New ASCO System Confronts Lack of EHR Interoperability

Daniel Martin, MD

With the recent launch of a free, open-source system for standardizing the electronic storage of patient records, the American Society of Clinical Oncology (ASCO) hopes to improve information flow across electronic health record (EHR) platforms and curtail the harms of information blocking by private aggregators and vendors of patient information.¹

Short for Minimal Common Oncology Data Elements, mCODE is ASCO’s response to private, nonuniversal systems of patient data collection that, owing to nonconforming data classification systems or vendor practices, have impeded oncology practices and investigators from sharing and accessing information. Findings from ASCO’s 2017 Oncology Practice Census indicated that almost half (40%) of oncology practices were unable to accept any patient information from other practices because of these barriers.² Providing an example of how nonconforming data systems can impede information sharing, ASCO noted that in a sampling of just 30 practices, total neutrophil count was listed under 76 different headings.³

Figure 1. EHR Interoperability in Oncology⁴ (Click to Enlarge)

Paywalls for access to patient data within private networks of oncology practices constitute significant hindrances for information circulation.¹ ASCO has expressed strong concern about physicians being charged by EHR vendors and document archiving services to access and exchange their own information. Charging for data access effectively constitutes a form of information blocking because it places “a clear roadblock in front of clinicians who are motivated to participate in quality measurement reporting and who wish to support research and improvements to quality of care,” ASCO said.³

Healthcare analytics and the aggregation of data pools have become big business: The practice of mining patient data and trading it for payment is not only common, but also profitable. To gain access to the valuable data that are accumulated, organized, and stored within EHRs, information seekers often must join a vendor’s network, which comes at a price.

Flatiron Health and Cota Healthcare, both in New York, New York, are 2 companies that have sought to build healthcare information businesses that depend on information collected from EHRs. These pools of patient data are analyzed for important trends in the use of oncology drugs that can give physicians insight into optimal patient selection and potential outcomes. Companies such as these could potentially lose exclusivity if mCODE is widely adopted.

Flatiron did not respond to requests for comment, but a representative of Cota said he believes his company can work with ASCO to promote freer access to data without impairment of its business model. Andrew Norden, MD, MPH, MBA, chief medical officer of the company, said that if mCODE is widely adopted, it will likely make it easier for Cota, physicians, and investigators to extract and analyze data components from large patient data sets.

Such corporate activities provide value to oncologists because a large data pool by itself is of little use, said Daniel Martin, MD, chief medical information officer of Seattle Cancer Care Alliance, in an interview with OncologyLive^®.

“Outside of aggregating various secondary databases, tumor registries, or billing databases, it’s hard to get an aggregate picture of how an oncology patient is doing [relative to others] across the country,” he said. “The data aren’t granular enough to do that and that’s exactly why companies like Flatiron decided they were going to manually unearth this data, because it has both health and commercial value.”

A nonproprietary offshoot of ASCO’s CancerLinQ data exchange and analytics venture, mCODE was created in collaboration with the MITRE Corporation and the Alliance for Clinical Trials in Oncology Foundation.⁵ It is being piloted at cancer centers nationwide, including Partners HealthCare in Boston, Massachusetts, and Intermountain Healthcare in Salt Lake City, Utah. In an attempt to deploy the system more broadly, ASCO has invited practices across the country to adopt the open-source code.

Robert S. Miller, MD, the medical director of CancerLinQ, said he doesn’t expect mCODE to actually compete with private vendors. “Commercial companies won’t need to abandon what they’re doing,” he said in an interview with OncologyLive^®. “Our philosophy with mCODE is that it’s a data standard, a way of describing which particular data elements we deem to be the most important, and then defining the specific terminologies or vocabularies that we use to delineate those elements.” In other words, mCODE represents a bridge toward deep analysis.

“Publishing mCODE is an important step toward realizing incredible insights into treatment that the 97% of [patients with] cancer not participating in clinical trials could generate,” Jay J. Schnitzer, MD, PhD, MITRE’s chief technology officer, told The ASCO Post in a recent interview.⁵ “Shared treatment experiences could be turned into information that patients and clinicians can use to better navigate care options.”

However, it’s unclear how much overlap there will be between what mCODE does and what the commercial sector would like to keep proprietary. Providing deep insights into patient treatment is also a stated objective of some commercial platforms.

