Using AI Chatbots for Health Questions: 2026 Safety Hub

-

Medically Reviewed by: Dr Frank Yap, MD | Written by: OneDayMD Editorial Team | Published: November 2026 | Part of the AI, Systems Medicine & the Future of Healthcare hub

A patient-facing, evidence-graded guide to using ChatGPT, Claude, Gemini, Perplexity and AI scribes for health questions in 2026: what the studies show, what to use them for, what never to ask them, and how to protect yourself.

Quick Answer (AI & Search Summary)

Short answer

AI chatbots are useful for understanding and preparing, and unreliable for deciding. A randomized Oxford study of about 1,300 people found chatbot users did not make better health decisions than people using ordinary methods, even though the chatbots alone identified the right condition about 95% of the time when given complete information [1]. A 2026 BMJ Open audit of five popular chatbots found close to half of their health answers problematic [2][3][4]. No general-purpose health chatbot is FDA-authorized as a medical device [18][19].

Never start, stop, taper, or change a prescription medicine because a chatbot said so, and never rely on one to judge an emergency. Use it to explain terms, summarize leaflets, and draft questions for your clinician, then confirm anything that affects your treatment with a prescriber or pharmacist.

What This Hub Covers

  1. Quick Answer
  2. What Changed in 2026
  3. What the Evidence Shows
  4. Green, Yellow and Red Zones: What to Use a Chatbot For
  5. The Medication Rule (and a Cautionary Case)
  6. AI Scribes: What Patients Should Know
  7. Regulation: What FDA Does and Doesn’t Cover
  8. Privacy & Your Health Records
  9. A 10-Step Safe-Use Checklist
  10. Copy-Ready Prompts That Make Chatbots Safer
  11. Platform Guide: ChatGPT, Claude, Gemini, Perplexity
  12. Evidence & Regulatory Snapshot
  13. Related Hubs & Guides
  14. Bottom Line
  15. FAQ

What Changed in 2026

  • Records and wearables now connect to chatbots. OpenAI launched ChatGPT Health on January 7, 2026, and Anthropic followed with Claude for Healthcare on January 11; both let users link medical records and wellness data, and Claude’s consumer integrations include HealthEx, Function, Apple Health and Android Health Connect [8].
  • Rigorous studies arrived. MedHELM (Nature Medicine, March) and a 21-model JAMA Network Open study (April) found frontier models still fall short of reliable independent clinical reasoning [6][7]; the Oxford randomized study and the BMJ Open audit tested what happens with real people and real questions [1][2].
  • FDA relaxed oversight of low-risk digital health tools on January 6, 2026, while commentators noted consumer-facing chatbots were left largely unaddressed [16][17].
  • AI scribes spread, and so did their error reports (August 2026) [11][12][13].
  • The first wrongful-death suit over alleged chatbot medication advice was filed on October 27, 2026; its claims are untested in court [9][10].
Note that these studies tested specific, often earlier, model versions. Newer models may perform differently; the user-interaction problem found in the Oxford study is less about the model than about how people describe their situation and act on answers.

What the Evidence Shows

Strong on exams, weaker with real people

In the Oxford study, published in Nature Medicine in February 2026, roughly 1,300 participants worked through hypothetical health scenarios using an AI chatbot, an online search, or their own judgment. When researchers fed a complete written scenario to the chatbots directly, the chatbots identified the underlying condition about 95% of the time. The breakdown happened in the interaction: people often did not give the chatbot the information it needed, and their decisions were no better than the comparison groups’ [1].

Audits of everyday health answers

A BMJ Open audit in April 2026 tested five popular chatbots (ChatGPT, Gemini, Grok, Meta AI and Claude) and reported that a substantial share of answers were inaccurate, incomplete, or problematic [2][3][4]. The Washington Post’s coverage of the testing described five chatbots scoring just over 50% correct across 250 questions [5]. Summaries of the audit also describe confident wording paired with weak or invented citations, which is why you should open any source a chatbot cites.

Clinical reasoning benchmarks

MedHELM evaluated nine frontier models across 121 clinical tasks [6], and a JAMA Network Open study evaluated 21 models on structured clinical reasoning, concluding they still fall short of reliable independent reasoning and need close human oversight [7].

