An AI deep learning model based on patient health care records has been found to predict patients’ health outcomes with a high degree of accuracy, including length of hospital stay, time of discharge and death. The trial at two US hospitals has been reported in the highly regarded Nature journal, and used data such as patients’ medical history, vital signs and clinicians’ notes.
Using AI in healthcare by trained clinicians could significantly improve healthcare delivery and thus, patient outcomes. However, without regulatory frameworks set in place to monitor how healthcare data is used and shared, private companies in the US could misuse such data. Further, feeding a tech giant such as Google with huge sets of healthcare data from multiple sources could contribute to its gain in monopoly over healthcare.
Is this a sign of much broader use of big data and AI in healthcare to come? Should we welcome the use of our personal data for the benefit of greater performance of such models? And how can such advances in technology be scaled rapidly and transparently at the same time?