The goal of creating a rapid learning system for oncology advanced this summer when the American Society of Clinical Oncology announced it is developing a breast cancer-specific prototype. This first step, scheduled to be completed by year's end, will help ASCO assess the challenges of building CancerLinQ—the official name of its rapid learning system—and refine its approach before expanding to include other cancers.
Since the term “rapid learning system” first emerged in 2005, many top-level players in the health care field have come to embrace it as the vision for the future of health care research and clinical care. Many organizations are working to advance the concept of rapid learning systems, but none in quite the way ASCO is doing.
“What distinguishes it is that this is truly a bold and national vision. I don't know of any other professional society or government agency that has quite put all the pieces together—and, more than that, said, ‘We're going to go do this because it's essential to the practice of oncology,’” said Lynn Etheredge, Director of George Washington University's Rapid Learning Project and the nation's leading advocate for rapid learning systems (RLS).
It is a massive undertaking that will require partners from health care information technology, government agencies, pharmaceutical and biotechnology manufacturers, advocacy organizations, and others. And noted George W. Sledge Jr., MD, a member of the CancerLinQ Advisory Group, ASCO's ultimate role in the enterprise is not clear because the cost and scope of the initiative is not yet known.
For the moment, the society's role is as initiator of what ASCO leaders think will be a transformation in the way oncologists treat their patients, said Peter Yu, MD, Chair of ASCO's Health Information Technology Work Group.
“We are looking ahead five to 10 years to see where medicine is going—not just oncology and not just patient care, but the entire health care system,” he said. “This is bolder than anything ASCO has ever done before.”
Even as it takes the first step on its own, ASCO is reaching out to a wide range of potential partners. “We are trying right now to put this out because this is going to require support from oncologists, as well as regulators, as well as payers, as well as health care IT,” Yu said. “We need to build consensus early on because we want to have a partnership.”
Creating the Prototype
To create the RLS prototype, ASCO will draw on its breast cancer clinical practice guidelines and Quality Oncology Practice Initiative (QOPI) measures to develop quality measurement and clinical decision support tools. De-identified breast cancer patient data will be imported from the electronic health records (EHR) of a sample set of oncology practices.
The goals for the prototype initiative are to:
- provide “lessons learned” about the technological and logistic challenges involved in a full-scale RLS implementation;
- provide real-time clinical decision support integration in a demonstration EHR system;
- demonstrate value-added tools, such as the ability to measure a clinician's performance on a subset of QOPI measures in real-time;
- demonstrate the ability to capture and aggregate data from a variety of sources; and
- create new ways of exploring data and generating hypotheses related to breast cancer.
The prototype is scheduled to be unveiled later this year at ASCO's first Quality Care Symposium (Nov. 30-Dec. 1 in San Diego).
The Big Picture of Rapid Learning
An economist by training, Etheredge conceived and started promoting the idea of a rapid-learning health system that uses EHRs to create large, searchable national databases of de-identified patient data in 2005. Since then, the Robert Wood Johnson Foundation has funded his work as chief cheerleader for RLS—writing, speaking, coaching, networking, and educating about the potential of these databases.
He tracks what is being developed around the country and can reel off a long list of database development projects that are early forms of an RLS. For example, the Society for Thoracic Surgery has collected data from more than 4.3 million surgical procedures to create its cardiothoracic surgery outcome and quality improvement program, and a group of health systems including Kaiser Permanente and Geisinger Health System has developed a data warehouse network that has access to some 16 million patient records.
Meanwhile, the federal government's financial incentives for EHR adoption are helping build the technology infrastructure needed for an RLS to become a reality. Most recently, National Institutes of Health Director Francis Collins, MD, laid out a vision for the future of medical research that, to Etheredge's ears, sounded like it is based on an RLS.
“It's the first time that someone at that level has endorsed the core concepts of building an electronic health record-based system with tens of millions of patients,” he said. “In his view, it's to enable precision medicine that addresses the wide genetic variations across people. Clearly, the report he was keying off of was the precision medicine report from the National Academy of Sciences last fall.”
