All articles tagged with #chatgpt health
A Nature Medicine study tested ChatGPT Health with 60 real-life medical scenarios and found it under-triaged 51.6% of emergencies and over-triaged 64.8% of non-urgent cases, with some life-threatening events misclassified, though stroke was correctly triaged 100%. OpenAI says ChatGPT Health isn’t designed for diagnosis and emphasizes clinician involvement; researchers urge more rigorous testing before wider use.
A Nature Medicine study tests ChatGPT Health with 60 clinician-authored vignettes across 21 clinical domains under 16 factorial conditions (960 responses). Performance follows an inverted U-shape, with the most dangerous errors at extremes: 35% for non-urgent cases and 48% for emergencies. Among gold-standard emergencies, 52% were under-triaged (e.g., could misdirect diabetic ketoacidosis or impending respiratory failure to 24–48 hours instead of ED), while classic emergencies like stroke and anaphylaxis were correctly triaged. Anchoring by family or friends shifted edge-case triage toward less urgent care (OR 11.7). Crisis-intervention messages activated inconsistently across suicidal ideation presentations. No significant effects by patient race, gender, or barriers to care. Overall, the findings raise safety concerns and call for prospective validation before consumer deployment of AI triage tools.
OpenAI has launched ChatGPT Health in the US, a new feature that analyzes users' medical records and app data to provide personalized health advice, raising privacy concerns despite assurances of data separation and protection; the tool is not yet available in the UK or Europe and aims to support, not replace, medical care.