Case Study

History 2017-2018

Origin: The lucky spark was a chance encounter between former medical colleagues, Dr. Jay Schnitzer and Dr. Monica Bertagnolli, on an airplane in 2017. Bertagnolli was a preeminent cancer surgeon at Brigham and Women’s and was leading the American Society of Clinical Oncology (ASCO) and Alliance for Cancer Trials in Oncology (ACTO). Dr. Schnitzer was Chief Medical Officer and Chief Technology Officer at MITRE, with a growing portfolio of health information technology research.

Drs. Bertagnolli and Schnitzer had been circling around how to achieve a Learning Health System, where caring for every patient enables learning from every patient. Every person with any disease should have access to high-quality care and have the opportunity for their experience to help others in the future.

Drs. Bertagnolli and Schnitzer had also been circling around one of the fundamental challenges to achieving a Learning Health System: Lack of a single standard for collecting and sharing patient data electronically. They organized the first clinical requirements team convened under the auspices of ASCO. This team was co-chaired by Drs. Jim Chen and Doug Blaney, staffed with world-class experts from across the oncology space. The team's mandate was to agree on first use cases (a narrow scope was important), and on a minimal set of clinical data to address use case requirements. This data dictionary later seeded a new standard for cancer data: the minimal Common Oncology Data Elements.

Learnings

ASCO and MITRE, led by Drs. Bertagnolli and Schnitzer, became what we now call Champions for the new oncology specialty initiative. As champions, they:

• Committed their leadership
• Offered many of the key capabilities needed to start and succeed:
  • ASCO and ACTO – clinical
  • MITRE – standards and systems engineering
  • Both – convening power
• Had access to potential members of the community and the resources needed to develop and leverage mCODE:
  • ASCO and ACTO – access to almost all clinical cancer organizations and vendors
  • MITRE – deep expertise in standards and systems engineering
  • Both – strong links to Federal health agencies.
• Established the need for clinical expert consensus on very specific use cases and associated minimal data requirements.

History 2018-2019

First Project: EHR Endpoints for Cancer Clinical Trials (ICAREdata study). The original mCODE clinical team envisioned a small set of use cases in its initial, conceptual work. It was clear that success required executing a project where standardized data collection and sharing in the real world would test the value of the new data standard. The use case for the ICAREdata project was to capture and share, using the mCODE standard, high quality clinical treatment data from electronic health records (EHRs) that could be used for oncology clinical trials and other research.

ACTO and MITRE became the ICAREdata Use Case Champions. Over time, multiple health systems, clinical trials, Epic, pharmaceutical companies, NCI, and FDA participated in, hosted, and/or funded development and testing.

Initial work focused on standardizing and collecting just three key data elements (cancer type, disease status (e.g., progressing, stable...), and reason for treatment change), all modeled within the fledgling mCODE IG.

Learnings

ACTO, and MITRE understood that the best way to develop a cancer data standard is to focus on the small set of core data elements needed for specific use cases (e.g., specific challenges) that, if empowered by standardized data, could have a substantial impact.

ACTO and ASCO’s credibility and clinical network facilitated the engagement of outstanding clinical experts and vendors.

Adoption and Value in real-world settings was in the Plan from the start. This ultimately helped refine mCODE and workflows and led to greater participation and substantial funding.

This use case also helped identify the skillsets necessary during different phases of use case work.

mCODE Design: The ICAREdata Use Case provided helped establish an approach and structure for mCODE (around patient, assessments, disease, treatment, outcomes, genomics), and design principles.

mCODE V 1.0 IG was balloted in the Health Level Seven (HL7) standards organization in 2019 and benefited substantially from the ICAREdata experience.

mCODE design principles started to become clear during the ICAREdata Use Case and have improved and served mCODE’s evolution since. Principles include: (a) Focus on discrete data collected during patient encounters. (b) Acquire clinical expert input first and engage other experts as the IG is developed. (c) Base the IG on government and industry requirements and trends (e.g., FHIR, US Core). (d) Leverage the best existing clinical data modeling efforts. (e) Adopt global standard code systems. More details in the Design Principles section.

History 2019-2020

Cancer Data Summit: Interest in mCODE was starting to build and new use cases were being proposed. A small group of leaders from across the oncology space (patients, providers, researchers, payers, gov’t agencies, vendors, and others) convened for two days in October 2019 to brainstorm on the future of standardized cancer data, obstacles to success, and next steps. Below is one of the many storytelling roadmaps from the Summit. One outcome was support for creating a clinical requirements group (which became the mCODE Executive Committee) and an implementation group (which became the CodeX HL7 FHIR Accelerator).

Learnings

The Summit provided substantial benefits. Extremely useful feedback was provided by a diverse set of leaders. Noticeable alignment between diverse views happened during the two-day event. Strategies for next steps were shared. Many who attended the Summit expressed interest in staying involved, and many of those became Champions and participants in the future mCODE Executive Committee and CodeX Community.

mCODE Executive Committee and Technical Review Group: This body was organized under the leadership of ASCO and included clinical leaders and experts from major cancer-related societies, government agencies, and others as needed. This group followed from the original clinical requirements team convened by ASCO in 2019 (see above). The mandate continues to be to maintain and evolve the clinical data dictionary that underpins mCODE.

It has proven beneficial to have this strictly clinical group focusing on clinical requirements and interpretation. An expert can be an invaluable contributor to mCODE without being deeply involved in the technical and logistical work that CodeX would take on. However, a valuable, small subset of those involved in the Executive Committee have joined CodeX standards and use case work.

CodeX Community To complement the Executive Committee’s clinical leadership, MITRE led the creation of CodeX within HL7’s then-new FHIR Accelerator Program. CodeX was established, and still operates, with a focus on creating FHIR IGs, but also just as importantly a focus on implementing FHIR IGs into software and workflows, testing these in the field, and supporting adoption and value measurement in the real world. A lightweight process, Governance structure and Membership fee model were established. There were 9 founding members and other categories members by the end of 2020.

Establishing the goal of adoption and value of a standard, rather than focusing primarily on standards development, was and is not easy. This generally takes a greater commitment of resources and time from more parts of the community to achieve the goal. However, this goal resonated with the growing community, and ultimately paid off in new use cases, members, adoption, and value in the real world.

Initial Vendor Implementation: Epic and others started to integrate parts of mCODE into their software.

Demand for mCODE from large cancer centers (through ASCO, ACTO and the Summit) provided substantial motivation for vendors to deliver mCODE-infused systems to their customers.

History 2021-2024

Growth in number of CodeX Use Cases:

• Dec 2020: 5 Use Cases
• Dec 2021: 8 Use Cases (including first genomics Use Cases)
• Dec 2023: 12 Use Cases (including first cardiovascular Use Cases)
• CodeX Use Cases today

Learnings

CodeX leadership agreed to a use case selection process to help ensure use cases have the ingredients (e.g., the Tracks described on this website) to succeed. The table below lists use case start dates and the champions that drove them. Medical societies and government agencies were the most frequent champions.

Growth in CodeX Membership:

• Feb 2020: 9 Members
• Feb 2021: 20 Members
• Dec 2021: 30 Members
• Jun 2022: 40 Members
• May 2024: 50 Members

Two main drivers for attracting new members included promising progress within ongoing use cases, as well as new use cases which engage different segments of a given specialty (e.g., the radiology segment of the oncology space). Successful use cases engaged at least one (generally more) of each of the types of organizations needed to execute the use case.