Practice Point: Universal negative pressure surgical wound therapy is not quite ready for prime time, yet.
EBM Pearl: Evidence-informed care has three pillars: clinical expertise, application of the best available research, and patient values and preferences.
"If you get something free on the internet, you might be the product."
AI applications can search the medical literature and may systematically emphasize certain kinds of evidence, such as studies with positive findings over those with negative findings.
For example: The healing of large surgical wounds can be a laborious process, and although negative pressure devices can be painful and expensive to apply, if they help wounds heal, they might be worth it. A meta-analysis of over 3,000 patients in 2023 in the International Wound Journal (IWJ) suggested that these devices help in routine healing of large surgical wounds. Unfortunately, this paper was retracted (this spring, 15 months later) due to concerns about a compromised peer-review process. Last month, a much smaller study by another group of researchers published in The Lancet demonstrated that negative pressure devices were not associated with a shortened time to healing for surgical wounds healing by secondary intention. If an AI search engine is biased toward larger studies or those that show a positive effect or failed to exclude those with retractions, the IWJ meta-analysis would lead you to recommend routine negative pressure therapy for large postoperative wounds, particularly if the AI engine doesn’t identify and exclude retracted articles.
It’s easy to produce an elegant answer with an online search engine; it’s harder to produce a trustworthy one. For example, the instructions given to large language models might prioritize newer studies and those with positive results. In the case above, some search algorithms might prioritize this larger but retracted IWJ meta-analysis over the arguably more accurate Lancet article. In addition, not all AI tools have transparent conflict-of-interest policies in place, something that is standard for any trustworthy medical journal. If you are using a free AI engine to answer your medical questions and there is no transparency about how the engine arrives at the answer, you might want to ask yourself who is benefiting from the answers you are getting.
DynaMed EBM Focus Editorial Team
This EBM Focus was written by Dan Randall, MD, MPH, FACP, Senior Deputy Editor at DynaMed. Edited by Alan Ehrlich, MD, FAAFP, Executive Editor at DynaMed and Associate Professor in Family Medicine at the University of Massachusetts Medical School; Katharine DeGeorge, MD, MS, Senior Deputy Editor at DynaMed and Associate Professor of Family Medicine at the University of Virginia; Gayle Sulik, PhD, Senior Medical Editor and Team Lead for Palliative Care at DynaMed; McKenzie Ferguson, PharmD, BCPS, Senior Clinical Writer at DynaMed; Rich Lamkin, MPH, MPAS, PA-C, Clinical Writer at DynaMed; Matthew Lavoie, BA, Senior Medical Copyeditor at DynaMed; Hannah Ekeh, MA, Senior Associate Editor II at DynaMed; and Jennifer Wallace, BA, Senior Associate Editor at DynaMed.