Posted Mon, May, 16,2016
This author interview is by Dr. John Hornberger, of Stanford University. Dr Hornberger's full paper, The VA Point-of-Care Precision Oncology Program: Balancing Access with Rapid Learning in Molecular Cancer Medicine, is available for download in Biomarkers in Cancer.
Please summarize for readers the content of your article.
The article outlines the initiation of a novel, practical clinical program, called the Precision Oncology Program, intended to;
(1) assure access and remove disparities in cancer care for all patients enrolled in a health system, using the VA as pilot setting,
(2) increase the quality and speed of learning using new approaches for analyzing and interpreting complex medical information,
(3) increase patients' abilities to participate directly in their care.
Ultimately, the intent of this program is to more rapidly test and assess the approaches that will lead to substantial improvements in cancer survival and in the experiences of cancer survivors.
How did you come to be involved in your area of study?
I did my medical training at Stanford, followed by a 4-year post-doc in health outcomes and economics research. I became interested in how different disciplines collected, analyzed, and interpreted data, and then sought to put those learnings into practical action to solve real-world public health problems related to promotion of health. Throughout the 1990s I focused on statistics, decision analytics, and economics, and even implemented a rapid-learning system in many ways similar to the VA's POP applied to a dialysis-care setting. The project applied the state-of-the-art genomics of the time. Dialysis centers were an excellent setting to pilot such a program because they had the most sophisticated, online data systems available and patients were willing to participate because their treatments required attendance in the center 2 -3 times per week. We sought to analyze the data routinely collected develop better and more efficient decisions about how and when to adopt technologies. A key principle was to verify novel innovations in our own setting. Homocysteine lowering was picked as the initial project. We learned the program could be efficiently set up and we could quickly arrive at actionable findings, sometimes years before other groups that applied more traditional learning models.
What was previously known about the topic of your article?
We already had set up principles for implementing a program in a dialysis setting, using much less sophisticated data analytic tools than we have now. What hampered scaling of that program were the incentives of the dominant fee-for-service healthcare system that had few rewards for efficient learning in real clinical practice. Fortunately, reforms in payment systems are changing how people see the benefits of the program. They can see the potential for great improvements in health and wellness, and in simultaneously saving costs. To me, the VA POP program is natural progression of that initial dialysis program, inspired by people of a similar temperament to improve quality of care and bridge disparities as rapidly and efficiently as possible.
How has your work in this area advanced understanding of the topic?
There were several topics in designing the POP that needed to be explored. Not all of our experiences and learnings could be conveyed fully in the paper. We believe some of these may be important for non-VA entities who are pondering whether to implement their own programs. First, physicians and other staff came face-to-face with the challenge of how to interface with so much evidence generated in real time on their patients. Broadly speaking, the issues that arose centered around communications and decision making in the digital age. As a result, I have taken an interest in Sherry Turkle's - Director at MIT on Technology and Self-work on studying in a variety of setting the various nuances of the evolving digital age. Second, arriving at interpretations from the data that are clinically actionable in the new genomic and digital age requires a whole new set of understandings about how to analyze and interpret data. As precision medicine segments diseases into ever smaller groups of patients, innovations are needed in analytics (and I see novel ideas emerging from the statistical, economic, and computer science disciplines) to decide what is real (‘causation') and what is just along for the ride (‘association'). Last, the VA is focused on an important set of priorities that others will share, such as patient survival and experiences. I am delighted the VA is put so much focus on disparities across its system. We also discussed the cost or efficiency of care as this is important to the VA. Yet I can see how the systems consider this will differ and there are tools not discussed in the article which can bring insight to the program.
What do you regard as being the most important aspect of the results reported in the article?
The most important aspect, by far, is that an entity like the VA with a focus on quality of the care for veterans can integrate into routine practice a systematic, well-coordinated data-driven approach to learning what works (and what doesn't). It no longer is just an idea, but a reality. It is time for other healthcare entities to give it a try. I am excited to see a germ of an idea implemented 2 decades ago start to take hold - in its own way - in the VA. It means the idea is tangible and generalizable. I am excited to be an ongoing participant in what I see is a monumental change in how medicine is practiced and to see how this change leads to improvements in cancer care and the patient experience.
Posted in: Authors
Thu 09 Feb, 2017Public Science Insights: Where did we come from? Tracing the migrations of the first humans