No longer interested? Unsubscribe |
|
|
HEART SURGEON BEGS AMERICANS: "STOP DOING THIS TO YOUR AVOCADOS"
|
|
|
| Are you turning healthy fruits into highly unhealthy fruits, without even realizing it? Fruit can be one of the healthiest things you can put into your body, but the majority of Americans are guilty of making this single breakfast mistake that can counteract most of the health benefits of fruit. Some fruits can help with occasional constipation, reduce bloating, and even help burn calories. Others can wreak havoc on your digestive system and energy levels. In this viral presentation, world-renowned heart surgeon Dr. Steven Gundry showcases the three fruits to stop eating today. Click here to learn the top 3 common breakfast foods that are the cause of your fatigue. |  | P.S. Eliminating these fruits could forever change the way we help increase energy levels, decrease brain fog, support digestion, and even lose weight. | | *All individuals are unique. Results can and will vary. | | | | | If you no longer wish to get these emails, you can unsubscribe by clicking here. | |
|
|
An AI Chatbot Took Patients' Histories Before Real Doctor Visits. Here's What a Lancet Study Found In a supervised feasibility study of 98 primary care patients in Boston, Google's AMIE completed pre-visit interviews with no safety stops and often listed the eventual diagnosis. | | | The Capital Brief | October 10 | | | Digital Health Edition | | An AI Chatbot Took Patients' Histories Before Real Doctor Visits. Here's What a Lancet Study Found | | | | Millions of people already ask chatbots about their symptoms. A study in The Lancet tested what happens when a diagnostic AI talks with real patients before they see their doctor. Researchers from Google and Beth Israel Deaconess Medical Center in Boston published the study in The Lancet on October 8. Adults with scheduled urgent primary care visits chatted by text with AMIE, a diagnostic conversational AI built on Google's Gemini models, before their appointments. The analysis included 98 patients enrolled from April to November 2025. The AI took each patient's history and produced a list of possible diagnoses and a summary for the doctor. A physician watched every conversation live and could stop it for safety reasons. No conversation had to be stopped for safety. The final diagnosis, confirmed in medical records eight weeks later, appeared in the AI's list of possible diagnoses in about 90% of cases. In blinded reviews described by the authors, the AI's diagnostic lists were rated similar in quality to doctors', while doctors' management plans were judged more practical and cost-effective. This was a small, single-center feasibility study without a control group, run and funded by Google. It shows the approach can be tested safely under supervision, not that AI can replace a doctor. The useful idea is that AI may be most valuable as preparation for a clinical visit, not a substitute for one. This edition explores the evidence, not a recommendation to use any AI tool for medical decisions. Symptoms deserve assessment by a clinician. | | | | AMIE, short for Articulate Medical Intelligence Explorer, is a research system developed by Google. Patients chatted with it by text on a computer up to five days before their visit. It asked about symptoms, history and concerns, much like a clinician taking a history. It then generated a list of possible diagnoses and a summary for the treating doctor. The AI did not examine patients, order tests or access medical records. It is a research tool, not a product available to patients. Doctors made all actual diagnoses and treatment decisions. | | 02 | How the study was designed | | | The study took place at a primary care practice affiliated with Beth Israel Deaconess Medical Center. Adults with a scheduled non-emergency urgent care visit for a single main complaint were eligible. Participants needed to speak English, use the patient portal and have a laptop or desktop computer. Pregnancy-related and mental health complaints were excluded. Of 1,452 scheduled visits during the study period, 140 patients consented, and 98 completed both the AI chat and the doctor visit. About two-thirds of participants were women, and about half were under 50, younger than the clinic's typical urgent care patients. There was no comparison group of patients who did not use the AI. | | | | Safety was the main outcome of this feasibility study. None of the conversations required a safety stop by the supervising physician. Supervisors intervened a few times to clarify emergency criteria or correct details. Live supervision means the results do not show how the AI would perform alone. The authors and an accompanying Lancet commentary emphasize what happens after the AI speaks as a key