Children's National Hospital in Washington, D.C., is partnering with Microsoft to explore how generative AI can improve pediatric care.
"We've always been very interested in the power of generative AI," Jessica Herstek, MD, chief medical informatics officer at Children's National Hospital, told Becker's. "Microsoft has been pushing the envelope in clinical AI, and we saw an opportunity to work with them to prototype tools that could really benefit both our staff and patients."
During a recent two-day workshop with Microsoft experts, Children's National focused on prototyping AI applications to address key challenges in the hospital. Dr. Herstek explained that developing these use cases was a collaborative effort involving clinical, operational, and technical leaders. Input from frontline staff helped identify pain points and prioritize areas where AI could make the most immediate impact.
"Collaboration is key when developing AI tools in healthcare," Dr. Herstek said. "We made sure to include diverse voices — from clinicians to technical experts — to ensure the prototypes were grounded in real-world needs. The more input we get, the more successful these tools will be."
Addressing healthcare documentation and workflow challenges
One prototype explored during the workshop was "Beacon," an AI-powered tool designed to simplify access to policy and procedure documents. Beacon enables staff to search internal documents in real time.
"As we all know, having quick access to information is critical in healthcare," Dr. Herstek said. "Tools like Beacon could drastically reduce the time it takes for our team to find and reference the documents they need, especially during stressful moments."
Another focus of the workshop was improving documentation efficiency. Dr. Herstek highlighted the burden clinicians face in documenting patient information in EHRs.
"In my role as a hospitalist, writing discharge summaries is an ongoing challenge," Dr. Herstek said. "I often find myself spending too much time documenting the same things over and over. The hope is that with AI, we can automate some of that documentation and allow clinicians to focus more on patient care."
The team worked on generative AI tools to summarize patient information more effectively. AI could convert medical language into understandable summaries for patients and families while generating concise reports for primary care providers, coders, and insurers. According to Dr. Herstek, this could significantly reduce repetitive documentation tasks.
The workshop also explored using AI to extract data from physicians' notes, addressing challenges with free-text documentation in EHRs. The prototype could identify critical follow-up appointments, streamlining the process of transferring information to case managers or scheduling systems.
Additionally, the team experimented with AI to improve pediatric medication management.
"The challenge in pediatrics is that our medication alerts are often too broad, and clinicians become desensitized to them," Dr. Herstek said. "With AI, we could develop more targeted alerts that make sense for individual patients, reducing alert fatigue and improving safety."
By integrating generative AI, Children's National would be able to create personalized, context-sensitive alerts based on patient history, medications, and specific needs.
Next steps
Children's National plans to refine the four prototypes created during the workshop with Microsoft.
"We're still in the very early stages of this, but we hope to start testing these prototypes in clinical settings within the next few months," Dr. Herstek said. "Once we gather more data, we can refine the tools and roll them out more broadly."