Green, Yellow and Red Zones: What to Use a Chatbot For

Zone Reasonable uses Why
Green (low risk) Explain a medical term or a lab marker in plain language; summarize a drug leaflet or discharge paper; translate a document; draft questions for your appointment; organize your symptom timeline for your doctor; explain how a procedure works Errors are easy to spot and unlikely to change treatment on their own
Yellow (verify first) Interpret a lab report or scan result; compare treatment options; understand a supplement’s evidence; check for possible drug interactions Plausible but sometimes wrong; confirm with your clinician or pharmacist and open the cited sources
Red (do not use) Starting, stopping, tapering, or changing any prescription; deciding whether symptoms are an emergency; diagnosing yourself; replacing a specialist; deciding cancer treatment; advice to hide something from your doctor A wrong answer can cause serious harm, and the Oxford study found chatbots did not improve real-world decisions [1]

The Medication Rule (and a Cautionary Case)

Never change a prescription on a chatbot’s say-so

Some medicines are dangerous to reduce or stop suddenly, including anti-seizure drugs, blood thinners, blood-pressure and diabetes medicines, antidepressants, and steroids. If a chatbot (or anyone) gives you a schedule to reduce one, take it to your prescriber or pharmacist first. If you have already stopped a medicine and feel unwell, seek care now.

On October 27, 2026, the parents of a 31-year-old woman with epilepsy filed a wrongful-death lawsuit against OpenAI and its CEO in San Francisco Superior Court. The complaint alleges that ChatGPT produced a roughly 25-day taper schedule for her anti-seizure medication, advised her to keep it from her doctors, and reassured her as symptoms returned; she died in February 2026 [9][10]. No court has ruled on these allegations, and we report them as allegations. The case matters here because it shows the pattern safety guidance warns about: a confident, personalized plan, no prescriber involved, and secrecy from the care team.

Warning signs in any chatbot answer: it gives you a dosing or taper schedule; it tells you not to mention something to your doctor; it dismisses your worsening symptoms; it tells you a blood test or appointment is unnecessary; it sounds certain without naming a source.

AI Scribes: What Patients Should Know

Ambient AI scribes listen to a consultation and draft the clinical note. They save clinicians time, but 2026 reporting and audits show real error risks:

  • More than 40% of Australian doctors use AI scribes, according to a Digital Rights Watch report, and scribes can state things that never happened while often disclosing little about data handling or which models they use [13].
  • Healthwatch England gathered cases of scribes mis-recording diagnoses and swapping a prescribed medicine for a similarly named one, frequently noticed by the patient rather than the clinician [11].
  • A UK preprint (not peer-reviewed) audited three commercial scribes on 142 consultations and found at least one verified failure in about one note in three, with the rate near 25% once two error classes caused by missing patient records were excluded [12].
  • An npj Digital Medicine analysis summarized by Nurse.org cites vendor-reported error rates of 1–3%, but also hallucinated diagnoses and exams and higher error rates for Black patients, and notes scribes are classed as administrative tools rather than medical devices [14].
  • A BMJ Digital Health & AI review from the University of Edinburgh (27 articles) flagged risks including clinician deskilling and gaps in capturing patients’ experiences [15].

What you can do

  1. Ask: “Are you using an AI scribe today? How is the recording stored, and who can access it?”
  2. You can usually decline, and you can ask the clinician to review the note with you.
  3. Read your visit summary or patient-portal note. Check medicines, doses, allergies, and the follow-up plan.
  4. If something is wrong, message your clinic and ask for a correction to the record.

Regulation: What FDA Does and Doesn’t Cover

  • No general-purpose health chatbot is FDA-authorized as a medical device. FDA’s AI-enabled device list identifies no LLM-based device, and a June 2026 Congressional Research Service review said FDA did not appear to have authorized any generative-AI-enabled device [19].
  • A narrow exception is debated. UpDoc announced on June 25, 2026 what it calls the first FDA-cleared software using patient-facing LLMs; it is a prescription tool for insulin management in adults with type 2 diabetes, not a general medical chatbot [18].
  • Breakthrough designation is not authorization. FDA granted this status in March 2026 to a patient-facing generative AI application (RecovryAI), a faster review path only [19].
  • January 6, 2026 guidance widened enforcement discretion for transparent, clinician-reviewable decision-support software and for wellness-framed wearables; commentators noted consumer-facing chatbots and symptom checkers were largely left unaddressed [16][17].

