Nickol Georgy
GenerativeAI is rapidly altering medical diagnostic machinery, notably to interpret complex information like electrocardiograms (ECGs). Harvard Medical School is integrating AI into the school curriculum, training future physicians to work alongside machine-learning software. Notably, an interesting component of AI use within the healthcare industry is a model, which was trained on close to a million ECGs, that can identify dozens of conditions, sometimes even better than cardiologists.
Such systems can be even more potent when combined with real-world evidence from electronic health records (EHRs) and patient wearables. By drawing on this larger dataset, AI can spot patterns that are subtle and might otherwise be overlooked by traditional diagnostics. But whereas the potential is enormous, there certainly exists current biases in these datasets. Furthermore, hallucinations that generative AI still produces poses a concern. However, in collaboration with healthcare workers, this can be an excellent tool to help save thousands of people. Additionally, generative AI can be used to help with documentation so that doctors can spend more time socializing and connecting with patients rather than with administrative work. This has already been implemented in Brigham and Women’s Hospital—helping the doctor stay connected when talking to the patient rather than looking away midway to write notes on the computer software.
This Case Study is part of a larger essay on Artificial Intelligence in Health Care.