That report (OT, 1/10/12) addressed the potential breakthroughs from integrating electronic health record data with genomic data so that disease subtypes and the efficacy of treatments for specific subpopulations can be rapidly understood and the new knowledge applied.
It was that potential that prompted Etheredge early on to identify oncology as an important focus for RLS development. In 2009, he helped organize an Institute of Medicine workshop on rapid learning in cancer care (OT, 10/25/10).
“It's imperative for the future of cancer research and cancer care to have this kind of system,” he said. “And, if we can do it in cancer—because cancer is always the most difficult area—we're miles ahead in almost every other problem in the health system. So it's a wonderful model to learn from.”
What ASCO is Doing
ASCO's first step toward creating CancerLinQ occurred last year when it created a quality department, now known as the Institute for Quality. A few months later, it hired Robert Hauser, PharmD, PhD, former director of operations and informatics at the International Oncology Network, as senior director of the Institute, responsible for expanding ASCO's clinical practice guidelines, helping oncology practices adopt EHRs and other health IT and expanding QOPI.
Those three efforts are all needed to support CancerLinQ, Sledge explained. For example, up until now ASCO has limited its cancer care guidelines to situations in which there was the highest level of scientific evidence; thus, it does not have guidelines for a wide array of disease scenarios. Because CancerLinQ will need to provide electronic clinical decision support, it needs to be populated with guidelines that cover most of the situations a practitioner encounters.
“We're certainly trying to increase the breadth of coverage for our guidelines . . . by hiring new staff within the guidelines group and by having strategic relationships with other organizations that help create guidelines,” he said.
Meanwhile, ASCO is looking for ways to speed the adoption of technology. Fewer than half of all oncologists are using EHR systems, and the only way for a clinician to participate in CancerLinQ will be via EHR technology.
Further, Yu said, ASCO must relaunch QOPI as an electronic system. Currently, QOPI requires manual submission of quality measures, which is too labor-intensive and time-consuming. The work will require some QOPI measures to be redesigned so they can be collected and analyzed electronically. Additionally, the system needs to be able to accept data outside of QOPI.
Even as ASCO works on those prerequisites for CancerLinQ, leaders are networking externally to identify partners that can make the RLS a reality. Among others, ASCO has had discussions with IT giants like Microsoft, Oracle, Google, and IBM; several EHR vendors and pharmaceutical firms; government agencies including the Veterans Administration and the National Cancer Institute; nonprofits such as the National Comprehensive Cancer Network and Komen for the Cure; and patient data-sharing companies such as PatientsLikeMe.
The society has already contracted with some IT companies and is negotiating with several others to develop some of the component parts that allow the collection and analysis of large numbers of patient records, Sledge said.
“This is a rapidly evolving area, and we don't know whether we are going to be one among many players who have something similar going on, whether we are going to be the leader, or whether we are going to be an organization that provides decision-support guidelines and standards to perhaps another organization. It is way too early for us to be able to say.”
What It Means to Oncologists
The breast cancer-specific prototype that ASCO plans to unveil in November will not immediately be of use to clinicians— rather it will be a proof of concept that comes with a list of challenges to overcome before the full RLS can be built.
“The intent is to build a sandbox that we can play in . . . as a first step toward the rapid learning system,” he said.
When the RLS is functioning, oncologists who do not use EHR technology will be unable to provide their patients with the same level of evidence-based care that their wired colleagues can.
“I can't imagine how an oncologist in the future could practice without this kind of support,” Etheredge said. “I think people will look back and say ‘How could we have ever practiced oncology without access to real-time knowledge bases?’”
PODCAST: Listen on the iPad edition of this issue as George W. Sledge, Jr., MD, a member of the CancerLinQ Advisory Group, describes the practicalities of how a rapid learning system for oncology will affect patient care.
To receive our iPad issues, download the free Oncology Times app from the App Store today! Visit http://bit.ly/OT-iPadApp or follow the link on oncology-times.com