Healthcare Standards Development
The Value of Nurturing Collaboration
Two standards developing organizations—CDISC and HL7—demonstrate that collaboration, while challenging, delivers benefits to standards developers, end users, and ultimately patients.
by William Edward Hammond , PhD, Charles Jaffe , MD, PhD, and Rebecca Daniels Kush , PhD
Technical standards are essential to improving healthcare. For health IT to reduce medical errors and risk to patient safety, improve access to medical records, and support innovations in “individual-based” care, its tools must adhere to certain data interchange standards. Standards also enable aggregation of information from disparate sources and sophisticated reviews of such information to glean knowledge that can inform clinical decisions.
Since the mid-1980s, a number of organizations have formed to develop such standards. Often these standards developing organizations (SDOs) form because of a perception that a new requirement is not met by an existing organization. The unfortunate result is a multitude of SDOs, which risks generating industry confusion over standards adoption rather than enhancing interoperability among disparate communities.
The multiplication of organizations happens naturally enough. At the time of an SDO’s creation, the reasons and need for it seem realistic, at least to the individuals creating the organization. Over the subsequent years, its scope widens to meet the needs of its domain. Globally, similar SDOs are created to meet national and regional requirements.
There has been limited collaboration to date. However, collaboration in the generation of harmonized standards from the very onset of development can be mutually beneficial, complementary, and valuable to the entire industry. The success of this approach is exemplified by the relationship of Health Level Seven (HL7) and the Clinical Data Interchange Standards Consortium (CDISC).
Recognizing common interests, the two organizations agreed to share resources and build upon common interests. Joint work groups serve as forums in which members of both organizations harmonize standards and work together on common products.
The Value of Collaboration
Collaboration can save the standards developing community resources and improve the quality of the standards they develop.
Most of the healthcare and informatics community feel that standards developing efforts have been unnecessarily delayed and often are achieved with disappointing results. There are many gaps in the set of needed standards. One challenge is limited resources.
Most standards development organizations only meet face-to-face a few times each year. Clearly, volunteer time is a precious commodity, and staff contributions to the production and support of standards generation varies. The need to expand scope to address the gaps and to produce standards more efficiently requires an increase in paid staff, which cannot be achieved without an increase in revenue. In existing models, these funds have been almost exclusively generated from membership dues and educational fees.
Simple management of an SDO also requires increasing resources. Unfortunately, resources are limited, as funding, staff, and volunteers are stressed. This is invariably exaggerated with SDOs competing for those funds.
However, there is significant value to working together. This is true not only from the point of view of resource management and increased efficiencies, but also by ensuring harmonized and noncompetitive standards. For example, CDISC and HL7 have effectively increased the number of face-to-face meetings and teleconferences through collaboration.
Collaboration also leads to greater consistency in the standards produced. A global common information model and a common data model invariably increase consistency across standards. Reuse of models and tools provides a direct benefit for the organization and the end user. Consistent and compatible standards are a consequence of common models.
Collaboration also leads to the development and sharing of common tool sets. This is particularly true in those circumstances in which standard development is based upon common models and a common methodology.
Most SDOs in the US have a different standards development approach based upon the origins and focus of the organization. For example, the National Council for Prescription Drug Programs includes principally individuals from the prescription drug reimbursement domain. CDISC is largely comprised of biopharmaceutical and academic clinical researchers and technology and service providers for clinical research. HL7 is represented by government agencies, health IT vendors, and care providers. The standards organization X12 is almost exclusively limited to payers focused on claims reimbursement.
As a consequence, a natural duplication or overlap of requirements and desired uses for standards evolved. At times, standards developed by each entity became more duplicative and ultimately more disparate and competitive. When SDOs collaborate in the production of standards, the end user always profits from the shared expertise across these often artificial boundaries.
Benefits to End Users
Ultimately, SDO collaboration results in the creation of a single standard for a single specific purpose, thereby reducing or eliminating competing and often unnecessarily duplicative standards. For the developer or end user, this collaboration removes the artificial decision of which standard to choose and implement. In the final analysis, the industry benefits and the motivation to more rapidly adopt and implement appropriate standards becomes much greater. Both vendors and users are able to plan a path for both adoption and deployment of these common standards.
