A recent study conducted by Dutch researchers has found that an AI bot called ChatGPT may be just as effective, if not better, at diagnosing patients in emergency care compared to real doctors. ChatGPT, developed by OpenAI, uses advanced language models and has previously demonstrated success in passing medical exams and displaying a better bedside manner than human doctors.
In this study, the researchers analyzed the performance of ChatGPT in an emergency department setting. They anonymized the details of 30 patients who were treated at Jeroen Bosch Hospital's emergency department in 2022 and inputted physicians' notes on patients' signs, symptoms, and physical examinations into two versions of ChatGPT. The AI bot was asked to generate a list of possible diagnoses for each patient, which were then compared to the lists created by real doctors.
The results showed a significant overlap of around 60% between the shortlists generated by ChatGPT and the lists created by doctors. In 87% of cases, doctors had the correct diagnosis within their top five likely diagnoses, compared to 97% for ChatGPT version 3.5 and 87% for version 4.0.
One interesting case highlighted in the study involved a patient with joint pain that was relieved by painkillers but recurred along with other symptoms. While doctors suspected rheumatic fever, ChatGPT correctly diagnosed vasculitis as the most likely condition based on the patient's physical exam and additional tests.
However, it is important to note that ChatGPT is not a medical device, and concerns exist regarding privacy when using it with medical data. Nevertheless, researchers believe that AI could potentially save time and reduce waiting times in the emergency department, particularly in supporting less experienced doctors or identifying rare diseases.
Independent experts have welcomed the study's findings, suggesting that AI could eventually be used to assist doctors in hospitals. While the adoption of ChatGPT in clinical settings is still far off, further research in this area could ultimately support the work of busy health professionals and improve patient care.
In conclusion, the study indicates that ChatGPT performed well in diagnosing emergency department patients, demonstrating the potential for AI to support doctors and reduce waiting times. However, it is crucial to proceed with caution and address privacy concerns associated with using AI in healthcare settings.