Utilizing the Power of Rapid Learning Health-care Systems to Improve Patient Care 

A Conversation with Lynn Etheredge


Get Permission

CancerLinQ is a flagship health-care model not only in the treatment of cancer but for other disease specialties as well. If we learn how to make this kind of system successful in the treatment of cancer, it will become the model for every area of medicine.

—Lynn Etheredge

Lynn Etheredge’s career in shaping national health-care and social policy spans more than 4 decades and four Presidential administrations. He was the lead analyst in the development of health insurance proposals for Medicare and Medicaid while working in the White House Office of Management and Budget (OMB) during both the Nixon and Ford administrations and later headed OMB’s professional health staff in the Carter and Reagan administrations. In 1986, he was a founding member of the National Academy of Social Insurance, and during the 1990s, he continued to focus his efforts on crafting health-care reforms to reduce costs and improve health insurance coverage and quality, especially for low-income and elderly people.

In 2007, Mr. Etheredge proposed the concept of a “rapid learning” health-care system. His aim was to harness the information collected in patient electronic health records and utilize the power of the Internet and computing systems to build national medical databases and research registries, to facilitate best practices in clinical care and biomedical studies.

Today, he is continuing his work in the development of a national rapid-leaning health-care system with support from the Robert Wood Johnson Foundation. He is an advisor on health-care policy for a number of organizations, including ASCO, and is the Director of the Rapid Learning Project at George Washington University.

The ASCO Post talked with Mr. Etheredge about ASCO’s health information technology initiative CancerLinQ™ and how this type of resource is changing the way oncologists will treat patients in the future.

Benefits of CancerLinQ

How can rapid-learning health systems like CancerLinQ improve cancer care?

Most importantly, they will give every oncologist immediate access on their computer to the latest published research information, to help them make better treatment decisions. For example, the CancerLinQ prototype, which includes “de-identified” data from over 100,000 patients with breast cancer and is being tested now, should make it possible for oncologists to make a much more accurate diagnosis and determine more effective personalized therapies to improve the quality of care.

Health information systems like CancerLinQ will also provide oncologists with the ability to consult with each other on difficult cancer cases. For example, there are relatively few specific questions about safety and effectiveness that get addressed in the U.S. Food and Drug Administration (FDA) drug approval process. While we may know that a new drug is safe and effective in general, there are many questions that come up in clinical practices that don’t get addressed in clinical trials.

One example is how chemotherapies affect pregnant women. These women get cancers but are not included in randomized clinical trials, so there isn’t a lot of information about how to treat them. Having the ability to look up a database of thousands of patients with a particular characteristic or a particular problem will give physicians a lot more information to make a better decision regarding treatment. So this technology really fills in a lot of the gaps with regard to the issues that individual physicians and individual patients face that are rarely addressed in the academic literature.

I think community oncologists are going to be the greatest beneficiaries of CancerLinQ, because it is in those private practices where most cancer care takes place. Community oncologists will have the ability to ask not just one or two colleagues about their experiences in similar clinical situations, but thousands of physicians participating in CancerLinQ.

Impact on Survivors

In addition to providing oncologists with resources on individualizing treatment strategies, how else will CancerLinQ improve patient care?

CancerLinQ will also give us insights into the progress of the understudied long-term cancer survivor. Most of the research by drug companies and academic institutions is on the treatment and cure of current patients.

Cancer survivors are a very understudied group and a population that is growing rapidly. According to the National Cancer Institute, there are nearly 14 million cancer survivors in the United States, and that number is expected to climb to almost 18 million by 2022.1 I think CancerLinQ is one of the tools that will give us a unique set of insights and information about how to do a better job of monitoring and improving the quality of health status of cancer survivors long-term.

Research and Cost Issues

Will the use of health information technology also speed the dissemination of information on biomedical and comparative effectiveness research?

Absolutely. By building a database of patient health records, including treatment outcomes, as well as a database of national research registries, biobanks, and genomic data, and linking them, our understanding of science will advance very quickly.

In addition to the potential for improving patient care, will health information technology also be instrumental in reducing health-care costs?

We hope so, but I wouldn’t want to forecast that. It’s a complicated issue. The hope is that if we can prevent cancer and treat it more effectively once it is diagnosed, both of those accomplishments will reduce health-care costs. We are entering a world of precision diagnostics and gaining the ability to have targeted therapeutics and personalized medicine. The hope—and the expectation—is that we will save a lot of money in the process.

CancerLinQ vs Watson

How does CancerLinQ differ from IBM’s Watson supercomputer now being tested at Memorial Sloan-Kettering Cancer Center in New York and at other institutions?

IBM’s Watson is a very large software program that doesn’t start with medical knowledge and medical logic, so it needs to learn those things. The supercomputer may get to where it is a very useful medical decision tool eventually, but CancerLinQ is being built specifically for cancer care, and it will have the ability to look up existing patient databases, existing research registries, and existing scientific literature and analyze the information.

At the moment, I view Watson as a very interesting exercise in basic research that may ultimately prove very useful. Currently, however, I think it’s more useful for hypothesis-generating and -testing then it is for coming up with solid scientific information that can be useful in treatment decision-making.

Key Goal

Will rapid-learning health-care systems make overall health care in the United States more cohesive and uniform?

I think the technology is going to help standardize the approach to scientific research and delivery of care around best practices. Health information technology should help oncologists deliver the very best care for each patient, and that is the key goal here. I’m confident that we will make progress in that area. Whether it will alter the business side of medicine at health insurance and drug companies, I’m not sure.

What is extraordinarily important about ASCO’s approach in the development of CancerLinQ is that the physician community is taking the lead to build the databases that they and their patients need. Clearly, no other entity is going to be building the kind of resource CancerLinQ provides, so I think it’s imperative that we find a way to make this system successful.

Also, we have a lot to learn about how to build rapid-learning systems like CancerLinQ. CancerLinQ is a flagship health-care model not only in the treatment of cancer but for other disease specialties as well. If we learn how to make this kind of system successful in the treatment of cancer, it will become the model for every area of medicine. ■

Disclosure: Mr. Etheredge reported no potential conflicts of interest.

Reference

1. National Cancer Institute: Survivorship-related statistics and graphs. Available at http://dccps.nci.nih.gov/ocs/prevalence. Accessed May 10, 2013.


Advertisement

Advertisement



Advertisement