SDOs working together in meaningful and complementary ways will benefit a number of different communities and stakeholders. For example, in a large technology company, standards enable faster development and more rapid time to market. This almost always achieves lower cost structures. Deploying standardized interfaces rather than customized ones in a vendor solution results in a software solution that also will be more scalable.
In a global biopharmaceutical company, the development and adoption of health information standards will enable better decision making, reduce redundancy and rework, enable more effective clinical research at lower cost, and bring better therapies to patients more rapidly. For regulatory authorities, standards improve reviews, enabling the use of standard tools for viewing data.
Within academic research institutions, sharing of research information is being encouraged through governmental funding agencies. These grants encourage improvement in the way biomedical research is conducted and promote the engagement of communities in clinical research. The most effective means of exchanging research data is through the use of common standards. Standards harmonization that enables the secure flow of clinical information from electronic health records can benefit clinical research studies, safety reporting, and other use cases.
For example, technical standards can ensure that a subset of appropriately de-identified data can be aggregated at an academic center or study sponsor site to assess the effectiveness or safety of a therapy. Standards for regulated clinical research (such as the CDISC Operational Data Model) are accompanied by an audit trail, allowing clinical study monitors or regulatory authorities or auditors to trace the data they receive in support of a marketing application back to its origin to ensure accuracy and integrity of the information they are reviewing before approving a new product.
Barriers to Collaboration
The barriers to collaboration are many. Most SDOs depend on volunteers, who often may have a limited view of the broad impact of standards development, either nationally or internationally. In addition, all SDOs take direction from their stakeholders, whether it be government, directorates, or customers. These stakeholders perceive value in very different ways, which places very competitive forces on overarching development strategy.
In addition, every SDO is motivated to produce key standards that are implemented widely. That drive for survival is perhaps the most important barrier to collaboration. In simple terms, an SDO can be driven by a fear of becoming irrelevant. Most groups are willing to address the issues of cooperation and collaboration, but turning that willingness into action is a consistent challenge. At the most simplistic level, the proposition of a merger may only mean that the more dominant SDO would be willing to oversee the dissolution of its competitor and the adoption of its own long-term strategy, leadership, rules, and governance.
Another barrier is the inherent belief that one organization’s end users are fundamentally different from those of another organization—for example, more limited in scope or more diverse. Most frequently this leads an SDO to reject initiatives that appear to be less relevant. Inclusion of nonhuman (animal) subjects is one such barrier that often leads to contentious opposition to collaboration.
Finally, a significant, often critical barrier is the difference in governance models, including policies for balloting and for publishing standards. In scope, one organization may support draft standards for trial use, while another does not provide for this option. Questions about leadership, work process, and joint work structures may be difficult to resolve. The treatment of intellectual property, particularly when shared, may become contentious. Even access to the shared standard, particularly if it is freely given to the public domain by one organization, may play havoc with the business model of the other.
Examples of Productive Collaboration
The value of collaboration almost always outweighs the barriers, regardless of the seeming significance they seem to impose. The leadership of the respective standards development organizations and associated groups must coalesce to define fruitful collaboration and to minimize or completely overcome the multiple barriers, either real or imagined. If it cannot collaborate appropriately and productively, the health data standards industry risks increasing irrelevance and could ultimately fail to meet the real-world needs of its stakeholders. The collaboration between HL7 and CDISC offers one example of how SDOs can work together for the benefit of their end users.
HL7 was formed in March 1987. At that time the ASTM group E31 was focused on a standard to transfer laboratory data from commercial laboratories to healthcare facilities. The individuals who founded HL7 were interested in creating a larger scope of activity that would address all data interchanges required to support the functionality of hospital information systems. As a consequence, HL7 created a standard that overlapped with the E31 efforts. Eventually, this redundancy had to be resolved.
