
Eric Neil, CIO, UW Medicine

Todd Burstain, MD, Chief Clinical Informatics Officer, UW Medicine
UW Medicine is advancing digital transformation with a focus on AI-powered decision support, strong governance frameworks, and clinician collaboration. CIO Eric Neil and Chief Clinical Informatics Officer Dr. Todd Burstain recently discussed their journey to achieving HIMSS Stage 6, their measured approach to AI adoption, and the ongoing challenge of balancing innovation with standardization.
Scroll down to watch or listen to the full interview; or subscribe to healthsystemCIO on your favorite podcasting channel
Achieving HIMSS Stage 6: A Commitment to Patient Safety
Earning HIMSS Stage 6 recognition is no small feat, requiring organizations to meet rigorous benchmarks in digital maturity and EHR optimization. But for UW Medicine, the achievement was never about the award itself—it was about improving patient outcomes.
“The goal is not just to get an award,” Burstain emphasized. “The metrics were chosen because they reflect organizations that provide better care, and where providers and nurses are more satisfied using the EHR.”
A key component of their success was achieving high compliance rates in two critical areas:
- Barcode Medication Administration (BMA): Nurses scan both the patient and the medication to verify accuracy before administration. HIMSS requires 90% compliance for Stage 6 and 95% for Stage 7. UW Medicine currently maintains a 95-96% compliance rate across its hospitals.
- CPOE: Physicians directly enter orders into the EHR rather than relying on verbal instructions, reducing the risk of errors. UW Medicine boasts a 98-99% CPOE adoption rate.
Reducing Sepsis Mortality with Machine Learning
One of UW Medicine’s most impactful technology initiatives has been its sepsis prediction model, which has reduced sepsis-related mortality by 17%.
Sepsis remains the leading cause of in-hospital deaths in the U.S., but studies show early intervention can significantly lower mortality rates. UW Medicine customized Epic’s sepsis predictive model, training it on their own patient data over six months before deployment. The system continuously analyzes key clinical indicators—such as vital signs, white blood cell counts, and lactate levels—and triggers alerts when a patient is at high risk for sepsis.
“When a patient reaches a certain threshold, a ‘Code Sepsis’ is called,” Burstain explained. “Clinicians are immediately notified, and a standardized order set ensures the right fluids, antibiotics, and labs are administered without delay.”
The impact has been substantial. “A 17% drop in mortality may not sound like a lot, but when you consider how many patients develop sepsis, it means one more patient survives every five days who otherwise wouldn’t have,” Burstain said.
AI and Machine Learning: Cutting Through the Hype
The rise of AI in healthcare has created a mix of excitement and skepticism. While some organizations rush to implement AI solutions, Neil cautions against adopting AI without clear strategic value.
“We don’t do AI just to do AI,” Neil said. “We use it to advance our mission and strategic goals. Otherwise, you’re just throwing money away.”
UW Medicine has taken a measured approach to AI, focusing on projects that provide clear benefits. One promising initiative is ambient documentation technology, which captures provider-patient conversations in real-time and generates structured clinical notes.
“This is one of the most exciting things we’re doing,” Neil said. “If it improves provider efficiency, enhances documentation quality, and reduces pajama time—where doctors finish notes at home—it’s a win.”
Beyond documentation, AI is helping clinicians manage the growing complexity of patient data. Burstain noted that AI can organize and surface relevant information based on a provider’s specialty and clinical context.
“There’s so much data now that it’s easy to miss something,” he said. “AI can present the right information at the right time, ensuring nothing critical is overlooked.”
The Governance Challenge: Balancing Innovation and Standardization
With AI pilots and new technologies emerging at an unprecedented pace, UW Medicine has prioritized strong governance to ensure resources are used effectively.
Neil emphasized the importance of structured decision-making frameworks to guide IT investments. “Decisions on technology investments should be made by the health system’s leadership, not just IT,” he said. “IT is here to facilitate, not dictate.”
UW Medicine has implemented a three-tier governance structure:
- Enterprise Leadership Council: Co-chaired by the Chief Medical Officer and Chief Operating Officer, this group sets overall IT priorities and approves major initiatives.
- Clinical and Business Committees: Ten committees focus on specific domains, such as clinical applications, cybersecurity, and data analytics. Each committee is co-chaired by an IT leader and an operational leader.
- Specialized Councils: Groups focused on department-specific needs, such as radiology or surgery, ensuring granular input before broader decisions are made.
By embedding governance within clinical and operational leadership, UW Medicine helps ensure IT investments align with system-wide priorities.
Avoiding ‘Pilot Fatigue’ and Prioritizing the Right Innovations
With so many AI solutions entering the market, health systems risk pilot fatigue, where too many projects overwhelm staff and dilute resources. UW Medicine has taken steps to mitigate this risk by clearly defining what constitutes a pilot.
“If we’ve already decided to move forward with something, we stop calling it a pilot,” Neil said. “A pilot should be a real-world test to determine feasibility, not just the first phase of implementation.”
Burstain stressed that evaluating potential pilots requires a rigorous selection process. “Before we pilot anything, we do extensive industry research, conduct an RFP, and ensure we’re choosing the right solution,” he said.
Striking a Balance Between EHR Standardization and Innovation
One of the ongoing challenges for health system CIOs is balancing enterprise EHR standardization with the need to adopt new, innovative solutions. UW Medicine follows a general rule: integration trumps interfacing.
“We prefer integrated solutions within our existing platforms rather than standalone tools that require complex interfaces,” Burstain said. “More systems mean more risk, more security concerns, and more maintenance overhead.”
However, Neil acknowledged that sometimes innovation requires going beyond core EHR functionality. “We started this conversation by talking about ambient AI and clinical decision support—both are outside our native EHR,” he said. “In these cases, we carefully evaluate vendors and ensure strong governance controls.”
The key, he noted, is ensuring new tools align with long-term system priorities. “Some AI companies are brand new, and the landscape is evolving fast,” he said. “We don’t want to lock ourselves into something we’ll regret later.”
To-Do List
- Align AI with strategic goals. Avoid AI adoption for its own sake; focus on solutions that enhance patient care, provider efficiency, and operations.
- Governance should be a partnership. Effective governance requires clinical, operational, and IT collaboration, ensuring technology investments align with system-wide priorities.
- Be selective with pilots. Define clear criteria for pilots and avoid launching projects without a structured evaluation process.
- Prioritize integration over interfacing. Whenever possible, use solutions that integrate with existing systems rather than requiring complex, ongoing interfaces.
- Leverage AI to reduce clinician burden. AI should support decision-making, automate administrative tasks, and enhance clinical workflows—not add complexity.
The Future: Adoption Over Implementation
Looking ahead, UW Medicine is shifting its focus from deploying new technologies to ensuring widespread adoption and optimization.
“We built the Ferrari,” Neil said. “Now we need to teach our users how to drive it.”
Podcast: Play in new window | Download (Duration: 44:16 — 30.4MB)
Subscribe: Apple Podcasts | Spotify | Android | Pandora | iHeartRadio | Podchaser | Podcast Index | Email | TuneIn | RSS
Share Your Thoughts
You must be logged in to post a comment.