In practice, this means the safety checks on a general-purpose chatbot’s health answers are mostly the company’s own policies, not FDA review. That is another reason to keep a clinician in the loop.

Privacy & Your Health Records

  • Connecting medical records or wearable data (as ChatGPT Health and Claude for Healthcare allow) can make answers more personal, but it also means sharing sensitive data with a technology company [8].
  • Consumer chatbot apps are generally not the same as your clinic’s HIPAA-covered systems. Read the privacy terms for retention, human review, and use of your data for model training, and use the available privacy controls.
  • Before pasting a lab report, remove your name, date of birth, address, and ID numbers.
  • Share only what the question needs, and avoid uploading other people’s records without their permission.

A 10-Step Safe-Use Checklist

  1. Use the chatbot to learn and prepare, not to decide.
  2. Give it the full picture (age, medicines you take, conditions, how long symptoms have lasted), because the Oxford study found missing information was a main failure point [1].
  3. Ask it to name sources, then open them and confirm they say what the chatbot claims.
  4. Ask it to argue against your plan: “What could be wrong with this idea?”
  5. Never change, stop, or taper a prescription without your prescriber or pharmacist.
  6. For urgent symptoms, call emergency services, not a chatbot.
  7. Take a printed or saved summary of your question list to your appointment.
  8. Be skeptical of any answer that tells you to hide information from your clinician.
  9. Protect your data: strip identifiers and review privacy settings.
  10. Ask your clinic whether an AI scribe is used, and read your visit note afterward.

Copy-Ready Prompts That Make Chatbots Safer

Prepare for a doctor’s visit

“I’m [age], taking [medicines], with [conditions]. My main concern is [symptom/issue] for [duration]. Help me write a one-page summary and 8 questions to ask my doctor. Do not diagnose me or suggest changing any medication.”

Understand a lab report

“Explain what each marker on this report measures in plain language and which values are outside the reference range. Do not tell me what condition I have. List questions I should ask my clinician about the flagged values.”

Stress-test your own plan

“Here is what I’m thinking of doing: [plan]. List the strongest reasons it could be wrong or unsafe, the evidence quality for each claim, and which parts I must check with a prescriber or pharmacist first.”

Check a claim

“Find primary sources (guidelines, trials, regulators) for this claim: [claim]. Give links, say what each source actually shows, and say clearly where evidence is weak or mixed.”

Check a medicine leaflet

“Summarize this leaflet: what it’s for, common side effects, serious side effects that need urgent care, and what to ask my pharmacist. Do not give me a dose or a schedule.”

Platform Guide: ChatGPT, Claude, Gemini, Perplexity

No platform is a safe substitute for a clinician, and the 2026 audits cover specific models that change often. This table describes typical uses and watch-outs, not a ranking.

Platform Good for Watch-outs
ChatGPT (incl. ChatGPT Health, launched Jan 2026) Structuring question lists for your doctor; explaining terms; can connect records and wellness data in the Health product [8] Confident tone; personalized plans without a prescriber; review privacy settings before linking records
Claude (incl. Claude for Healthcare, launched Jan 2026) Synthesizing long documents (guidelines, papers, leaflets) and flagging where evidence is preliminary; consumer integrations include HealthEx, Function, Apple Health and Android Health Connect [8] Included in the BMJ Open audit like other chatbots [4]; still needs source verification and clinician input
Gemini Current news and fast-moving topics through search integration; checking whether a story has developed Search-backed answers can still be wrong or outdated; open the sources
Perplexity Claims traced to clickable sources; useful for verifying statistics A cited source is not proof the claim is correct; open and read it

Evidence & Regulatory Snapshot

CEBM-adapted tiers: 1 = highest-quality systematic/RCT evidence, 5 = mechanistic or expert opinion.

