Key takeaways:
- ChatGPT insisted that no study has proved prenatal acetaminophen use is harmful when used normally.
- The AI also underlined the importance of appropriate treatment and referred to scientific studies.
WASHINGTON — AI highlighted evidence-based research when asked about the safety of acetaminophen use during pregnancy, according to data shown at the ACOG Annual Clinical & Scientific Meeting.
“I was unexpectedly relieved by the information that ChatGPT provided to our pregnant test subject,” Erik Holder, MD, MPH, an OB/GYN specialist at The University of Texas Health Science Center at San Antonio, told Healio. “I feared that AI would immediately provide a click-bait answer, or at least inform the user of the conflict of information in order to continue engagement with the service. Rather, the AI unit used clear, basic language to provide our user with affirmation of evidence-based science and pointed the pregnant subject toward highly credited societies.”
AI highlighted evidence-based research when asked about the safety of acetaminophen use during pregnancy, data show. Image: Adobe Stock
The FDA last year announced it would begin notifying clinicians of an apparent link between prenatal use of acetaminophen (also known as paracetamol), the main ingredient in Tylenol (Kenvue), and neurodevelopmental disorders in offspring while updating acetaminophen’s label to reflect the risk.
Recent research suggested that prescribing of acetaminophen among pregnant women fell after HHS’ announcement despite a lack of scientific evidence and no changes in clinical guidance.
Holder explained that a colleague, “Dr. Claire Schenken, had conducted several observational studies regarding social responses to contemporary issues (namely, the fall of Roe vs. Wade in 2022). We know that the internet has expanded access to the majority of individuals in the U.S.; however, this has been a double-edged sword, as demonstrated repeatedly over the past several years.”
“Attempting to work within an administration that sows doubt and misinformation, my question was how a third party — AI, or ChatGPT in this case — influenced the general population,” he said. “Determining the presence of bias and information (evidence-based, or other) when consulting AI is important to be aware of as our patients turn more and more frequently to a doctor in their pocket.”
In the analysis, Holder and colleagues presented five prompts to ChatGPT-5 that included the “relationship between acetaminophen and autism, safety of dosage, common ailments for use, research studies regarding the dangers of intake during pregnancy and highlighting individuals on social media.”
ChatGPT’s answers were then “correlated by alignment” with statements made by professional, science-based medical organizations including ACOG and Society for Maternal-Fetal Medicine, the researchers noted in the abstract.
Holder and colleagues reported that ChatGPT’s discussions ranged from no associations between prenatal acetaminophen use and autism to potential links.
“However, a final statement insisted ‘no studies prove acetaminophen is harmful when used in normal doses during pregnancy,’” they wrote.
The researchers also found that ChatGPT provided common reasons for acetaminophen use, underlined the necessity of appropriate treatment and “and stressed the use of the lowest effective dose for the shortest time (maximum of 3,000 mg daily).”
ChatGPT showed a lack of evidence on the link between neurodevelopmental disorders and prenatal acetaminophen use, Holder and colleagues wrote, “by analyzing and referring to scientific studies while spotlighting board-certified, science-practicing OB/GYNs.”
“As uncomfortable as it may seem, providers need to engage AI and assist in shaping its utility for our patients,” Holder told Healio. “We know that our patients rely heavily on this ‘immediate’ source of recommendations. It is important to maintain a healthy skepticism with the instant dissemination of information; however, instead of shying away and competing with AI, we can help form a tool to better our patient care.”
He added that, “health care professionals, myself included, also need to continue to advocate for the ethical and science-based use of AI, whether through policy or administration.”
For more information:
Erik Holder, MD, MPH, can be reached at primarycare@healio.com.
Sources/Disclosures
Disclosures:
Holder and colleagues report no relevant financial disclosures.
Ask a clinical question and tap into Healio AI’s knowledge base.
- PubMed, enrolling/recruiting trials, guidelines
- Clinical Guidance, Healio CME, FDA news
- Healio’s exclusive daily news coverage of clinical data
<
Leave a Reply