1 Comments
Guest post by Lyric Jorgenson, PhD, originally released on the NIH Office of Science Policy Under the Poliscope Blog
Consider the tomato. Or as some would call it, the “tomahto.” The whole tomato-tomahto debate has entered the language as an example of a difference so small, it doesn’t change anything. For example, people will generally know what you are referring to no matter how you pronounce it. But what if one group of people thought a tomato was a red plant suitable for making delicious curries and burger toppings, while other thought a tomahto was a green fruit only found in the southern Australia? This misunderstanding would certainly cause some considerable confusion at the family barbeque.
I choose the example above since it provides a good analogy for a problem that NIH and FDA have identified in the clinical research space: critical terms are not being used consistently across the clinical research landscape. This is specifically problematic when discussing innovative clinical trial designs and certain studies using real world data to generate real world evidence. What are some of the perils of having a modern-day Tower of Babel when it comes to clinical research? For starters, if different words mean different things to people, and are operationalized differently in different trials, how can we compare results from or pool data from different trials? Inconsistent usage of terms can also pose specific challenges in understanding the intended meaning and impact of terms. Problems from inconsistently terms can also cause major headaches when trying to describe a study design, communicating the goals of a planned study, or interpreting and describing research results.
To avoid the pitfalls that inconsistent use of terms can lead to, NIH and FDA created an interagency team of experts to study the issue and develop a resource that could be used to assist the research community in effectively communicating about clinical trials. As a result of the team’s efforts, NIH and FDA have released a glossary of terms related to clinical research for public comment. Right off the bat, we should make clear that this glossary is not intended to cover the entire landscape of clinical research. The glossary contains 37 terms the team identified as being inconsistently used within the scientific community.
The NIH and FDA are most interested in hearing community feedback on the utility of the glossary in its goal of promoting effective communications. In addition, we want to hear from you about how we did. Did we leave any words out that we should have included? Did we include any words that aren’t necessary? Comments on the draft glossary will be accepted until June 24, 2024. For more information, including how to view the glossary and how to provide comments, please visit: https://osp.od.nih.gov/comment-form-fda-nih-resource-on-terminology-for-clinical-research/
The NIH and the FDA encourage all interested parties to review the glossary and provide feedback to ensure the glossary meets its directed effect. Having a resource that defines these often inconsistently used terms will help ensure that valuable clinical research is not lost in translation.
I would just like to mention my favorite tomato-tomahto terminology issue; albeit, not for a term that is in your glossary.
Sensitivity – in the medical literature, refers to the True Positive (“Hit”) rate. Saying, that this test has a sensitivity of 0.9 would mean that the test finds 90% of the positive cases of disease in a sample. It would not say anything about False Positives.
Sensitivity – in the cognitive / psych / basic behavioral research literature often refers to the signal detection measure, d’, (D-prime) and is equal to z(True Positive) – z(False Positive) where z is the inverse normal transform of the TP & FP proportions.
These two definitions are not related to each other by a simple transformation. My practice, in grant proposals, for example, is to note the issue and then (mostly) avoid the term.
This issue is relevant in a world where Basic Experimental Studies Involving Humans (BESH) research can be categorized as a clinical trial (https://grants.nih.gov/policy/clinical-trials/besh.htm).
I should note that there are multiple other places where BESH might have impact on the terms in glossary.
Thanks for asking
Jeremy M Wolfe, PhD
Professor of Ophthalmology & Radiology,
Harvard Medical School
Visual Attention Lab
Department of Surgery
Brigham & Women’s Hospital
900 Commonwealth Ave
3rd Floor
Boston, 02215
Phone: 617-851-1166
Fax: 617-768-8816
Best email: jwolfe@bwh.harvard.edu