3. While it is clear that real-world evidence helps to address an unmet need, it is also flawed. What are some of the challenges associated with real-world evidence?
In the real world, cancer care is a messy business. Fragmentation of care delivery across different sites of care, complexity of supply chain, and outdated data management and record keeping make real-world evidence analysis an imperfect science. Across claims data and EMR data, we see significantly more variability and lack of completeness than we do in clinical trial data.
This variability is one of the key reasons why some healthcare providers have been slow to trust real-world evidence. However, as only 3 percent of adults participate in cancer clinical trials and those that do are not representative of the general population, it is critical that the knowledge trapped within the 97 percent of patients who do not participate in clinical trials finds its way into value assessment. We have focused our research not only on performing the real-world comparative effectiveness research that is needed, but also on improving the quality of the research by identifying ways to solve the problems inherent to real-world data.
4. What have we learned from recent research studies by Cardinal Health about the challenges associated with real-world evidence and possible solutions for overcoming these issues?
We believe we have made significant contributions to the science of HEOR. This year, our work helps to establish best practices in two critical areas: (1) line of therapy (LOT) assignment in claims datasets and (2) disease stage in EMR datasets analyzed with natural language processing (NLP).
LOT designation is critical to value assessment, but changing treatment paradigms have challenged traditional methodology. We’ve modeled the impact of the increasing use of maintenance therapies and “stop and start” treatment schedules on current HEOR analyses. We’ve challenged our colleagues to reach consensus on new LOT definitions.
Similarly, we’ve addressed the issues related to redundant and conflicting staging data—which refers to the progression or “stage” of the disease — found with NLP analyses of EMR. It is not unusual to see conflicting entries in a patient’s EMR on different dates or by different doctors, and sometimes even by the same doctor on the same day. Through our recent research studies, we have identified and proposed a methodology to validate the correct stage.
5. You led a symposium at ISPOR on the challenges of standardizing value algorithms for cancer care. As you speak to your colleagues across the oncology industry about the role of real-world evidence in defining value, do you see perceptions changing?
The symposium gathered experts from across the healthcare spectrum. With me were panelists representing providers, pharma, payers and policy makers. Although the panelists brought different perspectives, everyone agreed that our healthcare system was in transformation, current cost trends were unsustainable and that a transition to value-based care was inescapable. They praised nascent efforts at value calculation, but believed that first-generation tools are too limited in focus. There was also a shared belief that successful value assessment would need to be patient-centered and have underpinnings in real-world evidence.
We were gratified by the warm reception from a standing-room-only audience and, as a result of the high level of interest in this topic, we have added one-on-one interviews of our panelists to the current issue of our FOCUS magazine.
To learn more about Cardinal Health Specialty Solutions’ work in HEOR, please read the recent research abstracts or download the 2016 issue of FOCUS magazine.