239 Health claim verifications avg. score 4.4/10 69 rated true or mostly true 169 rated false or misleading
“In controlled tests, fewer than half of experienced radiologists were able to reliably detect AI-generated deepfake X-ray images.”
The claim conflates two different study conditions. When radiologists were not told deepfakes were present, only 41% spontaneously flagged something unusual — but this measures unprompted suspicion, not detection accuracy. When explicitly told synthetic images were included (the standard controlled detection task), radiologists achieved 75% mean accuracy, well above the "fewer than half" threshold. The claim cherry-picks the lower figure and mischaracterizes it as a controlled detection result.
“Corticosteroid injections are more effective than physiotherapy or rehabilitation in treating chronic tendon injuries.”
This claim is not supported by the evidence. Multiple systematic reviews and meta-analyses consistently show that corticosteroid injections provide only short-term pain relief (weeks) for tendon injuries, while physiotherapy produces equal or superior outcomes at 3–12+ months. For chronic tendon injuries specifically, a PMC-NIH review found "no good evidence" supporting corticosteroid use, and a 2025 PubMed meta-analysis confirmed injections are not superior to physical therapy beyond the short term. Clinical guidelines treat injections as adjuncts to rehabilitation, not replacements.
“Corticosteroid injections provide effective long-term relief for musculoskeletal injuries such as tendinopathy and rotator cuff tears.”
This claim is not supported by current medical evidence. Multiple recent systematic reviews and meta-analyses (2022–2025) consistently show that corticosteroid injections provide short-term pain relief — typically lasting weeks to a few months — but do not deliver effective long-term relief for tendinopathy or rotator cuff injuries. At intermediate and long-term follow-up, corticosteroids perform no better than placebo or physical therapy, and may worsen structural integrity in some cases. The only supporting evidence is a 2005 meta-analysis now superseded by stronger, more recent research.
“Social media use is as addictive as controlled substances such as cocaine or heroin, producing comparable neurological and behavioral dependency.”
Social media and controlled substances like cocaine or heroin share some overlapping dopaminergic pathways and reward-circuit activation, but the claim that they produce "comparable" addiction overstates the evidence. Peer-reviewed research consistently describes "similarities" and "overlap" — not equivalence. Cocaine and heroin directly hijack neurotransmitter systems through pharmacological mechanisms fundamentally different from social media's behavioral reinforcement. The American Academy of Pediatrics explicitly calls this comparison "not accurate," and the scientific consensus classifies social media overuse as a behavioral addiction, categorically distinct from substance dependence.
“Hormonal contraception, including birth control pills and hormonal IUDs, increases the risk of idiopathic intracranial hypertension.”
The best available evidence does not support this claim. A comprehensive meta-analysis published in Neurology in March 2026, along with earlier population-based case-control and large cohort studies, found no significant association between hormonal contraception — including birth control pills and hormonal IUDs — and idiopathic intracranial hypertension. Clinical guidance from neuro-ophthalmology specialists explicitly states there is no convincing causal evidence. The signals cited in support come from weaker study designs or apply only to specific products, not the class as a whole.
“Anktiva (nogapendekin alfa inbakicept) is approved and clinically effective for treating, curing, or preventing all types of cancer, not solely bladder cancer.”
Anktiva (nogapendekin alfa inbakicept) is approved only for BCG-unresponsive non-muscle-invasive bladder cancer with carcinoma in situ, as confirmed by the FDA, EMA, and all major clinical references. It is not approved for any other cancer type. While early-phase trials have explored its use in other malignancies, no regulatory body has recognized it as effective for treating, curing, or preventing "all types of cancer." The claim dramatically overstates both the drug's approved scope and its demonstrated efficacy.
“Consuming carbohydrates in the evening has a calming effect on the body.”
