CloudMedx Inc., a software developer, offers Cloud-based predictive health analytics and care coordination platform, announced earlier this week, that its Artificial Intelligence-powered platform, CloudMedx AI, took a modified version of the United States Medical Licensing Examination (USMLE) Step 1 a few weeks ago that is performed better than human doctors in a simulated study. After this performance, it indicated that AI can not only comprehend medical concepts but also offer insights that may help physicians.
The study also illustrated that when human intelligence is coupled with machine intelligence, it may create advanced outcomes for tasks that would otherwise be very time-consuming and difficult for humans. The USMLE is an examination process where physicians take in order to get licensed in the country. The modified version of USMLE had questions like case studies where a situation of a patient is explained, and the examinee, both humans, and the AI were anticipated to implement medical knowledge and reasoning to answer multiple choice questions. The simulated exam didn’t ask simple facts that may be answered through a keyword search, but rather it explained compound scenarios. CloudMedx’ Clinical AI assistant amplifies clinical operations and revenue cycle management for health systems. The system utilizes NLU (Natural Language Understanding) and deep learning and can be incorporated into EHR (Electronic Health Record) to provide clinical insights within workflows to expand hospital staff for advancing operations and documentation. Currently, Comcast is working with some of the top healthcare centers in the country.
CloudMedx AI outperformed all the doctors by achieving a score of 85 percent correct answers. While 81 percent was the highest human score and low score was 68 percent, with an average score of 75 percent. Eventually, this first round study revealed that AI may boost human doctors so that their combined medical intelligence brings efficiencies and advances the overall process.