Research Terminology Must Harmonize to Advance Science

What one researcher describes as asthma, another might call reactive airway disease. Unfortunately, such examples demonstrate the lack of a common pediatric research terminology, which presents a barrier to data reuse, sharing, and integration.

Thought leaders at The Children’s Hospital of Philadelphia published a state-of-the-art review article in the March issue of Pediatrics that encourages the research community to start talking the same language. They described a project under way, called the Pediatric Terminology Harmonization Initiative, that aims to provide uniform definitions and descriptions of clinical observations and data in pediatric research.

The National Institute of Child Health and Human Development is coordinating the effort in collaboration with the Enterprise Vocabulary Services of the National Cancer Institute, in order to fill a gap in the pediatric research infrastructure in the U.S.

“It is of critical importance to address this gap if we’re going to advance our scientific understanding of the causes of childhood disease,” said Christopher B. Forrest, MD, PhD, professor of pediatrics at CHOP. “Science advances through reproducibility. And when we all use different terms, we’re not able to reproduce research.”

Dr. Forrest compared the initiative to the Rosetta Stone, an ancient tablet with passages of identical meaning written in three languages that helped to decode the riddle of hieroglyphics.

“The Rosetta Stone was a big advance that allowed us to translate different languages,” Dr. Forrest said. “That’s what the terminology project does. It’s a common language that eventually will act as a Rosetta Stone for different research projects. It will allow communication across research studies.”

Traditionally, when a research team sets up a study, they have many questions in mind to tackle. But perhaps another study team across the country with the same study concept phrases their questions in an entirely different way. This could convolute attempts to pool the results.

Michael A. Padula, MD, MBI, medical director of informatics for the Division of Neonatology at CHOP, gave an example from three studies that the National Institutes of Health had funded about the off-label use of nitric oxide in premature infants. The New England Journal of Medicine published the three studies’ results: Two found it not to be beneficial, but the other found it was beneficial. The NIH convened a panel to interpret the results, but they were unable to make any conclusions because each study used slightly different definitions of bronchopulmonary dysplasia, which was the main outcome that the studies measured.

“They spent a great deal of money to fund these studies, but in the end they weren’t much farther along,” Dr. Padula said. “It was a missed opportunity. Had they been able to retrieve a granular level of data to allow them to harmonize the different definitions, then they would have been able to make a more definitive conclusion about the therapy.”

It can be a challenge for the research community to reach consensus on an accurate definition of a complex disease such as bronchopulmonary dysplasia or even on a common condition such as asthma. Some researchers may be reluctant to embrace a set vocabulary because they prefer certain customized terms. L. Charles Bailey, MD, PhD, an assistant professor in CHOP’s divisions of oncology and hematology, used the analogy of how people describe different shades of colors — red vs. rose, for instance.

“The research community must make a conceptual leap to say that we’re all going to agree to call this red in certain contexts, even though we can tell the shades apart, because it is more important for us to collaborate than it is for me to say exactly that my color is rose,” Dr. Bailey said.

That is where subject matter experts assembled through the Pediatric Terminology Harmonization Initiative jump in to develop semantic interoperability. They systematically go piece by piece through each field of medical specialization to align like concepts, define them appropriately so that they relate to both research and clinical care, and then tether them to reference terminology sets that researchers can access easily.

Although the scope of the terminology project is substantial and time-consuming, CHOP’s experts explained why it is important to pursue it now. First, it is increasingly becoming important to perform collaborative pediatric research to answer questions from broader perspectives.

For example, PEDSnet is a national initiative of children’s hospitals and disease specific networks that are working together to reduce the barriers to rapid and better clinical research. PEDSnet will tap into the pediatric research terminology effort as it begins to create a common data model, said Dr. Forrest, who is principal investigator of PEDSnet.

A common pediatric research terminology also will be an essential part of the National Children’s Health Study, led by Steven Hirschfeld, MD, PhD, who co-authored the Pediatrics article. The National Children’s Health study will be collecting huge amounts of data that must be aligned to allow translation of the evidence into clinical practice.

Second, the era of health care information exchange is creating a sea change because the care that clinicians provide is now linked to electronic health records (EHRs) that operate using specific languages and terminologies. However, most EHR systems currently do not describe pediatric conditions thoroughly or support pediatric research. Organizers of the pediatric terminology project want to get their foot into the door as EHRs are constructed so that they all map to the same standardized concepts when referring to pediatric patients.

“From the research side, this is a fabulous opportunity because all kinds of information are being collected,” Dr. Bailey said.

Yet, one of the biggest barriers to widespread adoption of a common pediatric research terminology is raising awareness, Dr. Bailey said. The NICHD website already has harmonized several terminology reference subsets for pediatric specialties that researchers can consult and begin to deliberately design studies to be consistent and interoperable.

“Start thinking about how the research you’re doing connects to the questions that your colleagues are trying to answer and how everyone could benefit from shared terminology,” Dr. Bailey said.