A recent study conducted by Washington State University’s Elson S. Floyd College of Medicine sheds light on the potential obstacles to using artificial intelligence (AI) during medical emergencies. The study focused on OpenAI’s ChatGPT program and its ability to assess the cardiac risk of simulated patients experiencing chest pain.

One of the key findings of the study was the inconsistency in ChatGPT's conclusions when presented with the same patient data. Dr. Thomas Heston, the lead researcher, noted that ChatGPT's performance was not consistent, often giving different risk assessments for the same patient data.

This inconsistency is especially concerning in critical cases where accurate and timely assessments are crucial for providing appropriate medical care. Patients presenting with chest pain could be suffering from various conditions, requiring healthcare providers to quickly and accurately diagnose and treat them.

The study also compared ChatGPT's performance to traditional methods used by doctors to assess cardiac risk, such as the TIMI and HEART protocols. The results showed that ChatGPT had a lower agreement rate when evaluating patient data against these established protocols, raising concerns about its reliability in high-stakes medical situations.

Dr. Heston emphasized the potential of AI to enhance healthcare but also stressed the need for further research to address its limitations. While AI technology can offer benefits like quick access to patient data and assistance in generating differential diagnoses, there are still areas where human nurses and doctors play a crucial role in providing quality care.

In conclusion, as AI continues to evolve, it is essential to evaluate its performance rigorously, especially in high-risk healthcare settings, to ensure patient safety and improve medical decision-making.