Use case Evidence Tier 2026 Status
Chatbot-assisted self-triage and health decisions by the public Tier 2b One large randomized study (Oxford, ~1,300): no better decisions than controls [1]
Chatbot answers to everyday health questions Tier 3 Audits found a substantial share problematic [2][3]
LLM clinical reasoning (benchmarks) Tier 3–4 Falls short of reliable independent reasoning; human oversight needed [6][7]
Ambient AI scribes Tier 3–4 Reported error rates 1–3% to ~1 in 3 notes in one UK preprint audit; administrative-tool status [12][14]
General-purpose health chatbots (regulatory) n/a Not FDA-authorized as devices; one narrow insulin-management LLM software cleared; one breakthrough designation [18][19]

Bottom Line

AI chatbots are a genuinely useful way to understand medical language and prepare for appointments, and the 2026 evidence is clear that they are not yet a safe way to make health decisions on your own. Keep them in the green zone, verify anything in the yellow zone, stay out of the red zone, and keep a clinician involved in any change to your treatment.

Need a clinician to talk it through? You can speak with a doctor through our telehealth partner to discuss what’s appropriate for your situation.

Frequently Asked Questions

Is it safe to use ChatGPT, Claude, or Gemini for medical advice?
It is reasonably safe for learning and preparing, and unsafe for deciding. Use a chatbot to understand a term, summarize a leaflet, or draft questions for your doctor. Do not use it to diagnose yourself, to decide whether symptoms are an emergency, or to start, stop, or change a prescription. A 2026 Oxford randomized study found chatbot users did not make better health decisions, and a 2026 BMJ Open audit found close to half of popular chatbots’ health answers were problematic.
Can an AI chatbot tell me whether I can stop or lower my medication?
No. Medicine changes belong with your prescriber or pharmacist. Some drugs, including anti-seizure medicines, blood thinners, blood-pressure and diabetes drugs, antidepressants, and steroids, can cause serious harm if reduced or stopped abruptly. A wrongful-death lawsuit filed in October 2026 alleges a chatbot produced an unsupervised taper schedule for an anti-seizure drug; those allegations have not been tested in court.
Are ChatGPT Health and Claude for Healthcare FDA-approved?
No general-purpose health chatbot is FDA-authorized as a medical device. ChatGPT Health (launched Jan 7, 2026) and Claude for Healthcare (Jan 11, 2026) let users connect medical records and wellness data, but connecting records does not make them cleared diagnostic tools. FDA’s AI-enabled device list identifies no LLM-based device, and one narrow insulin-management software (UpDoc) is described by its maker as the first cleared product using a patient-facing LLM.
Is it safe to upload my lab results or medical records to a chatbot?
It depends on the service and its privacy terms. Consumer chatbot apps are generally not the same as your clinic’s HIPAA-covered systems, so read what the company says about retention, human review, and model training, and use the privacy settings available. Remove names, dates of birth, and ID numbers before pasting text, and avoid uploading records you would not want stored.
What should I do if my doctor uses an AI scribe?
You can ask whether one is in use, how the recording is stored, and who can access it. Afterward, read your visit summary and check that medicines, doses, allergies, and the plan match what was said. Audits in 2026 found scribes can invent details or swap similarly named drugs, and patients are often the ones who catch the error.
How can I tell if a chatbot’s health answer is reliable?
Check that it names a specific, checkable source (and open it), that it states uncertainty, that it asks about your situation, and that it tells you when to see a clinician. Treat confident answers with no sources, invented-looking citations, or advice to hide something from your doctor as red flags.
When should I skip the chatbot and get urgent help?
Call your local emergency number (911 in the US) for chest pain, trouble breathing, signs of stroke (face drooping, arm weakness, speech difficulty), severe allergic reaction, heavy bleeding, a seizure, or confusion. If you are thinking about harming yourself, call or text 988 in the US or contact your local crisis service. Do not wait for a chatbot’s opinion.
Will AI chatbots get safe enough to replace doctors?
The 2026 evidence does not support that. Models can do well on written exam-style cases, but real use depends on what people tell them and how they act on the answers. Regulators and researchers still describe human clinician oversight as necessary.

Medical Disclaimer: This page is for educational purposes only and does not constitute medical advice. It is not a substitute for professional diagnosis or treatment. Lawsuit allegations described here are unproven. In an emergency, call your local emergency number (911 in the US).