HL7, from its beginning, has been an open organization, welcoming any group to join to create standards that were within its scope. In some cases, HL7 expanded its scope to accommodate new activities when circumstances demanded it.
CDISC was founded to fill an identified need for data interchange standards within the clinical research domain. In 1997 an initial CDISC volunteer group began working with the US Food and Drug Administration, biopharmaceutical companies, academic research organizations, and other clinical research stakeholders including technology and service providers and academic research institutes to develop clinical research standards.
CDISC immediately became a growing and productive organization. However, in 2000 HL7 and CDISC leadership identified common interests and formed an affiliation. This affiliation grew stronger with the creation of the Clinical Trials Special Interest Group—now the Regulated Clinical Research Information Management (RCRIM) Work Group—a forum in which members of both organizations harmonize standards and work together on common products. The BRIDG domain analysis model, described below, is one product of that shared interest.
The collaboration has not always been easy or straightforward, but it has been productive. Since the earliest efforts to establish common benefits, both organizations have sought to complement rather than dominate the relationship. Very early in the process there was a recognized need to leverage the unique skill sets that the respective membership offered.
Historically, CDISC and HL7 had already produced essentially competing XML standards. An initial attempt to harmonize these was unsuccessful. In exploring what failed in greater depth, the two organizations identified a number of solutions. These resulted in a recommendation by an HL7 expert, which led to the development of a clinical research domain analysis model. This model represents the domain of clinical research within the context of the HL7 Reference Information Model (RIM). Equally important was that the collaboration process also served to ensure harmonization among the various CDISC content standards.
In the interim, an HL7 V3 message was developed as a transport method for the CDISC LAB Content Model. This standard facilitates the electronic transport of clinical research laboratory data in bulk from central laboratories to research sponsors based upon a CDISC standard content format. In addition, end users can choose from multiple implementations (transport methodologies) including CDISC XML, HL7 V3 message, ASCII, or SAS. In an effort to extend the LAB standard to include genomics data, CDISC and HL7 initiated a collaborative Clinical Genomics Work Group that has developed clinical genomics standards harmonized from the start.
The BRIDG Model
The organizations developed more than specifications and technology. They created a framework for collaboration that has grown in concept and in scope. Only months following the onset of a unified effort to create a domain analysis model for clinical research, the National Cancer Institute joined with CDISC and HL7 on the effort. Because the scope and goals were consistent with the development of applications for the National Cancer Institute’s Cancer Bioinformatics Grid (caBIG), further collaboration became both practical and self-sustaining.
The model became known as the BRIDG (Biomedical Research Integrated Domain Group) model, and in 2005 it was deemed to be the domain analysis model for the HL7 RCRIM Technical Committee (now called the RCRIM Technical Work Group). Shortly thereafter, the Food and Drug Administration became a key stakeholder.
Although modeling is initiated outside of the BRIDG and it is sequentially harmonized by a joint harmonization workgroup, the BRIDG model represents the domain of clinical research in the context of healthcare. Its scope is protocol-driven research and related regulatory artifacts; it also includes safety reporting. At present, the BRIDG model is under submission within ISO through the Joint Initiative Council, described in more detail below.
The Protocol Representation Standard
Another joint project between HL7 RCRIM and CDISC is the development of a protocol representation standard. Begun in 2002, the project’s specific objective was “publication of a standard, machine-readable model for protocol representation that will enable interchange of this data among systems and stakeholders.” The project covers the lifecycle of the protocol, from development through reporting, including trial registration and study design.
This global standard began as a simple spreadsheet that included a set of elements found in all research protocols, as well as the element definitions, the rational for their inclusion (e.g., international guidelines), attributes, and similar relevant information. As the work progressed, it became clear that a model for expressing the relationships among these elements was needed.
This CDISC-HL7 RCRIM standard was one of the first to form the basis of the BRIDG model, which was in its early stages at that time. The National Cancer Institute Study Calendar work was also harmonized into BRIDG at the same time. The Protocol Representation model is the first CDISC standard to have emanated directly as a subset of the BRIDG model, rather than an existing standard that was later harmonized into BRIDG. After all, the protocol is the core of a clinical research study.
