Healthcare leaders were given an advanced preview of MEDITECH’s latest AI initiatives during last week’s MEDITECH LIVE leadership event, as the company looks to build upon its ongoing partnership with Google Cloud to build generative AI capabilities into other areas of the Expanse EHR. MEDITECH LIVE is the organization's foremost annual customer event that attracts leaders from healthcare organizations worldwide to network, share stories of innovation, and strategize for the future.
MEDITECH and Google have been working together on AI initiatives since 2021. At MEDITECH LIVE, MEDITECH announced the next phase of this partnership, which includes the companies working on the auto-generation of clinical documentation. MEDITECH is focusing on the hospital course narrative at the time of discharge.
“Clinical documentation is one of the most burdensome areas of care delivery,” explained MEDITECH Director of Marketing Rachel Wilkes. “We see great potential in leveraging generative AI to reduce the time it takes to document at discharge, particularly for patients with long stays or high acuity levels. By using AI to automatically generate an initial draft of this hospital course narrative that a clinician will review within the Expanse discharge plan, our customers will be able to improve the physician's workflow.”
Discrete, structured data elements such as medications and test results can already be pulled into the discharge summary. However, the hospital course narrative is free text and often requires a time-consuming review of multiple days of multidisciplinary documentation to succinctly summarize the most salient points. Generative AI can save clinicians time by creating summaries, which clinicians then review and edit, helping streamline documentation.
Search and Summarization Capabilities Next Steps
Google Health Senior Clinical Specialist Dr. Peter Clardy, MD, joined Randy Brandt, PA-C, a primary care physician assistant at Mile Bluff Medical Center, to provide MEDITECH LIVE attendees with a demonstration of the EHR integration of Google Health’s search and intelligent summarization capabilities. This technology, which leverages natural language processing to intelligently pull information from both structured and unstructured data, is currently being implemented at Mile Bluff, with Brandt serving as team lead.
“Embedding Google Health’s search and summarization capabilities into our EHR is going to bring Mile Bluff to the next level regarding healthcare innovation and patient satisfaction,” said Brandt. “We sometimes receive scanned documents that can be over 150 pages long, and the information I need may only be in one paragraph. Having a tool to search through all these pages for me and find the information I need will save me so much time.”
Beyond simply synthesizing information, the search and summarization capabilities intelligently organize and surface information so that the most important conditions are always highlighted.
Dr. Clardy also demonstrated the potential in leveraging LLMs to further synthesize a patient’s medical history, to give clinicians a better understanding upfront.
“In collaboration with MEDITECH, we are exploring ways generative AI can summarize complex unstructured medical information, giving providers a high-level overview of each problem. The summaries will be linked (or "grounded") to evidence in the medical record, which allows immediate verification by the provider as to the source of the information surfaced,” explained Dr. Clardy. “These new tools combine improved medical language understanding with search and summarization, and they will empower providers to ask complex questions about a patient's history. We hope the development of new interactive tools will provide clinicians with the right information, at the right moment in time, with a direct way to review and validate the details.”
Additional AI Use Cases
Beyond its work with Google Cloud, other AI use cases highlighted throughout the event include MEDITECH’s integration with multiple voice assistant vendors to leverage ambient listening devices as digital scribes. By taking advantage of large language models and natural language processing, Expanse can use this information to automatically generate clinical visit notes, which may then be reviewed and approved by the physician.
MEDITECH also showcased its Expanse Virtual Assistant solution, which is now LIVE and available for general delivery. Virtual Assistant uses conversational AI to streamline chart navigation, and now also includes the ability for physicians to use AI to place orders.
MEDITECH’s Executive Vice President and COO Helen Waters has long advocated for innovations like artificial intelligence and large language models in healthcare delivery.
“We’ve actually spent the last several years exploring AI and partnering with experienced industry leaders to help inform clinical decisions, improve efficiency, and ease staff burdens,” she said.
“While interest in AI (and specifically generative AI) has certainly accelerated over this past year, MEDITECH has gotten out ahead of the hype and is turning our vision into a reality. These technologies have the potential to dramatically improve the healthcare experience in the coming months and years, so I’m glad we’ve hit the ground running.”
