Dr. Roden says the bank is being used to research genetic variants that could be important in determining which patients have a predisposition to a certain disease, their responses to different drugs and any susceptibility to side effects.

Using Prior Authorizations

Steve Miller, MD, senior vice president and chief medical officer for Express Scripts, in St. Louis, says that using a set of algorithms, the PBM looks at a patient’s medical history, demographics, ability to pay for drugs and behavior to predict medication adherence. The predictive modeling helps determine who might be a good candidate for a specific drug.

Express Scripts uses prior authorization as a tool related to personalized medicine to ensure that patients who are responsive to a specific medication have access. The policy applies to Herceptin for breast cancer and Selzentry (maraviroc) for patients with HIV-1 infection, for example, as well as 19 other drugs.

Dr. Miller says there is a lack of uniformity in diagnostic tests, inconsistent availability and concerns that the tests may over- or understate risks and benefits. The Personalized Medicine Coalition suggests that pharmacy benefit teams can use real-world observational data taken from clinically validated tests, rather than a controlled trial, to determine the clinical value and cost-effectiveness of a test.

However, many observers believe FDA must provide more oversight of genetic tests.

Jon Minken, senior director, global reimbursement for Qiagen, provider of sample and assay technologies in Gaithersburg, Md., says the prevailing strategy for determining who might be a prime candidate for a specific therapy has been a shotgun approach, with physicians using every therapy available to treat a disease instead of tailoring for effectiveness.

“With the high cost of specialty drugs, it is critical to use a diagnostic test to accompany a drug to determine if a patient will be responsive,” Minken says.

He recommends that pharmaceutical companies submit both the diagnostic test and the drug application to the FDA at the same time when a test is warranted.

“Companion diagnostics are win-win,” says Carole Welsch, director, personalized healthcare for North America for Qiagen. “Trial and error is unnecessary; instead, you can rely on test results.”

If there is a known population that would benefit most from a drug, there will be a more judicious use of resources even if the therapy costs more, according to Minken. He advocates for prior authorization for selective therapies.

Qiagen developed a diagnostic gene test to identify the best responders to Erbitux (cetuximab), a colorectal cancer drug approved by FDA in 2004. Analysis indicated that patients with a mutation of a certain cancer-associated “KRAS gene” should not receive the drug because it would provide no benefit. Patients with a normal version of the gene are more likely to respond to Erbitux plus chemotherapy.

Welsch says 40% of patients have the mutation.

New drugs with companion diagnostics are reaching the market with a even more in the pipeline. For example, FDA recently approved Gilotrif (afatinib), a drug for non-small cell lung cancer along with a companion diagnostic to identify EGFR mutation-positive patients.

Physician reaction

Eric Topol, MD, chief academic officer, Scripps Health, a not-for-profit health system in La Jolla, Calif., prefers to call personalized medicine, “individualized medicine,” to reflect its ability to capture all essential data on every person.

Scripps uses routine genotyping to predict responses to Plavix (clopidogrel) in patients with stents and  sequences tumors and the DNA of cancer patients to discover if any mutations are present.

Dr. Topol says that testing patient DNA for a mutation might be limiting.

“The same mutation may be causing more than one cancer so testing should not be by the organ but rather by the mutation itself,” he says.

He says he is bit skeptical about the reaction of physicians to personalized medicine. Too many see the concept as potentially threatening their authority not to mention their wallets if they are not compensated for testing and application of results.

Cover Genetic Tests

Response Genetics, which develops and sells molecular diagnostic tests to determine how cancer patients respond to drug therapy, recently initiated a provider contract with Blue Cross and Blue Shield of Illinois (BCBSIL).

“It's another component of our provider network that enables BCBSIL members to make the most of their in-network benefits for these services, and we feel it's a value to offer them,” says Michael Deering, director, media and public relations for the health plan.

Although most insurers will cover diagnostics if there is evidence to support them, some patients will ask for a drug even if tests show they are unlikely to respond to it. And insurers must be prepared to provide evidence as well as utilization management.

Tested in Court

The recent Supreme Court ruling that naturally occurring genes cannot be patented should breathe some life into diagnostic tests used to identify which patients would be good candidates for certain therapies.  

Myriad Genetics claimed to hold patents for the BRCA1 and BRCA2 genes, which increase the risk of breast and ovarian cancer, respectively, and for a diagnostic test to identify whether patients have these mutations. The ruling invalidated parts of the patents, enabling other labs to use new technology to develop and sell broader one-time tests to identify cancer risks.

Supporters of the decision say it opens the door to more clinical care and research and the ability to develop multi-gene panels.

Drugs with pharmacogenomics biomarkers

Abacavir

Ado-Trastuzumab Emtansine

Aripiprazole

Arsenic Trioxide

Atomoxetine

Atorvastatin

Azathioprine

Boceprevir

Brentuximab Vedotin

Busulfan

Capecitabine

Carbamazepine

Carisoprodol

Carvedilol

Celecoxib

Cetuximab (2)

Cevimeline

Chlordiazepoxide and Amitriptyline

Chloroquine

Cisplatin

Citalopram (2)

Clobazam

Clomipramine

Clopidogrel

Clozapine

Codeine

Crizotinib

Dapsone (2)

Dasatinib

Denileukin Diftitox

Desipramine

Dexlansoprazole (2)

Dextromethorphan and Quinidine

Diazepam

Doxepin

Drospirenone and Ethinyl Estradiol

Eltrombopag (2)

Erlotinib

Esomeprazole

Everolimus

Exemestane

Fluorouracil (2)

Fluoxetine

Fluoxetine and Olanzapine

Flurbiprofen

Fluvoxamine

Fulvestrant

Galantamine

Gefitinib

Iloperidone

Imatinib (4)

Imipramine

Indacaterol

Irinotecan

Isosorbide and Hydralazine

Ivacaftor

Lansoprazole

Lapatinib

Lenalidomide

Letrozole

Maraviroc

Mercaptopurine

Metoprolol

Modafinil

Mycophenolic Acid

Nefazodone

Nilotinib (2)

Nortriptyline

Omeprazole

Panitumumab (2)

Pantoprazole

Paroxetine

Peginterferon alfa-2b

Perphenazine

Pertuzumab

Phenytoin

Pimozide

Prasugrel

Pravastatin

Propafenone

Propranolol

Protriptyline

Quinidine

Rabeprazole

Rasburicase

Rifampin, Isoniazid, and Pyrazinamide

Risperidone

Sodium Phenylacetate and Sodium Benzoate

Sodium Phenylbutyrate

Tamoxifen (3)

Telaprevir

Terbinafine

Tetrabenazine

Thioguanine

Thioridazine

Ticagrelor

Tolterodine

Tositumomab

Tramadol and Acetaminophen

Trastuzumab

Tretinoin

Trimipramine

Valproic Acid

Vemurafenib

Venlafaxine

Voriconazole

Warfarin (2)