Patients skip clinician follow-ups
New data show 42% of patients who consult AI for health advice don't follow up with a human clinician—raising safety and continuity questions just as U.S. healthcare organizations report implementation and hiring lagging behind AI strategies. The split between consumer use and clinical integration suggests risks from unsupervised AI consultations and uneven institutional readiness. (techtarget.com, htworld.co.uk)
KFF’s March 25 poll was administered Feb. 24–Mar. 2 among 1,343 U.S. adults and reports a margin of sampling error of ±3 percentage points for the full sample. (beckershospitalreview.com) The poll finds 32% of adults say they’ve used AI chatbots for health information in the past year, with 29% citing physical-health use and 16% citing mental-health use. (kff.org) Asked why they turn to AI, 65% of users named speed/immediacy as a major reason, 41% said to look up information before seeing a provider, and 36% said privacy or comfort asking questions was a major factor. (kff.org) Younger adults and uninsured people are disproportionately represented among AI health users: 28% of those ages 18–29 used AI for mental-health information, uninsured users report higher mental-health use (30% v. 14% for insured), and Black and Hispanic adults report higher mental-health AI use than White adults. (kff.org) A large majority of the public—77%—expresses concern about the privacy of medical information given to AI tools, while about one in five AI users cite not having a health care provider or an appointment as a major reason they relied on AI. (kff.org) Industry surveys show implementation gaps at institutional level: a Guidehouse/HIMSS survey found 78% of health systems engaged in AI projects but only 52% felt operationally ready to implement them. (guidehouse.com) Research on hospital-level deployment shows adoption is clustered and unequal: an analysis of 3,560 U.S. hospitals found “hotspots and coldspots” of predictive-AI implementation and lower adoption in regions with greater healthcare access needs. (nature.com) Separate polling and reports document real-world risk signals, including a consumer survey in which roughly 22% of respondents said they had followed AI medical advice later proven wrong, underscoring the potential for safety and misinformation when users do not seek clinician follow-up. (newsweek.com)