Flatiron’s 4-part OncoCloud suite, composed of OncoAnalytics, OncoEMR, OncoTrials, and OncoBilling, tracks and analyzes data from more than 2 million patients with cancer.⁶ Flatiron organizes unstructured real-world clinical data and markets the information that it amasses and stores in the software platform to cancer care providers and life science companies. Early this year, Flatiron announced an extension of an agreement with the FDA to sift through deidentified patient datasets curated from EHRs to support regulatory decision making. The stated goal was to “generate insights into cancer treatment trends and clinical outcomes in the United States” using real world data (RWD) on patients.⁷

A Shared Data Language

mCODE is ASCO’s answer to the need for a common data language. In drawing from a library of standard, widely available medical terminologies, mCODE permits physicians to retrieve data by querying familiar clinical terms, no matter the EHR used to conduct the search. In their design of mCODE, ASCO and collaborators identified 6 core classification domains of critical importance: patient demographics and other characteristics; disease details; genomic characteristics of cancers; relevant laboratory tests and vital signs; treatments such as surgery, radiation, and drugs; and patient outcomes.⁵

ASCO asserts that adoption of the categories, which span the stages of cancer diagnosis and treatment, will homogenize how patient data are registered in EHRs.⁵ This initial set of standards and specifications, coupled with mCODE’s incorporation of Fast Healthcare Interoperability Resources (FHIR), a data standard for the electronic exchange of healthcare information, will allow a seamless transference of patient data across EHRs, ASCO said.^5,8 mCODE’s common data language and interconnectivity that incorporate FHIR design will also support EHRs in capturing genomic data yielded by molecular diagnostics and treatment.⁵

“We tried very much to use terminologies that were in the public domain so there would not be proprietary terminologies used that cancer centers and others [who] want to implement mCODE would have to pay for, or would be restricted by licenses,” Miller said.

Figure 2. Most Commonly Used EHRs in Oncology⁴ (Click to Enlarge)

Figure 3. Oncology Practice Managers Hesitate to Share Data for Research or Commercial Purposes⁴ (Click to Enlarge)

The variance of EHR implementation noted in Genentech’s report further complicates data sharing; 20% of practices said they had more than 1 EHR.⁴ Epic, OncoEMR, and Cerner were the 3 most widely used EHRs among a population of 194 practices, with 35%, 15%, and 13% of respondents indicating that they employ these systems, respectively (Figure 2⁴).

Patient Data Aggregation: A Commercial Hotbed

Broad implementation of mCODE will be vital to the success of ASCO’s project of improving data circulation. “We’re looking to have mCODE permeate the cancer ecosystem: practitioners, cancer centers, and commercial companies, whether those are laboratories or tech companies,” Miller said. “Our aspiration is that everyone will look at the core set of mCODE data elements and agree that they’re critical for describing cancer data in [a] standard way.”

This includes commercial data aggregators, Miller added: “We welcome private [vendors] to participate [in mCODE] and hope that they are willing to adopt mCODE as a data standard.” ASCO has already “started conversations” with for-profit companies to explore how mCODE and independently owned platforms may collaborate in the future, he said.

“[Cota is] engaging with mCODE leaders about the potential to work together,” Norden said. “I think [our] discussions will continue because we’re enthusiastic about this. It’s in all of our interests to ensure that these [data] elements are routinely captured from realworld data RWD sources.”

Like Flatiron, Cota applies analytics to fragmented RWD to interpret patient information. Data are refined into the Cota Nodal Address system,⁹ which isolates homogeneous cohorts of patients with cancer to identify what Norden terms “unwarranted” variation in treatment plans, outcomes, and costs. Cota accumulates data from physician notes; radiology, pathology, surgical reports; and genomic testing results, then assigns each patient a digital numerical identifier. The system is designed to facilitate precision medicine and lower costs of care.

Increased data sharing will also depend on participation from oncology clinics. Nearly half (40%) of oncology practice managers are not currently willing to share their de-identified EHR data with pharmaceutical companies or others for research or commercial purposes, according to Genentech’s report.⁴ However, 37% said they would consider sharing the information if it were part of a software solution that solved practice needs, aligned incentives with an integrated delivery network or payer, or involved “sufficient monetary compensation” (Figure 3⁴).