The claim captures a partial biochemical truth — carbohydrates can promote tryptophan uptake and serotonin production — but the unqualified statement that evening carbs "have a calming effect" is misleading. Peer-reviewed evidence shows outcomes depend critically on carbohydrate type, quality, and quantity. High-glycemic or large carbohydrate meals before bed are associated with sleep fragmentation, melatonin suppression, and blood sugar disruption. Only high-quality, low-glycemic carbohydrates in moderate amounts show associations with improved sleep and reduced anxiety, and even then, the evidence is mixed on evening-specific timing.
“Dietary supplements containing undisclosed pharmaceutical drugs — including steroids, thyroid hormones, and amphetamine-like stimulants — have been sold to consumers for years with limited regulatory consequence due to insufficient FDA enforcement capacity.”
The claim is substantially accurate. Peer-reviewed research documents over 1,000 dietary supplements adulterated with undisclosed pharmaceuticals — including synthetic steroids and stimulants — sold from 2007 through 2021, with some products remaining on shelves years after FDA warnings. However, the specific inclusion of "thyroid hormones" as a central adulterant pattern is not well-supported by the evidence. Additionally, the enforcement gap stems primarily from DSHEA's statutory design (classifying supplements as foods), not purely from insufficient FDA capacity — a meaningful distinction the claim obscures.
“Collagen supplements in the United States are largely unregulated by the Food and Drug Administration due to the Dietary Supplement Health and Education Act of 1994.”
The claim is substantially accurate. DSHEA (1994) does exempt collagen supplements from FDA premarket approval and shifts the burden of proving unsafety to the FDA, which multiple peer-reviewed and medical sources confirm. However, "largely unregulated" overstates the situation: the FDA retains meaningful post-market authority including cGMP manufacturing standards, labeling enforcement, adulteration removal powers, and premarket safety review for new dietary ingredients. A more precise framing would be "largely exempt from premarket approval requirements" rather than "largely unregulated."
“Most studies reporting benefits of collagen supplements are funded by the supplement industry or by researchers with financial ties to the supplement industry.”
Industry funding is widespread in collagen supplement research, and a major 2025 meta-analysis found that positive results were concentrated in industry-funded, lower-quality trials while independent, higher-quality studies showed no significant benefit. Harvard and peer-reviewed reviews flag conflicts of interest as a pervasive concern. However, no source in the evidence base actually counts the proportion of benefit-reporting studies that are industry-funded, so the specific claim that "most" such studies meet this threshold is plausible but not directly demonstrated.
“The scientific evidence supporting the benefits of collagen supplements for non-cosmetic body systems, such as bones, joints, and digestion, is weaker or less established than the evidence for cosmetic benefits.”
This claim oversimplifies a complex evidence landscape. While digestive benefits of collagen supplements do rest on thin, mixed evidence, joint and osteoarthritis outcomes are supported by multiple reviews and meta-analyses — making them comparably or even better established than cosmetic claims. Critically, recent high-quality analyses show that positive cosmetic results are largely driven by industry-funded, lower-quality studies, with independently funded trials finding no significant skin benefits. Grouping all non-cosmetic domains as uniformly "weaker" misrepresents the actual state of the science.
“COVID-19 vaccines cause sudden death in young, healthy people.”
The claim that COVID-19 vaccines cause sudden death in young, healthy people is not supported by the evidence. Multiple large-scale population studies — including a CDC analysis, a 2026 PLOS Medicine case-control study, and surveillance data covering tens of millions of people — consistently find no increased risk of sudden death among vaccinated young individuals. While vaccine-induced fatal myocarditis has been documented in extraordinarily rare cases (28 deaths identified globally against billions of doses), this does not support the sweeping causal claim as stated.
“The SARS-CoV-2 BA.3.2 variant has significant immune escape potential and has been confirmed in 23 countries.”
CDC and WHO data confirm BA.3.2 was detected in at least 23 countries and demonstrates enhanced antibody escape in laboratory testing — both factual pillars of the claim hold up. However, describing the immune escape as "significant" without qualification overstates the real-world picture: WHO rates BA.3.2 as low additional public health risk, vaccines are still expected to protect against severe disease, and the variant shows reduced infectivity with no consistent growth advantage. The core facts are accurate, but the framing omits important context.