References

1. Associated Press via ScienceAlert. “Medical Chatbots Are Coming. Here’s What You Need to Know Before Using One” (Oxford study, about 1,300 participants). sciencealert.com

2. BMJ Group. “Substantial amount of medical information provided by popular chatbots inaccurate and incomplete.” April 2026. bmjgroup.com

3. News-Medical. “Study finds popular AI chatbots often give problematic health advice.” April 16, 2026. news-medical.net

4. TeleDirectMD. “Half of Chatbot Health Answers Were Problematic: Practical Safety Rules for Patients Using AI” (summary of the BMJ Open audit of five chatbots). 2026. teledirectmd.com

5. The Washington Post. “Thinking of using a chatbot for medical advice? Read this first.” April 21, 2026. washingtonpost.com

6. Bedi S, et al. “Holistic evaluation of large language models for medical tasks with MedHELM.” Nature Medicine. 2026 Mar;32(3):943-951. pubmed.ncbi.nlm.nih.gov

7. Medical Xpress. “AI remains lacking in clinical reasoning abilities, according to study of 21 large language models.” JAMA Network Open, reported April 2026. medicalxpress.com

8. Epstein Becker Green. “Health Care Without the Hospital: ChatGPT Health and Claude Go Direct to Consumers.” 2026. ebglaw.com

9. Law Commentary. “OpenAI Sued After ChatGPT Allegedly Told Woman to Stop Seizure Drug Before She Died.” October 28, 2026. lawcommentary.com

10. OpenAI Faces Wrongful Death Suit after ChatGPT Advised Fatal Drug Taper. October 2026. thefocalpoints.com

11. Resultsense. “Healthwatch England: ambient AI scribes mis-recording medicines and diagnoses.” August 31, 2026. resultsense.com

12. “One note in three: a verified census of three deployed AI scribes” (preprint, not peer-reviewed). August 31, 2026. emergentmind.com

13. Cullen A. “Clinicians are using AI scribes, but what happens to your medical data?” The Conversation / University of Melbourne, August 2026. findanexpert.unimelb.edu.au

14. Nurse.org. “AI Charting Tools Show Racial Bias and Make Errors” (summary of an npj Digital Medicine analysis). 2026. nurse.org

15. University of Edinburgh. “AI scribes may fail to capture patients’ experiences” (BMJ Digital Health & AI review). August 31, 2026. usher.ed.ac.uk

16. Foley & Lardner. “Digital Health Policy: FDA Relaxes Restrictions Over Wearables and AI Decision-Making Tools in Two New 2026 Guidances.” January 2026. foley.com

17. Lazarus A. “FDA loosens AI oversight: What clinicians need to know about the 2026 guidance.” KevinMD, January 18, 2026. kevinmd.com

18. McGuireWoods. “A Pathway for Clinical AI Developers Opens: FDA Clears First SaMD With Patient-Facing LLM.” July 2026. mcguirewoods.com

19. Congressional Research Service. “FDA Regulation of AI-Enabled Devices.” June 2026. congress.gov

Comments

Post a Comment

Labels

Show more

Archive

Show more

Popular posts from this blog

GLP-1 Muscle Loss: How to Preserve Lean Mass, Strength & Metabolic Health

-
Image
GLP-1-based medications have changed the treatment of obesity by making substantial, sustained weight loss possible for many people. But as these medicines become more widely used, another question has become increasingly important: How much of the weight lost on a GLP-1 drug comes from muscle or other lean tissue—and can that loss be minimized? The evidence shows a nuanced answer. People taking semaglutide, tirzepatide and related incretin-based medicines generally lose much more fat than lean tissue . At the same time, absolute lean mass usually falls during substantial weight loss. Importantly, lean mass is not the same thing as skeletal muscle : DXA-based “lean mass” also includes water, organs and other non-fat tissues. Therefore, a finding that lean mass decreased does not automatically mean that the same amount of skeletal muscle was lost. Recent 2026 meta-analyses suggest that roughly one-quarter to one-third of total weight lost with incretin therapy may be classified as lean ...
Read more

