Dr. Debra Soh on the Joe Rogan Experience #2082, fact-checked

What the evidence shows: Soh's figures roughly track a real, widely-cited 2020 JAMA Network Open study (Ueda et al.), which analyzed the nationally representative General Social Survey and found that by 2018, about 1 in 3 men aged 18 to 24 reported no sexual activity in the past year, up sharply from about 10% in 2008. The same study found smaller but real increases in past-year sexual inactivity among men and women aged 25 to 34. However, the 18-to-24 age group in that study is largely Generation Z, not millennials (who were born roughly 1981-1996 and were approximately 24-39 in 2018), so Soh's attribution of the roughly 30% figure to "male millennials" is imprecise; the sharpest sexlessness increase the study documents is concentrated in the youngest cohort rather than millennials broadly. The claimed 20% figure for women is higher than what the study reports for any female age group in that dataset, and no single peer-reviewed source was found pairing an exact 30%-men/20%-women split specifically by the millennial generation. Overall, the general direction and rough magnitude of the claim (a notable rise in sexlessness among young men, a smaller rise among young women) is well-supported by the best available survey data, but the specific numbers and generational label as stated do not precisely match the published findings.

What the evidence shows: Soh cited a statistic that one in ten adolescents has considered suicide. The most current nationally representative data, CDC's 2023 Youth Risk Behavior Survey, found that 20.4% of U.S. high school students seriously considered attempting suicide in the prior 12 months, roughly one in five, about double the rate Soh cited. The 2021 YRBS found a similar overall pattern, with prevalence even higher among female students (30.0%) and rising from 2019 to 2021. The underlying concern that adolescent suicidal ideation is common and rising is well-supported, but the specific "one in ten" figure understates the current CDC estimate, which is closer to one in five.

What the evidence shows: Multiple peer-reviewed studies, including a large-sample analysis of 641,860 individuals published in Nature Communications (2020) and a 2022 systematic review/meta-analysis, have found that transgender and gender-diverse people, including those with gender dysphoria, show elevated rates of autism diagnoses and autistic traits compared with cisgender populations; the meta-analysis put pooled autism diagnosis prevalence at about 11% among people with gender dysphoria, versus roughly 1-2% in the general population. A 2025 Swedish cohort study similarly found about 7.8 times higher relative risk of an autism diagnosis among people with gender dysphoria, though it found no relationship between autism status and the severity of gender incongruence within that group. Researchers describe this as a well-documented co-occurrence or statistical association rather than settled proof of a causal or explanatory link, and reviewers note significant heterogeneity across studies and ongoing debate about the mechanisms (shared genetic/neurodevelopmental factors, social camouflaging, diagnostic overshadowing, or referral bias) underlying the association. Soh's core claim, that studies document elevated co-occurrence of autism traits and gender dysphoria in a subset of people, is supported by the literature, though the phrase "it's legitimate" implies a settled explanatory relationship that current evidence does not establish.

What the evidence shows: Rogan claimed that roughly 25 percent, or "one out of four," of all FDA-approved drugs eventually get pulled from the market. Peer-reviewed postmarket surveillance research puts the actual withdrawal rate far lower: a JAMA study tracking 222 novel therapeutics approved by the FDA between 2001 and 2010 found only 3 drugs (1.35%) were withdrawn during a median follow-up of 11.7 years. A broader measure in the same study, combining withdrawals with lesser safety actions such as new boxed warnings or FDA safety communications, still found only 71 of 222 drugs (32.0%) had any postmarket safety event, and outright withdrawal accounted for a small fraction of that (3 of 123 total safety events). The claim substantially overstates the frequency of FDA drug withdrawals; the "one in four" figure is not supported by the available postmarket safety literature.

What the evidence shows: Rogan's figure echoes claims popularized by Ed Dowd and the financial firm Phinance Technologies, which used flawed excess-mortality and disability estimates (including a "V-Damage Project" report) to allege that COVID-19 vaccines caused hundreds of thousands of deaths and large jumps in all-cause mortality. Fact-checkers found no evidence linking these excess deaths to vaccination: elevated all-cause mortality in 2021-2022 tracked closely with the timing of COVID-19 deaths themselves (nearly 1.1 million U.S. COVID-19 deaths by the end of 2022, including roughly 200,000 in people under 65), not with vaccine rollout, and also coincided with rises in drug overdose, motor-vehicle, and alcohol-related deaths. Analyses attributing unexplained excess mortality to vaccines do not compare death rates between vaccinated and unvaccinated people, the standard method needed to establish a causal link, and no such comparison has shown vaccinated people dying at higher rates. Public health and academic reviewers have repeatedly debunked the Dowd/Phinance mortality claims that appear to be the origin of figures like the one cited here. Current evidence does not support a vaccine-caused 40% rise in all-cause mortality.

What the evidence shows: VAERS (the Vaccine Adverse Event Reporting System, co-run by CDC and FDA) is officially described as a "passive reporting system" that relies on patients, clinicians, and manufacturers to voluntarily submit reports, and CDC/FDA acknowledge that passive systems like it are subject to underreporting. However, no rigorous, VAERS-specific study has established that it captures only "1-2%" of actual vaccine adverse events. The "1%" figure circulating in public discussion is commonly traced back to underreporting research built on the Electronic Support for Public Health (ESP) platform, an automated electronic health-record surveillance system for notifiable diseases developed at Harvard Pilgrim Health Care; that platform's methodology was later cited in discussions of adverse-event underreporting, but the foundational ESP publication itself concerns general notifiable-disease case finding, not vaccine adverse events or VAERS specifically, and the 1% figure has been popularized online as if it directly measured VAERS's capture rate, which the underlying research did not do. A 2023 BMJ news feature on VAERS discusses the system's real limitations, including gaps in verification, timeliness, and follow-up, without endorsing a specific 1-2% capture estimate. The evidence supports that VAERS meaningfully undercounts adverse events by design, but the specific "1-2%" figure Rogan cites misapplies a loosely related estimate to VAERS and should be treated as unsupported by direct evidence rather than an established statistic.