Research institute · Pew Research Center
Pew: 36% use social media for health info, 7% trust it
Pew's April 2026 study of 5,111 US adults maps where Americans actually go for health information and how much of it they trust. The reach-versus-trust gap is the finding UK marketers in any health-adjacent category should be optimising against.
Source published Summarised
Key findings at a glance
- 85% of Americans get health information from health care providers at least sometimes, the most common of seven sources Pew tested.
- 51% say they often or extremely often turn to doctors and other medical workers for health information.
- 66% get health information from people facing similar health issues at least sometimes, peer testimony rivals expert sources for reach.
- 60% turn to major health websites such as WebMD or Mayo Clinic at least sometimes.
- 36% get health information from social media at least sometimes, meaningful reach despite a poor trust profile.
- 22% get health information from AI chatbots at least sometimes, a new digital channel that has reached more than one in five US adults.
- 47% of Americans who use social media for health information say that information is not too or not at all accurate.
- 7% of social media health-info users rate that information as highly accurate, the lowest credibility ceiling of any source tested.
- 65% of Americans who get health information from providers rate that information as highly accurate, the trust gold standard.
- 76% say they encounter conflicting health information at least sometimes; 27% see it often or extremely often.
- 54% say it is at least somewhat difficult to know which information to trust when they come across conflicting health claims.
- 73% get health information from three or more different sources at least sometimes, multi-source consumption is the default behaviour, not the exception.
What the report covers
Pew Research Center asked 5,111 US adults how often they turn to seven different sources for health information, health care providers, major health websites, people with similar health issues, journalists and news organisations, government health agencies, social media, and AI chatbots, and how accurate they consider each one. The fieldwork ran 20-26 October 2025; results published April 2026.
Two findings dominate. First, the medical establishment still leads by a wide margin: 85% of Americans get health information from providers at least sometimes, and 65% of that group rate provider information as highly accurate. Second, the newer digital channels, social media (36% reach) and AI chatbots (22% reach), have built meaningful audiences but have not built meaningful trust. Only 7% of social media health-info users rate that information as highly accurate; AI chatbots sit in a mixed-trust middle.
The crowded environment also has a cost. Half of Americans say judging the accuracy of health information is at least somewhat difficult, and 76% say they regularly encounter conflicting health claims. The headline reach numbers therefore overstate channel quality, audience scepticism is now baked into how people consume this content.
What this means for UK marketers
The data is US-specific but the consumer behaviour transfers cleanly to UK markets. Three operational reads for UK brands and the agencies running their digital marketing programmes:
Reach is not permission. The 36% of Americans who get health information from social media are not 36% who believe it. When 47% of that audience considers the information inaccurate and only 7% considers it highly accurate, organic social and paid social health content is starting from a credibility deficit. Brand work running through Meta Ads in supplements, fitness, sleep, women's health, mental wellness, and any adjacent category, should treat the platform as a discovery layer that needs to hand the audience off to higher-trust assets (clinically-cited landing pages, qualified-expert content, recognised retail listings) before asking for a conversion. Single-platform creative that tries to close on the in-feed unit is fighting the audience's prior.
The 22% AI chatbot number is a planning input, not a novelty. More than one in five US adults are now consulting AI chatbots for health information at least sometimes, and the audience is growing fastest among under-30s. UK brands need to start treating LLM responses as a distribution surface, that means structured content, factual product pages, schema-marked FAQs, and authoritative third-party citations that LLMs can crawl and reproduce. The same investments that compound for traditional SEO are now compounding for AI answer surfaces. Compliance teams should also note that an LLM summarising your product page is not bound by the ASA Code on health claims, your source content needs to be defensible on its own.
Conflicting information is now the default consumer experience. 76% of Americans encounter conflicting health claims at least sometimes; 54% find it difficult to judge which to trust. For brands, that is both a risk and an opening. The risk is that one viral counter-claim can collapse a category narrative, vitamin D, statins, oat milk all have recent examples. The opening is that audiences are actively looking for resolution. Owned content that explicitly addresses common contradictions ("is X really good for you?") and shows working, methodology, sample size, source, is the format that performs against this audience posture, and it ages well in search.
Read the trust gap alongside the two follow-up Pew studies on health and wellness influencers, where half of US under-50s now report getting health advice from influencers but only 10% of that audience fully trusts what they hear. The same pattern repeats across every channel Pew tested: consumption is broad, trust is narrow, and the brands that win are the ones designing for the gap.
How they got the data
Pew Research Center surveyed 5,111 US adults from 20-26 October 2025 via Wave 182 of the American Trends Panel. The ATP recruits via address-based sampling using the US Postal Service's Computerized Delivery Sequence File, with online and live-telephone interviewing in English and Spanish. Oversamples are weighted back to population proportions so the result represents the views of the full US adult population.
Respondents were asked how often they get health information from each of seven sources, how accurate they consider information from each source, and how difficult they find it to judge accuracy when they encounter conflicting health information. Detailed responses, question wording and the full survey methodology are published alongside the report. The report's lead authors are Giancarlo Pasquini, Galen Stocking, Emma Kikuchi, Isabelle Pula and Eileen Yam.
Research institute
Read the full report
From Pew Research Center · published 07/04/2026