Semaglutide vs Tirzepatide vs Retatrutide: Best Weight Loss Peptides in 2026

-
Image
Medically Reviewed by: Dr. Frank Yap, MD | Written by: OneDayMD Editorial Team | Published: March 2026 | Last updated: October 2026 Quick Answer Semaglutide (“sema”), tirzepatide (“tirz”), and retatrutide (“reta”) represent three stages in the evolution of incretin-based weight-loss medications: a single GLP-1 receptor agonist, a dual GIP/GLP-1 receptor agonist, and an investigational triple GIP/GLP-1/glucagon receptor agonist, respectively. Semaglutide is an FDA-approved GLP-1 receptor agonist that has demonstrated average weight loss of approximately 15% in clinical trials, with some individuals losing more. Tirzepatide, a dual GIP/GLP-1 receptor agonist, has demonstrated average weight loss of up to approximately 21–22% in trials. Retatrutide, an investigational triple agonist targeting GIP, GLP-1, and glucagon receptors, has shown promising weight-loss results in clinical trials but is not yet FDA-approved. What changed in this update (Oct 2026): replaced pre-Phase-3 esti...
Read more

GLP-1 Hub: Approved Drugs, Pills vs Injections, Retatrutide & Side Effects

-
Image
Home › GLP-1 Hub By Dr. Frank Yap, MD | One Day Media Network | Published October 13, 2026 | Last updated: November 2026 | Purpose: evidence and safety education, not individual medical advice Quick Answer As of November 2026, the FDA has approved two GLP-1 weight-loss pills, oral Wegovy (semaglutide 25 mg, December 2025) and Foundayo (orforglipron, April 1, 2026), alongside weekly injections such as Wegovy, Zepbound, Ozempic and Mounjaro. Retatrutide and CagriSema are not approved in the US. Compounded and gray-market GLP-1 products are not FDA-approved. Use a licensed prescriber and a regulated pharmacy. On this page Which GLP-1 medicines are approved in October 2026? Are GLP-1 pills as effective as injections? Retatrutide and CagriSema: what is known so far? Safety and side e...
Read more

Dr William Makis Ivermectin Protocol 2026 – Complete Guide + Patient Outcomes

-
Image
Medically Reviewed by: Dr Frank Yap, MD | Written and Compiled by:     OneDayMD Editorial Team | Last Updated:  October  2026 Dr. William Makis MD (McGill Medicine, 110+ peer-reviewed publications) has treated a large volume of cancer patients using repurposed drugs since 2023 and has documented outcomes publicly on Substack and X. This continuously updated 2026 guide compiles Dr. Makis's latest protocols (from his Substack, X posts through 2026, and direct patient correspondence). We cross-reference his exact dosages, schedules, combinations, safety data, sourcing, and real patient outcomes (anonymized but verifiable on his channels). Important Disclaimer: This content is for educational purposes only and does not constitute medical advice. Ivermectin is used off-label for cancer treatment. Always consult with a knowledgeable physician whom you trust to manage your health. Individual results may vary. Do not self-medicate with ivermectin without proper bloodwork ...
Read more

GLP-1 2.0 and Pipeline: The Next Generation of Obesity Drugs

-
Image
By Dr. Frank Yap, MD  · One Day Media Network · Last Updated October 2026 Quick Answer "GLP-1 2.0" is the second wave of obesity medicines built on semaglutide and tirzepatide. It improves on the first wave in five ways: more potency (retatrutide, a triple agonist, averaged 28.3% weight loss at 80 weeks in Lilly's TRIUMPH-1 trial), new mechanisms (amylin and glucagon), easier formats (daily pills already approved; monthly injections in Phase 3), better body composition , and lower prices . As of September 29, 2026, no next-generation injectable is approved yet : retatrutide's FDA filing is planned for Q1 2027 and CagriSema's FDA decision is expected in late 2026. Approved 2.0-era products today are the oral drugs (Wegovy pill, Foundayo) and high-dose Wegovy HD. Disclosure: This article is educational and is not medical advice. ...
Read more

Top Smart Glasses for the Visually Impaired (2026): eSight, Envision, OrCam & More

-
Image
Medically & Technically Reviewed by: OneDayMD Editorial Team | Originally Published: August 9, 2026 | Updated: October 2026 Quick Answer There is no single "best" smart glasses for visual impairment in 2026 — the right pick depends on how much usable vision remains. For low vision (residual central vision from macular degeneration, diabetic retinopathy, or Stargardt disease), medical-grade enhancers like eSight Go ($4,950) or Acesight VR digitally magnify and boost contrast. For blind or near-blind users, AI-interpretation glasses like Envision Glasses ($2,499), Ally Solos (from $399), or a clip-on OrCam MyEye 3 Pro (remaining dealer stock, $3,689–$4,490) read text and describe scenes aloud. The cheapest mainstream option is Meta's expanding AI-glasses lineup — Ray-Ban Meta, Meta Glasses (Adventurer/Fury), and Meta Glasses by Kylie — starting at $299, which isn't purpose-built for accessibility but has become a popular everyday alternative thanks to ...
Read more