question. Unsupervised use would require separate testing. Safety in a small, supervised study is a starting point, not a guarantee. | | | | The final diagnosis, checked against medical records eight weeks later, appeared in the AI's list of possible diagnoses in about 90% of cases. According to the authors' earlier preprint, it was the AI's top choice in about 56% of cases and in the top three in about 75%. Blinded evaluators found no significant difference between the AI and doctors in the quality of diagnostic lists. Doctors' management plans were rated more practical and more cost-effective. The AI could not examine patients, which limits diagnosis for many conditions. Some final diagnoses were presumptive rather than confirmed by tests. Exact final figures should be checked in the published paper. | | 05 | How patients and doctors reacted | | | Patients' attitudes toward AI improved after the chat and stayed improved after the visit. Fewer than half felt the AI would keep information confidential or seemed fully trustworthy. Most doctors who reviewed the AI's summaries found them helpful for preparing for the visit. Many said the summaries likely influenced what they did in the visit. Trust, privacy and clinician workload are central questions for real-world use. These reactions come from a small group at one practice. Broader acceptance would need to be studied. | | | | The study was small, at a single center, with no control group. Participants needed a computer and English, excluding many patients. About 7% could not finish because of device problems, according to the authors. Patients and doctors knew they were being observed, which can change behavior. The underlying AI model was changed partway through the study. The study was funded and largely run by Google, and Google-employed authors may hold company stock. Larger, controlled trials are needed before drawing conclusions about benefits. | | | | If you use a chatbot to think about symptoms, treat its answers as questions to bring to a clinician, not as a diagnosis. Seek emergency care for warning signs such as chest pain, trouble breathing or sudden weakness, without waiting for any app. Avoid sharing sensitive health information with tools whose privacy policies you do not understand. Writing down your symptoms, timeline and medications before a visit can help any appointment. Ask your health system whether it uses AI tools and how your data are handled. Tell your clinician if you have used AI tools to research your symptoms. Your doctor remains responsible for diagnosis and treatment decisions. | | 08 | The long view on AI in medicine | | | AI tools are moving from laboratory tests to real clinics, and evidence is beginning to accumulate. This study suggests a supervised diagnostic AI can safely gather histories and support doctors in preparing for visits. For patients, practical questions include: Who reviews what the AI says? How are my data protected? Can I opt out? For health systems, the question is whether such tools save time and improve care without adding risk. For regulators and researchers, controlled trials and independent evaluations are the next step. The long-view idea is that AI may be most useful as a partner in preparation, not a replacement for clinical judgment. A good conversation is the start of care, not the end of it. | | | | The Takeaway | | 1 | In a supervised feasibility study of 98 primary care patients, a diagnostic AI chatbot completed pre-visit interviews with no safety stops. | | 2 | The eventual diagnosis appeared in the AI's list of possibilities in about 90% of cases. | | 3 | Doctors' management plans were rated more practical and cost-effective than the AI's. | | 4 | The study was small, single-center, uncontrolled and funded by Google. | | 5 | The findings support further controlled testing, not using AI in place of a clinician. | | | Stay patient, stay curious,
T.C. Brown
Editor, True Capital Talk | | | | Sources & Further Reading | | | | For educational and informational purposes only. Nothing here is a diagnosis, treatment plan, or individualized medical advice. Do not rely on any AI tool for medical decisions; consult a qualified clinician about symptoms or health concerns. | | |
|
|
This message was delivered by True Capital Talk (TCT), a brand of TerraTrance Technologies, LLC.
Need assistance or have a question? Simply reply to this email or reach us at [email protected] All content in this email, including text and images, is the property of TerraTrance Technologies, LLC and may not be copied, redistributed, or shared without prior written permission.
© 2026 True Capital Talk (TCT). All Rights Reserved.
200 Broadway Blvd NE, Albuquerque, NM 87102
You may unsubscribe at any time if these emails are no longer useful to you.
|
|
|
|
