The slow, but inevitable evolution of electronic oncology health-care systems has already, at least conceptually, moved to the next generation of machines that not only store and process data, but also have the ability to provide real-time clinical decision support. At ASCO’s first Quality Care Symposium in San Diego, ASCO CEO Allen S. Lichter, MD, unveiled the Society’s ambitious addition to the evolution of the rapid-learning system.
Paving the Way Forward
“It’s my pleasure to launch the maiden voyage of ASCO’s contribution to the rapid-learning system, CancerLinQ. The idea of a rapid-learning system has been floating around for a long time. The Institute of Medicine [IOM] has dedicated an unprecedented amount of its attention to this idea, which culminated in a monumental report on the earning health-care system. The report gave rationale, data, and specific recommendations on how to move forward. ASCO has embraced the vision outlined by the IOM and others, and that’s the origin of CancerLinQ, the next evolution of our quality programs,” said Dr. Lichter.
Dr. Lichter explained that CancerLinQ’s core is the ability to measure quality in the oncology practice, measure outcomes, report in real time, and provide new methods of data exploration and hypothesis generation. “The real-time collection of quality and outcomes will allow the system to learn and improve the care of cancer patients by driving higher value and better outcomes,” said Dr. Lichter.
Rapid-learning System
“Although very complex, the core of the rapid-learning system is a circle that is really a classic feedforward-feedback cycle of quality improvement. It begins at the point of care where the data are collected by computer, transformed, and aggregated; then the data are analyzed and services are provided in compliance with important measures,” said Dr. Lichter.
The system, explained Dr. Lichter, has a second cycle attached to it. “Once the data are analyzed, it can go through peer review and feedback, generate new hypotheses, and do correlation and trend analysis. That second process begins to make our guidelines and guidance better,” said Dr. Lichter.
“Right now, as in most medical specialties, the majority of things we do are based on consensus or good solid hunches, not on level 1 clinical trial evidence. In fact, there are so many decision points in oncology that we will never fill in the grid solely through clinical trial data. However, with the second cycle of the rapid-learning system—where the correlation and trend analysis occurs—we can improve the effectiveness of our guidelines,” said Dr. Lichter.
The Prototype
After laying the foundation by engaging leaders in the oncology community, hammering out a business plan, and doing the legal and technology analysis, ASCO’s work on the prototype system began. Dr. Lichter said that the group set out to construct the prototype with five goals:
1. Provide “lessons learned about the technological and logistical challenges of a full-scale implementation of CancerLinQ
2. Demonstrate value-added tools, such as the ability to measure a clinician’s performance on a subset of the Quality Oncology Practice Initiative (QOPI®) in real time
3. Create new ways of exploring clinical data and hypothesis generation
4. Demonstrate our capability to develop rapid, real-time clinical support based on clinical guidelines and integrate them into an electronic health record system
5. Demonstrate the ability to capture and aggregate complete longitudinal patient records from any source and make use of the data.
“After taking the key points from the original feedforward-feedback cycle, we built a rather crude oncology [electronic health record system] from an open-source [electronic health record] so we could interact with the system and develop batch filing and processing of data. In other words, we developed automatic processing so we were able to understand the data that were being sent to us from various aggregates of medical records,” commented Dr. Lichter.
Analytic Component
Using an elegant visual platform called Galileo Cosmos™, Dr. Lichter demonstrated a drag-and-drop feature that allows the user to enter all pertinent clinical information about a particular patient and screen the information against a particular regimen. The program graphically illustrates the various outcomes in a multiplicity of clinical scenarios. Dr. Lichter toggled back and forth, visually comparing and contrasting the outcomes of different agents.
“With this kind of data capture, we can begin to show and analyze outcomes and answer interesting and complex clinical questions in real time,” said Dr. Lichter. After showing multiple slides that explored the interactive capabilities of the prototype electronic medical record—from practice guidelines to adverse events—Dr. Lichter noted that many of the key nodal points in a learning health-care system are currently feasible.
Work in Progress
Stressing that CancerLinQ is still very much a work in progress, Dr. Lichter said, “We have learned about the technological and logistical challenges involved in a full-scale CancerLinQ implementation. And we have demonstrated value-added tools, such as the ability to measure a clinician’s performance on a subset of QOPI measures in real time,” he noted.
“We have also shown our ability to develop rapid clinical decision support based on guidelines, and integrate them into a demonstration [electronic health record] system. And while there is much work to be done, we have shown that the core functions of a health learning system are within our grasp, setting the stage for fundamental and unprecedented transformation,” said Dr. Lichter. ■
Disclosure: Dr. Lichter reported no potential conflicts of interest.