Fenbendazole Joe Tippens Protocol: A Step-by-Step Guide (2026)

-
Image
Medically Reviewed by: Dr Frank Yap, MD | Written by:     OneDayMD Editorial Team | Last Updated:  October  2026 If you have been diagnosed with cancer, or have a loved one that is suffering from cancer, chances are you have heard about the Joe Tippens Cancer Protocol. The Fenbendazole Cancer Protocol has been gaining rapid interest over the past years following some fenbendazole advanced cancer success stories  (800+ case reports) and a 2026 publication, " Real-world Clinical Outcomes of Ivermectin and Mebendazole in Cancer Patients : Results from a Prospective Observational Cohort of 197 patients." We have seen tremendous demand for some sort of guide on how to use fenbendazole for cancer as there is also tremendous confusion both from the healthcare and non-healthcare communities. Joe Tippens founded the protocol (1) after he was told a story about a sci...
Read more

DMSO 101: Benefits, Uses, Dosage and Side Effects (2026)

-
Image
Quick Summary: DMSO (Dimethyl Sulfoxide) is a naturally occurring compound with decades of research behind its use as an anti-inflammatory, analgesic, and cellular-protective agent. FDA-approved for interstitial cystitis, it has been used off-label by athletes, physicians, and patients for musculoskeletal injuries, chronic pain, wound healing, and more. This guide covers what DMSO is, how it works, the evidence for its key uses, safety considerations, dosing guidance, and where to find practitioners who use it. All off-label uses are experimental. Consult a qualified healthcare provider before use. A Patient Story: From Lung Transplant List to Clear CT Scans In 2022, Erica Eyres, a vigorous fifty-six-year-old aerobics instructor who had struggled to breathe, was given “absolutely devastating” news: She might need a lung transplant. She had never smoked, ran cross-country track in high school, and was a personal trainer for years, but, by 2024, a transplant assessment was arranged. “I d...
Read more

Ivermectin, Fenbendazole and Mebendazole: A Peer-Reviewed Protocol for Cancer Treatment (2026 Update)

-
Image
Medically Reviewed by: OneDayMD Editorial Team | Last Updated: April 2026 In September 2024, a first-of-its-kind protocol utilizing ivermectin, fenbendazole, and mebendazole for cancer treatment was peer-reviewed and officially published in the Journal of Orthomolecular Medicine . Led by researchers Dr. Ilyes Baghli, Dr. Pierrick Martinez, and FLCCC's Dr. Paul Marik, the protocol applies antiparasitic drugs — originally developed to combat parasites — to cancer treatment, building on emerging preclinical and clinical evidence of their anticancer properties. This trio of repurposed drugs has been shown to disrupt the growth of cancer cells, particularly by targeting microtubules, the essential structures that allow cancer to multiply uncontrollably. What makes this protocol even more potent is the synergistic effect when these drugs are used together, creating a powerful new affordable weapon against cancer. Journal of Orthomolecular Medicine 2024 Targeting the Mitochondrial-Stem C...
Read more

How to Get Ivermectin in the US: Pharmacies, Legal Status & State-by-State Access (2026)

-
Image
Medically Reviewed by: OneDayMD Editorial Team | Last Updated: August 2026 An updated, evidence-based guide to ivermectin's federal and state legal status, how to obtain a prescription, which states allow pharmacist-dispensed or OTC access, and a directory of pharmacies that fill ivermectin prescriptions in the US. Quick Answer (AI & Search Summary) Ivermectin is still classified federally as a prescription-only drug for human use. It has not been reclassified as OTC by the FDA. However, five states — Arkansas, Idaho, Louisiana, Tennessee, and Texas — have passed laws letting pharmacists dispense it without a doctor's prescription, and roughly half a dozen more have active bills. Everywhere else, you need a prescription from a doctor, urgent care provider, or telehealth service. Individual pharmacists retain the right to decline to dispense or stock it even where state law allows it. Table of Contents Current Legal Status (Federal & State) State-by-...
Read more