The study, titled Acetaminophen Use During Pregnancy and Autism on TikTok: Disproportionate Exposure to Misinformation in a Content Analysis, was conducted by researchers including Shivani Srivastava, a medical student at Rutgers School of Medicine, with study author Katherine Poole scheduled to present the work at the meeting. The research team set out to understand exactly what information pregnant patients and their families were being exposed to online after seeing questions about Tylenol safety arise in their own clinics. Their results paint a striking picture of how algorithmic amplification can distort the public information environment around a sensitive reproductive health topic, even when a substantial share of the available content is scientifically sound.

TikTok has become a primary source of health information for adolescents and young adults, a population that is particularly vulnerable when algorithmic recommendation systems accelerate the spread of medical misinformation. Because the platform serves large numbers of people of reproductive age, misinformation about prenatal exposures poses direct risks to prenatal health decision-making. In September 2025, public attention to a proposed association between prenatal acetaminophen exposure and autism spectrum disorder, despite the absence of causal evidence, triggered a rapid surge of TikTok content on the topic. The researchers seized on this moment as a natural experiment in how health controversy unfolds on a recommendation-driven platform.

Methodologically, the team conducted a cross-sectional content analysis of 53 TikTok videos posted between September and November 2025, immediately after the FDA announcement. The videos were identified using the search terms tylenol autism, acetaminophen autism, and paracetamol autism, along with the hashtag tylenol, and the study received institutional review board approval under protocol Pro2026000574. Two reviewers independently coded each video for misinformation status, creator type, educational quality as measured by the validated Global Quality Scale, thematic framing, and engagement metrics. Inter-rater reliability for the misinformation classification was substantial, with a Cohen’s kappa of 0.86, indicating strong agreement between the independent coders.

The statistical analysis was rigorous for a study of this scale. Group differences were assessed using Fisher’s exact test and Kruskal-Wallis tests, and independent predictors of misinformation were identified using Firth penalized logistic regression, a method well suited to small samples and rare outcomes. All analyses were performed in R version 4.5.3. The headline numbers were sobering: of the 53 videos, 52.8 percent contained evidence-consistent information, 28.3 percent contained outright misinformation, and 18.9 percent were classified as mixed or uncertain. In other words, the majority of content was accurate, yet the misinformation punched far above its weight in reach.

That asymmetry between volume and visibility is the study’s most consequential finding. Despite comprising only 28.3 percent of the videos analyzed, misinformation accounted for 52.4 percent of total views across the 28.13 million total views recorded for all content examined. A single misinformation video reached 10.3 million views, representing 36.6 percent of all exposure across the analyzed videos. As Srivastava explained, the misinformation videos, despite being a smaller portion of the overall available subject content, accounted for over half of the total views of all the content the team examined. The results, she said, indicate that TikTok’s algorithm amplified low-quality, conspiracy-framed videos containing misinformation on a very sensitive health topic and allowed them to spread far wider than evidence-based videos explaining the lack of causal evidence between prenatal acetaminophen exposure and autism.

Who was creating the misinformation mattered as much as how widely it traveled. Health professionals in the sample produced exclusively evidence-consistent content, while influencers accounted for 73.3 percent of all misinformation videos, a difference the authors report as statistically significant. Nearly three-quarters of the videos containing misinformation were therefore created by influencers rather than medical professionals. The quality gap was equally pronounced: misinformation videos scored significantly lower on educational quality than evidence-consistent videos, with a mean Global Quality Scale score of 1.73 compared with 4.18. Misinformation videos were also far more likely to employ conspiracy framing, present in 46.7 percent of them, and pharmaceutical distrust themes, present in 53.3 percent.

The regression analysis added an important nuance to the creator story. When the researchers adjusted for other factors, lower educational quality and conspiracy framing each remained independently associated with misinformation, with p values of less than 0.001 and 0.012 respectively, but creator type was no longer independently significant after adjustment. This suggests that the defining features of misinformation on the platform are not simply who makes the content but what the content looks like: videos that are poorly sourced, sensational, and framed around conspiracy narratives are the ones that both constitute misinformation and attract algorithmic amplification. The study’s figures visualize this pattern, with one bubble plot mapping each video’s Global Quality Scale score against its view count on a logarithmic scale, and a second plotting the odds ratios and 95 percent confidence intervals from the Firth penalized logistic regression model.

The clinical implications are direct. As more people turn to social media, and especially TikTok, for medical advice, the researchers argue that it becomes increasingly important for pediatricians to proactively address misinformation with patients and their families, including patients of reproductive age who may be weighing acetaminophen use during pregnancy. Srivastava noted that the project grew out of frontline experience: after seeing pregnant patients in their own clinics question whether it was safe to take Tylenol during pregnancy, the team wanted to understand what information those patients were being exposed to online. The findings, she said, reveal how important it is for medical institutions to be part of the conversations on social media by sharing accurate, engaging content that reaches patients where they are, rather than ceding that space entirely to untrained creators.

Beyond the exam room, the authors conclude that broader solutions will require platform-level content moderation and coordinated public health responses to counter misinformation on platforms widely used by adolescents and young adults. The acetaminophen episode illustrates a structural problem that no single clinician can solve: an evidence base showing no causal link between prenatal acetaminophen exposure and autism coexisted with a recommendation algorithm that rewarded the most conspiratorial and lowest-quality content on the subject with the largest audiences. Until the incentives of algorithmic distribution change, the study suggests, accurate health information will continue to compete at a disadvantage against content engineered for outrage, and medical institutions will need to meet patients on the same platforms with content that is both scientifically sound and genuinely engaging.

Subject of Research: TikTok misinformation about prenatal acetaminophen exposure and autism following the FDA's 2025 label change announcement

Article Title: TikTok misinformation content on acetaminophen and autism surged after presidential announcement

Article References: TikTok misinformation content on acetaminophen and autism surged after presidential announcement. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: TikTok, acetaminophen, autism, misinformation, pregnancy, FDA, social media, pediatrics, public health, algorithmic amplification, influencers, content analysis