Knocking Down the Barriers

Along with its interest in a standard language for the exchange of oncology patient data, ASCO hopes to lower barriers to EHR interoperability such as per-transaction fees, imports and exports of information between platforms, and contractual requirements that give EHR companies exclusive license to use a healthcare provider’s data.¹

ASCO has supported provisions within the 21st Century Cures Act (Cures Act), signed into law on December 13, 2016, to advance interoperability; support the access, exchange, and use of EHR information; and address occurrences of information blocking.¹⁰ In an editorial, ASCO noted that the Office of the National Coordinator for Health Information Technology has a mandate to reduce information blocking, as established by section 3022(a) of the Cures Act.³ ASCO encouraged practitioners to report impediments “imposed by health information technology vendors to access and exchange electronically accessible health information for authorized use under state and federal law.”

ASCO has also supported a Centers for Medicare & Medicaid (CMS)—proposed companion rule to the Cures Act that would address interoperability issues to improve access to and quality of the information that CMS considers necessary to Americans’ ability to make informed healthcare decisions. This would include data about healthcare prices and outcomes.^11,12

Integrating mCODE

The first version of mCODE’s core set of data elements, released on June 1, can be accessed at mCODEinitiative.org. “Through the pilots, we’ll learn what works and what doesn’t work, and what needs to be modified,” Miller said. He added that there have been “several hundred downloads” of the data transfer and classification system since its launch.

References

1. Policy issue brief: interoperability of electronic health records. making patient healthcare information readily accessible to ensure high-quality care. American Society of Clinical Oncology website. asco.org/advocacy-policy/asco-in-action/policy-issue-brief-interoperability-electronic-health-records. Published October 11, 2017. Accessed September 12, 2019.
2. Kirkwood MK, Hanley A, Bruinooge SS, et al. The state of oncology practice in America, 2018: results of the ASCO Practice Census Survey. J Oncol Pract. 2018;14(7):e412-e420. doi: 10.1200/JOP.18.00149.
3. Rubinstein WS. CancerLinQ: cutting the Gordian knot of interoperability. J Oncol Pract. 2019;15(1):3-6. doi: 10.1200/JOP.18.00612.
4. The 2019 Genentech Oncology Trend Report. 11th ed. South San Francisco, CA: Genentech; 2019. genentech-forum.com/content/dam/gene/managedcare/forum/pdfs/Oncology-Trends/Genentech_2019_Oncology_Trend_Report_Digital_version.pdf.
5. 2019 ASCO: mCODE, a core set of common cancer data standards, established. The ASCO Post. June 6, 2019. ascopost.com/News/60120. Accessed July 26, 2019.
6. Community oncology. Flatiron website. flatiron.com/oncology/. Accessed July 26, 2019.
7. The FDA and Flatiron Health expand real-world data cancer research collaboration [press release]. New York, NY: Flatiron Health; February 25, 2019. flatiron.com/press/press-release/the-fda-and-flatiron-health-expand-real-world-data-cancer-research-collaboration/. Accessed September 16, 2019.
8. FHIR overview. HL7 International website. hl7.org/fhir/overview.html. Published December 27, 2018. Accessed September 13, 2019.
9. Providers. COTA website. cotahealthcare.com/providers. Accessed July 26, 2019.
10. 21st Century Cures Act: interoperability, information blocking, and the ONC health IT certification program. Federal Regist. 2019;84(42)7424-7610. To be codified at § 170.315(b)(11).
11. Medicare and Medicaid programs; Patient Protection and Affordable Care Act; interoperability and patient advantage for Medicare Advantage organization and Medicare managed care plans, state Medicare agencies, CHIP agencies and CHIP managed care entities, issuers of qualified health plans in the federally-facilitated exchanges and health care providers. Federal Regist. 2019;84(432)7610-7680. To be codified at 45 CFR 170.213.
12. ASCO submits comment letter on recent senate hearing on electronic health information. American Society of Clinical Oncology website. asco.org/advocacy-policy/asco-in-action/asco-submits-comment-letter-recent-senate-hearing-electronic-health. Published March 27, 2019. Accessed September 12, 2019.

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,” ASCO President Monica M. Bertagnolli, MD, said in a statement.⁵

Currently, data from most of the nearly 15 million people living with cancer in the United States are housed in some kind of EHR.⁵ The differences in the types of data that EHRs record, along with inconsistent terminologies and formats for describing and classifying this information, lead to cross-system incompatibility. The result: impaired care coordination and a lack of clinical insight that could lead to suboptimal outcomes and failures to perceive practice-advancing developments.

Two years after the census, interoperability remains a problem. A recent report from Genentech showed that just 13% of practices’ EHRs were interoperable with out-of-area hospitals and oncology centers of excellence (Figure 1).⁴