AI outperforms doctors in Harvard trial of emergency triage diagnoses
Summary
A Harvard study tested how well artificial intelligence (AI) systems perform in emergency room diagnosis compared to human doctors. The AI was more accurate in quickly diagnosing patients and making treatment plans when given medical records, showing potential to assist doctors in emergency care.Key Facts
- The study involved 76 emergency room patients at a Boston hospital.
- AI identified the correct or close diagnosis 67% of the time; human doctors were right 50%-55% of the time.
- With more detailed patient information, AI accuracy increased to 82%, compared to 70-79% for human doctors.
- AI outperformed 46 doctors in making longer term treatment plans, scoring 89% accuracy versus 34% for humans.
- The AI only reviewed written patient data and did not assess physical signs like appearance or distress.
- Researchers said AI is unlikely to replace doctors but will work alongside them in a team including the doctor, patient, and AI.
- Nearly 20% of US doctors and about 30% of UK doctors use AI regularly, mostly for clinical decision support.
- Experts highlighted the need for clear rules on responsibility if AI makes errors in medical care.
Read the Full Article
This is a fact-based summary from The Actual News. Click below to read the complete story directly from the original source.