Google Cloud Releases New Clinical Generative AI Tools at HIMSS24 – MedCity News

author
1 minute, 59 seconds Read

On Tuesday at the HIMSS conference in Orlando, Google Cloud unveiled new AI features aimed at helping providers, payers and any other healthcare organizations seeking to make better use of their clinical data.

The tech giant’s first announcement was the launch of its Vertex AI Search for Healthcare. The product is a generative AI-powered search function that can help clinicians quickly parse through patient data and clinical notes.

Searching through clinical data is difficult and incredibly time-consuming, pointed out Aashima Gupta, Google Cloud’s global director for healthcare strategy and solutions, in an interview at HIMSS. With its newly released search function, Google Cloud seeks to give clinicians a user experience similar to the one they get when they use Google.com to search the web in their daily lives, she declared.

This tool is available as an API, which allows users to customize their settings to best fit their EHR workflows, Gupta noted. She also explained that the tool inspires trust from its users by displaying footnotes that link back to the specific data points included within the generated search results.

In addition to the launch of Vertex AI Search for Healthcare, Google Cloud also announced that it is adding two new capabilities to its healthcare-specific large language model (LLM). The LLM, called MedLM, is a generative AI model designed to speed up workflows for clinicians and medical researchers. 

The first new capability is an API designed to help clinicians better classify chest x-rays for screening and diagnostic use cases. The second is an API that gives clinicians a chronological list of a patient’s conditions, along with short AI-generated summaries about each of them.

MedLM is currently being tested by Google Cloud customers including Highmark Health, Mayo Clinic and HCA Healthcare. Nurse handoffs is a major use case that health systems are exploring, Gupta noted.

“When a nurse gets off their shift, they have to give the new nurse a lowdown for each patient — they have to spend 10-15 minutes for each patient. If you have 10 patients, that’s an hour and half spent on handoff,” she stated.

In situations like this, MedLM is beneficial because it not only saves nurses time by automatically generating a briefing for each patient, but also cites the specific patient data points within the EHR that were used to craft that summary, Gupta explained.

Photo: zhuweiyi49, Getty Images

This post was originally published on this site

Similar Posts