Integrating the EHR for Precision Medicine Has Evolved

ACCC logo | Image credit: ACCC

For all the possibilities of precision medicine, whether a patient gets a targeted therapy still comes down the decision of a busy doctor to order biomarker testing—and then, upon getting the results, to properly use the information to prescribe the right therapy, if a patient has a mutation that can be targeted.

Data show that at each step, doctors fail to make the right choice about half the time, and for years, having an electronic health record (EHR) that could not integrate genomic testing information was part of the problem. But that’s changing, and a panel Thursday at the Association of Cancer Care Centers 50th Annual Meeting and Cancer Center Business Summit explored aspects of what technology is now available as well as the steps institutions can take to make the leap to an integrated system.

Sigrun Hallmeyer, MD | Image credit: Advocate Lutheran

Moderated by Sigrun Hallmeyer, MD, medical director, for Advocate Lutheran General Hospital Cancer Service Line and Survivorship Center, the panel featured:

• Jonathan Katchmore, associate vice president, Commercial Diagnostic Strategy, Loxo Oncology;
• Christopher McNair, PhD, associate director for Data Science and director of Cancer Informatics, Sidney Kimmel Cancer Center; assistant professor, Department of Medical Oncology, Thomas Jefferson University;
• Brian Davis, Genomics Implementation Lead, Epic;
• Nate Wade, PharmD, MBA, BCOP, associate director, Clinical Oncology, Flatiron Health.

“Precision medicine has really become standard of care” Hallmeyer said. “Indications are ever expanding,” she said, use of therapy to hit targets is “not just prolonging life and reducing toxicity, but actually saving lives.”

But there are still many barriers to patients receiving testing, the biggest one being that access comes down to the decision of individual physicians. “The clinician has to constantly understand which diagnosis is the appropriate diagnosis in which a particular test may be appropriate,” Hallmeyer said.

It’s not just having scientific knowledge, such as “Is this even a test that is actionable?” but also daily practicalities, such as knowing how to use a testing company portal and remembering the password. Ordering the wrong test can have financial ramifications, because then a patient’s testing coverage may be exhausted and a physician may have missed a critical gene.

Bringing information into the EHR is critical, but getting there takes time and commitment, speakers said.

Missed opportunities. Katchmore highlighted the growth of precision medicine in clinical trials, saying the rise of precision medicine has improved survival. For the physician in everyday practice, however, it’s “almost impossible” to keep up with the growing number of trials involving biomarkers.

Although physicians often think they are testing for an actionable biomarker, he pointed to a study that showed only 46% of patients with cancer are tested. What’s worse, a 2019 study found that among patients who were tested and found to have an actionable mutation, only 48% received the appropriate therapy. “We often refer to that as winning the genomic lottery and not being able to cash I on that lottery ticket,” he said.

Why are so many patients missing out? What are the gaps? Katchmore said having molecular tumor boards where community oncologists feel they are heard—and where challenging cases can be aired—can be very effective. “There are data showing now that many of these patients are now benefiting greatly,” Katchmore said.

One key is to make these boards multidisciplinary, he said, so all stakeholders feel part of these outcomes.

Real-world implementation. McNair described the process undertaken at Jefferson Health to fully integrate a systemwide precision medicine structure, which was needed as the health system grew to 14 hospitals across Pennsylvania and New Jersey. More physicians were using testing, and there was no consistency in the use of vendors, even within the same type of cancer—and none of the data was making its way into the EHR. “It was chaos,” he said.

McNair partnered with a medical oncologist who shared his vision for a fully integrated system, and they created a multistakeholder team that included members from information technology and those with access to health system leadership—an essential piece. He advises institutions and health systems to strike a balance between having “too many cooks” but not enough members to have the right voices involved. Doing this upfront prevents problems later on, he said.

The group developed a charter and developed information on which tests were most frequently used, as well as costs and materials that would be needed to make integration happen. McNair said the team didn’t do all their work themselves—they required vendors to collect some of it, although the process would necessarily reduce the number of vendors. Keys to success included finding ways to reduce complexity in the workflow and to make the integration reproduceable as tests were added. Education ahead of the launch was important, too.

He advised the group to develop metrics to track success—and to not try to impose external logic on how IT does things.

McNair said the first integration was the most painful, but once physicians and staff saw it succeed, “You’ll have people clamoring for more.”

An Epic evolution. To get an idea of Epic’s reputation, Hallmeyer introduced Davis by saying he would discuss how the EHR giant would tell “how to make Epic work for us instead of us working for it.” Davis just smiled, then described Epic’s genomic module, which he said is not so new anymore—and designed to help the community oncologist who has limited resources.

“It’s essentially liberating the genomic data from the PDF,” Davis said. “Imagine a place within the EHR that you can treat just like any other clinical data. You can query it; you can easily interact with decision support.”

As the EHR connects with testing laboratories, he said, “data harmonization is happening naturally.” Epic continues to work on issues such as coding systems and portability, and the need for data standards is evident.

He addressed the provider efficiency challenges that Hallmeyer mentioned. “Even if you’re integrating with a laboratory and all you’re getting back is a PDF, it’s still worth doing that—getting the discrete data is definitely beneficial,” he said. Electronic ordering is still better than “having to jump out into a web browser and order something in a portal, then try to find the scan and upload it into the media tab.” These steps can be eliminated with an electronic connection to the laboratory—it may be just a start, but it’s still an efficiency gained.

Biomarkers as a quality metric. Wade noted that for practices taking part in the Enhancing Oncology Model, use of appropriate biomarkers will be a quality metric, so he expects to see the practices that have signed on for CMS payment model to continue advancing in this area.

Flatiron, known for its OncoEMR and FlatironAssist products, is integrated with Caris, Foundation Medicine and Guardant, and Wade said there is another major integration coming shortly. He concurred with Wade that the ability to order tests within the EHR was more efficient, adding that was is also important is “making those results more actionable and easily retrievable.”

Having the PDF in the structured results is important, Wade agreed, but it’s also important to make it “more accessible and easier for the physician to find 1 week or 1 month or even 6 months or a year later. I think that’s really a tremendous value that we provide.”

Flatiron is experimenting with helping physicians with decision support features that reflect real-world realities, including the fact that payers drive many decisions. When it first rolled out a tool that included biomarker cohort regimens in the National Comprehensive Cancer Network (NCCN) guidelines, the feedback was that NCCN has many regimens that meet guidelines—and physicians would like to bump their preferences to the top. So that is now possible.

And now, Flatiron is experimenting with options that show which choices reflect which payer preferences. That might put some off, he said “but if you’re a physician, at the end of the day, the payer is ultimately who decides [how] to pay for that medication.”