Under the recently authorized fifth instance of the Prescription Drug User Fee Act (PDUFA V), the FDA has been directed to develop standardized clinical data terminology through open standards development organizations (i.e., the Clinical Data Interchange Standards Consortium (CDISC)).” For the FDA, these standards are expected to improve the efficiency of drug review by establishing a “common language through well-understood concepts and terminologies,” with the intent of facilitating data sharing and data pooling and enabling other improvements in pre-market analysis and safety signal detection.
While this is just one of many requirements included in PDUFA V, it offers the potential to become much more than a regulatory improvement initiative. In fact, it’s an important milestone in the evolution of clinical research in general, in offering the potential of developing a consistent level of knowledge and expertise among the research community.
If it even comes close to realizing its potential, the therapeutic area standards initiative may have a profound impact on nearly everyone involved in the clinical research process. In recognition of this, the recently-established TransCelerate Biopharma, Inc., a new, non-profit organization created by ten of the top pharmaceutical sponsors, has chosen to include active participation in this project among its initial investment portfolio of pre-competitive joint initiatives. Other organizations are also participating through CDISC and the Coalition For Accelerating Standards and Therapies (CFAST), a joint partnership of CDISC, the Critical Path Institute and the FDA, with the goal of developing and maintaining data standards tailored to individual diseases and therapeutic areas.
The key phrase in PDUFA V is the “common language” which should eventually influence the training, development and practice of clinical research professionals for years to come. While the use of advanced technologies have affected many aspects of bio-medical research in general, the practice of conducting clinical trials remains a comparatively low-tech and extremely time and labor-intensive endeavor. Often the success of a trial depends on how well it is designed (so that it can find sufficient patients and answer the right questions) and by the quality and integrity of the data collected. Both of these depend on knowledgeable researchers.
Standards Can Improve Training and Development
Typically, training programs for clinical research associates, study coordinators and even data managers center around explaining regulatory requirements such as Good Clinical Practices (ICH E6) and specific regulations such as 21 CFR 11. These education programs are often process-centric – what does each participant in the research process need to know in order to perform their part of the research study? Companies are required to document these standard processes for most research activities in the form of standard operating procedures (SOPs), which form the basis for much of the training provided.
Understanding the science of the research study is something that is assumed to come with academic education and on the job experience. But it’s hard to impart any consistency on previous academic studies and practical trial experience – the former depends on each individual student, teacher and curriculum, and the latter is picked up haphazardly along the path of a career.
But having a new set of therapeutic area standards can make a difference. These new standards can be much more than a prescribed way to name variables in a database – it can become an essential model framework for designing and setting up a new clinical study, and provide a consistent introductory context of what’s involved in each disease or therapeutic area, no matter who’s involved.
What's in the Box
This should become clearer by examining the proposed content of a full therapeutic area standard package. The process of defining a new standard begins by examining the range of experience of multiple trials, investigators, sponsors and protocols to try to determine the essential information concepts that characterize the disease and its treatment. These include the disease history, signs and symptoms of targeted patients, understanding the key measures and lab results that characterize disease progression or improvement, typical or expected adverse events, commonly used concomitant medications, and even information about potential genetic markers or possible environmental factors that may cause or exacerbate the disease.
This leads to a model describing the most essential elements of the disease, and how each element relates to others, which can be visualized in a mind map that shows how such concepts interrelate. A mind-map can quickly allow clinicians and research participants to grasp the essential information concepts that comprise the disease and determine at a glance what’s important and what’s missing. Thus, the mind map rapidly establishes a common level of understanding as a basic, non-technical orientation suitable for all.
The next essential step is the creation of definitions. Each concept must be precisely and uniquely defined. Of course, agreeing on common definitions is also one of the risk areas – it can certainly be a challenge to reach consensus on defining concepts that affect so many disparate clinicians around the world. Yet the lack of common definitions is one of the most significant barriers toward consistent interpretation and effective reuse of data – CRFs and databases are notoriously guilty or using the same variable or term to describe slightly different nuances of flavors of a single concept.
Well-defined research concepts can later be developed into questions on a CRF, which correspond to variables in a database. But the standard promises to provide traceable consistency between basic concept (what do we want to know?) to CRF (how do we collect it?) through database representation (how do we process, review and analyze the data?). The development of a concept into questions or variables needs to be performed against the backbone of a robust information model (provided for CFAST projects by the BRIDG model) to ensure consistency, reusability and a persistent structure. And these developed representations of the concepts need to be bound to precise medical and research terminologies – so the same words are used to represent the same meaning no matter where they originate.
Many concepts are hardly unique to research, but, in fact, are shared with the larger world of healthcare delivery. Yet observed healthcare data points may not be optimal for research purposes even if the core concepts are the same – or more likely, similar. So the representation of the standard must be capable of quickly building and characterizing relationships between identical and similar concepts – an application ideally suited for the semantic web.
But it’s not sufficient just to provide these concepts, relationships and terminologies as static reference documents – it’s necessary to have an online knowledge repository that manages and provides access to all of this information in multiple forms so it can be utilized efficiently at each stage of the process – and afterwards, to support secondary research uses of data.
Once all these individual molecules of information about a therapeutic area are made available, it’s possible to enrich the standard with standard CRFs, sample protocol study designs, and representative analysis models. What eventually should emerge is a comprehensive kit comprising a rich set of tools defining research in a specific therapeutic area – a user’s guide of patterns to follow, and the building blocks to assemble a study. And it is this knowledge resource that can and should be made available to the full research community to establish a basic level of proficiency in each disease area.
Taking advantage of information resources
Realization of this vision should have significant benefits toward efficiency of the research process and the quality of research data. So it’s good to have a congressional mandate to spur it along. And what a resource this can be toward the future training and development of clinical research professionals. It should now be possible for CRAs and study coordinators to gain a common understanding of each therapeutic area that their studies are exploring, which should improve communication and data quality – and the overall skills and knowledge of each participant. And by ensuring open access to these standards, the benefits can be extended globally, to academic research as well. The ability to access a complete kit that describes most of what you want to know about conducting a typical study in a particular therapeutic area will establish a common level of basic scientific proficiency among all study participants. No longer will training need to focus primarily on process and rules – it can now also include the scientific underpinnings that are relevant to each disease area.
Now, some fear that such comprehensive standards will inhibit and constrain research. No standard can possibly predict every question that may be relevant to a research project, but rather than constantly reinvent the same questions over and over, isn’t it more sensible to look into a single, gold standard global information resource and pick and choose what you need and only define what’s missing? And equally important is a feedback mechanism to ensure that newly identified concepts are added to the knowledge base over time, so that each subsequent study learns from the last.
Indeed, how can we envision effective secondary uses of research information otherwise?
The research professional of the future will need to be plugged into such a research semantic infrastructure. Knowledge must be captured consistently and reusably throughout the research process. And researchers must acquire the necessary skills to enable themselves to easily plug in, early and often.
Of course, all this requires that we actually succeed in developing these therapeutic area standards, and that we achieve near universal adoption. So we’ve got quite some work to do first.
You can learn more about Wayne Kubick, and CDISC here.