In an effort to improve data sharing about cancer and improve the quality and coordination of patient care, three organizations have established a core set of EHR data elements and recommended technical specifications for capturing and reporting the characteristics, treatments, and outcomes of cancer patients.
The Minimal Common Oncology Data Elements (mCODE) initiative is a collaboration between the American Society of Clinical Oncology Inc. (ASCO), its wholly owned nonprofit subsidiary CancerLinQ LLC, the MITRE Corp., and the Alliance for Clinical Trials in Oncology Foundation (Alliance Foundation). They released the initial set of common cancer data standards and specifications June 1 at ASCO's 2019 Annual Meeting in Chicago and published them online at mCODEinitiative.org.
The effort's leaders note that data from most of the nearly 15 million individuals living with cancer in the U.S. is contained in EHRs of some kind. But many EHR systems in use prioritize the collection of different types of data, or use different terms to describe the same type of data, or collect data in different formats, making them incompatible with one another. This incompatibility dramatically limits the ability of cancer researchers and doctors to learn from patient records, hinders care coordination, and adds to the administrative burdens and costs for all users, but especially practices and patients.
"Progress and quality care for patients with cancer can hinge on our ability to seamlessly share patient data among doctors, hospitals, and researchers. But data sharing is much more difficult, if not impossible, when EHR systems contain incompatible information," said ASCO President Monica M. Bertagnolli, M.D., in a prepared statement. "With mCODE, we're bringing the oncology community together around common data standards that will bring us one step closer to our goal of learning from every patient with cancer."
mCODE is being developed collaboratively and the mCODE elements are available for use free of charge. mCODE's data elements are designed using standard, widely available medical terminologies, which means that a physician's clinical query across different EHRs that each use mCODE should convey the same meaning and retrieve similar patient details. The mCODE data elements are grouped into six main areas to describe the patient's cancer journey: "patient" (demographic and other characteristics), "disease" (specific details regarding the cancer), "genomics" (molecular characteristics), "labs/vitals" (relevant laboratory tests and vital signs), "treatment" (surgery, radiation, and drug and other treatments), and "outcomes" (e.g., current cancer status and survival).
To incorporate these standards, mCODE will provide both a common data language and an open source, nonproprietary data model based on Fast Healthcare Interoperability Resources (FHIR) for interconnectivity across EHR systems. The standard elements also support EHRs in capturing and processing increasingly complex genomic data generated by advances in molecular diagnostics and treatment.
"The Alliance Foundation is pleased to be implementing mCODE data structures in a research program that integrates clinical data captured via the EHR with data obtained during conventional prospective cancer clinical trials," said Steven Piantadosi, MD, Ph.D., who is the Scientific Director of the Alliance Foundation ICAREdata Project, in a statement. "This work will validate the use of mCODE as a tool for clinical research and will also establish a platform for conducting research that includes many patients not currently represented in conventional prospective clinical trials."
mCODE is being piloted at cancer centers across the country, including Partners Healthcare in Boston and Intermountain Healthcare in Salt Lake City. In the coming months the results of these pilots will be used to refine mCODE as it is deployed more broadly.
To help guide the evolution of mCODE, the mCODE Executive Committee created the mCODE Council. The Council will advise the Executive Committee, sponsor use cases, and provide input regarding future data elements and use cases. More information, including an application to join the Council, is available at mCODEinitiative.org.