Epic adding AI, other capabilities to support clinicians
Artificial intelligence and improved capabilities designed to incorporate other sources of patient data were among the product improvements Epic customers were updated on at the vendor's annual users group meeting Tuesday.
Epic used the annual event at its Verona, Wis.-based headquarters to update an estimated 8,000 hospital IT executives and managers on progress it’s making to include upgrades in its products for 2018 and beyond, as well as achievements by the company and its customers.
The company highlighted its ability to use data and insights from its customers to improve both its products and hospital organizations’ ability to use those results to improve care. For example, Epic has developed 319 data-validated benchmarks that hospitals can use for improvement, says Carl Dvorak, Epic’s president.
The benchmarks “can help you understand what you can do to influence outcomes,” he says. Other comparative functions within the system “highlight areas where you seem to be an outlier among your peers, and highlight what high-achieving users are doing so you can ‘stand on their shoulders,’ “ using their practices to improve.
Artificial intelligence and machine learning capabilities are being sprinkled throughout Epic’s products next year to improve care and facilitate the use of the records system by clinicians. For example, the software is able to use machine learning to assist clinicians in documentation, which can help speed up work on the system, says Sumit Rana, Epic’s senior vice president of research and development. A tablet-based version of the program will enable it to be used to document all care in the emergency department setting, which will improve clinician efficiency and support better patient interaction.
The company also is making progress on improving the use of Rover, an app, giving clinical users of Epic’s EHR secure access to tools for clinical review, patient list management, medication administration and more on their smartphones. Artificial intelligence capabilities enable Rover to anticipate what clinicians most frequently request, assist in documenting care and even in locating medical equipment in a hospital building.
Improving the systems’ interactions with clinicians is both a function of making it easier to use while improving training methods and better enabling clinician support, executives said.
For example, Epic is creating 90-second videos aimed at specific tasks that clinicians need help with, Rana says; other capabilities in the system can help hospital administrators identify users who are struggling with the system, enabling them to be targeted with intensive intervention.
Epic is working to improve the ways it interacts with both users and hospital IT staff and executives, says Judy Faulkner, the company’s founder and CEO. “We need to highlight the value of our features in clear end-user terms and not in techie terms,” she notes. “We’ve found we need to go back and focus on physician training,” she added, reporting that the company has found adoption and quality of usage is best when specialists who know how to use the system well are able to assist other physicians in the same specialty at an organization.
According to Faulker, Epic and other EHR systems will be challenged in the near future to hold more than just the patient-generated data within the walls of healthcare organizations; other relevant information that influences patient health and care will be coming from outside sources.
Faulkner also highlighted EHR interoperability among Epic users, which exchange an average of 2 million patient records daily, with about 10 percent of those being exchanged across state lines.
“The network is impressive because the community of Epic clients have all opted into the network to share patient records,” says Christopher Longhurst, MD, CIO at UC San Diego Health, an integrated health system.
Epic is also continuing to migrate clients to cloud-based delivery of its product, hosted at the data center on its corporate campus; 40 clients have signed agreements to do so, and 20 of those are in production to make the move in the near future.
UC San Diego switched to the hosted model, with the aim of shrinking and eventually eliminating its data center, Longhurst says. “Holding data is not a core competency of healthcare organizations,” he says. “We can’t put together the level of data center that Epic has. We should be out of the data center business, and we can then move those operations to data centers that are more highly reliable.”