“ImmunityBio's drug N-803 (anktiva) is being investigated or has demonstrated efficacy in treating, curing, or preventing cancer types beyond bladder cancer.”
ImmunityBio's N-803 (Anktiva) is actively being investigated in multiple cancer types beyond bladder cancer, including pancreatic cancer, non-small cell lung cancer, glioblastoma, and other advanced solid tumors, as confirmed by ClinicalTrials.gov registrations and NCI-sponsored trials. Preliminary efficacy signals in NSCLC have been reported, though definitive Phase 3 proof of efficacy beyond bladder cancer has not yet been established. The claim's "being investigated" component is firmly supported by high-authority sources.
“Fathers are significantly more likely to be diagnosed with depression and stress-related disorders one year or more after the birth of a child than during the pregnancy period.”
This claim is grounded in a real finding from a large Swedish registry study showing a spike in fathers' clinical diagnoses at 12+ months postpartum. However, it overgeneralizes that single-country result into a broad rule. Multiple meta-analyses and systematic reviews place peak paternal depression at 3–6 months postpartum, not at one year or later. The Swedish study also compared the spike to pre-pregnancy baselines — not directly to the pregnancy period as the claim states — creating a key evidentiary gap.
“Taking caffeine before a period of sleep deprivation can fully restore social memory function that would otherwise be impaired.”
A 2026 peer-reviewed study did show caffeine reversed social memory deficits in male mice via a specific hippocampal CA2 mechanism. However, the claim's unqualified language — "fully restore social memory function" — overgeneralizes from a single animal model and one narrow social-recognition assay. No human evidence confirms this effect. Broader research shows caffeine often only partially rescues cognition under sleep deprivation and can disrupt recovery sleep. The core finding is real but the claim's framing is misleading.
“The diabetes drug metformin can suppress HIV replication, keep the virus dormant, and enable long-term remission without the need for daily antiretroviral therapy.”
This claim dramatically overstates the evidence. While metformin shows some ability to modulate HIV biology in laboratory and animal studies, no clinical evidence supports the assertion that it can enable long-term remission without daily antiretroviral therapy. Multiple peer-reviewed studies actually show metformin can increase HIV transcription and reactivate latent virus. All human studies tested metformin only as an add-on to ART, not as a replacement. The claim conflates early-stage, preclinical findings with established clinical capability.
“Holding in a sneeze can have negative health effects.”
The claim is well-supported. Multiple credible medical sources, including the Cleveland Clinic and ENT specialists, confirm that suppressing a sneeze can redirect pressure internally, potentially damaging eardrums, sinuses, throat tissue, or blood vessels. The claim uses "can have," which is a possibility statement — and documented case reports plus established physiological mechanisms are sufficient to validate it. While severe outcomes are rare, the possibility of negative health effects is real and medically recognized.
“Med beds are medically validated devices that can cure serious diseases using energy or frequency-based healing methods.”
No device called a "med bed" has been medically validated or shown to cure serious diseases in any clinical trial. The concept originates from conspiracy theories, not medical science. While some energy-based therapies (e.g., PEMF, sound stimulation) show limited benefits for specific symptoms, none constitute cures for serious diseases, and none involve "med beds." Major medical authorities, including the Cleveland Clinic and Cancer Research UK, confirm energy healing is unproven as a curative treatment. The FDA has issued warnings against unapproved medical claims for such devices.
“Waking a person who is sleepwalking can cause a heart attack or serious physical harm.”
This claim is a widely circulated myth. Major medical authorities — including the Mayo Clinic, Cleveland Clinic, Northwestern Medicine, and the American Academy of Sleep Medicine — explicitly state that waking a sleepwalker does not cause heart attacks, brain damage, or other serious medical harm. The only documented risk is temporary confusion or disorientation, which in rare cases may lead to minor accidental injury. The heart attack component is categorically unsupported by clinical evidence.