Another example of effective collaboration between HL7 and CDISC is the development of terminology that supports the requirements of clinical research and yet meets the vocabulary needs of the healthcare community. Other agencies and organizations have also been involved in global terminology development in order to achieve a global set of terminologies that are openly available for global clinical research studies and for those who need to collaborate and support interoperability worldwide. In addition, CDISC standards are now bound to a terminology framework that supports the safety data domains.
Finally, there is growing interest in developing this terminology to support specific therapeutic areas. For example, a terminology supporting tuberculosis diagnosis and treatment has been created through the collaborative efforts of CDISC, HL7, the Global TB Alliance, and the National Institutes of Health through a grant to Duke University. This successful approach is now being implemented in developing terminology for other therapeutic areas through CDISC and the HL7 Clinical Interoperability Council, which engages clinicians with domain expertise rather than knowledge of computer science.
The Continuity of Care Document
The Continuity of Care Document (CCD) is one of the most productive outcomes of the Healthcare Information Technology Standards Panel work to harmonize competing standards. For several years, HL7 has refined a unique computing environment, now referred to as the Clinical Document Architecture (CDA). During this same period, ASTM has developed the Continuity of Care Record (CCR), which provides XML representation of the health record of a single patient, including the diagnoses, medications, and allergies.
With incentive from the Healthcare Information Technology Standards Panel to serve the health IT community through harmonized standards, HL7, ASTM, and others have developed the CCD for the purpose of having a single standard for exchanging core information among physicians. With the CCD, different EHR products in different hospitals can exchange important information between physicians, such as in the case of a referral, in an efficient and accurate manner. This can save patients and caregivers trips between hospitals.
The IHE RFD Integration Profile
In another technical arena, collaborative development between IHE (Integrating the Healthcare Enterprise) and CDISC has led to the development and implementation of a mature integration profile. The Retrieve Form for Data Capture (RFD) profile was developed to facilitate the use of electronic health records for clinical research (electronic case report forms) and related secondary use cases such as safety reporting, disease registries, biosurveillance, lab and image exchange, and postmarketing surveillance.
By deploying the CDISC Operational Data Model (ODM) and linking the Clinical Data Acquisition Standards Harmonization (CDASH) standard for data acquisition in clinical studies with CCD in a functional integration profile, the contents of the electronic record derived from a patient care setting can be leveraged to pre-populate specified fields in a research form.
The RFD was developed to streamline workflow for the clinician who must report a clinical adverse event or wishes to participate in a clinical research study. This means that physicians who may not currently participate in research studies due to the associated administrative burden may be able to easily contribute to research in the future. This standard has now been balloted and approved through ISO TC215 and has been endorsed by the EHR Association.
New Collaborative Opportunities
There are several important collaborative opportunities that have recently been adopted by the Joint Initiative Council, which currently includes a broad cross-section of standards development organizations. The council was created to encourage collaboration among these SDOs, in particular “to enable the common, timely health informatics standards by addressing and resolving issues of gaps, overlaps and counterproductive standardization efforts.”
The work is enabled through a Joint Working Group and is supported by the respective standards development organizations. Projects recently identified include the harmonization of data types across models, the harmonization of reference models, and the development of a global reporting standard for adverse events (pharmacovigilance). BRIDG has been endorsed as a Joint Initiative Council project.
Important evolving relationships are well on their way to implementation. The value of collaboration for SDOs far surpasses the numerous challenges yet to be overcome. The benefits include the conservation of resources and the creation of significant efficiencies, and ultimately they extend to the measurable value to patients. Research is an important component of the healthcare continuum, because it both informs healthcare community and drives both quality and safety.
William Edward Hammond is chair of the HL7 board of directors. Charles Jaffe is CEO of HL7. Rebecca Daniels Kush is president and CEO of CDISC. The authors would like to acknowledge the support of the HL7 Outreach Committee for Clinical Research.