Originally published on Healthcare IT News
The health system achieved 100% adoption of an AI platform that consolidates multiple sources of clinical information into one easy-to-access mobile app. Clinician satisfaction with data access increased from 22% to 84%
At MetroHealth in Cleveland, a renal patient didn't receive dialysis before surgery, even though guidelines existed. The problem was critical protocols often were buried in a maze of portals, folders and emails – making them hard to find when they were needed most.
This near disaster made it clear to Dr. Luis Tollinche, department chair of anesthesiology, that the process needed a complete overhaul.
THE CHALLENGE
The anesthesiology department was facing a serious information overload problem that threatened both patient safety and clinician well-being. Each day, the team received around 12 different communications.
These included protocol updates, revised guidelines, institutional policies and safety alerts. The messages were scattered across six different systems, including email, SharePoint, policy databases, Epic messages, printed protocols and text messages.
"Clinicians often had only minutes, sometimes seconds, to locate critical information during complex procedures," Tollinche explained. "Research indicates that if the required information cannot be found within two minutes, people tend to rely on memory or intuition, which can increase the risk of errors.
"Our department experienced this when a patient suffered complications because a protocol was not followed," he continued. "The protocol existed, but it was lost in the fragmented communication systems. This incident highlighted that having excellent guidelines alone was not enough. If clinicians could not access the information quickly, adherence suffered and patient safety was compromised."
PROPOSAL
An AI platform from vendor C8 Health was proposed as a single point of access for all clinical knowledge. It promised to deliver the right information to the right clinician at the right time. The system was designed to understand context, including the clinician's role, the cases scheduled and the procedure type.
"The idea was to surface relevant protocols proactively, rather than requiring clinicians to hunt through multiple systems," Tollinche noted. "Unlike generic search tools, the AI platform was built specifically for healthcare. It included safeguards to ensure that responses came only from the institution's approved knowledge base.
"The platform could ingest information from all existing sources, convert it into a searchable, mobile-friendly format, and respond to plain-English queries," he added. "It was designed to integrate seamlessly with the EHR and other hospital systems without requiring major IT infrastructure changes."
The AI also was intended to learn from usage patterns, automatically prioritizing the most relevant information for specific procedures. The goal was to reduce reliance on memory, improve protocol adherence, and allow clinicians to focus on patient care rather than searching for information.
MEETING THE CHALLENGE
Implementation went smoothly, with the vendor migrating more than 750 knowledge items from SharePoint, policy databases and other repositories into the AI platform within 24 hours. The system was integrated with the EHR, allowing clinicians to access protocols directly from patient charts. It also pulled schedule information to proactively surface relevant protocols for upcoming cases.
"Every member of our anesthesia department uses the platform," Tollinche reported. "Residents rely on it heavily during night shifts. Attending physicians consult it for less common procedures or recent protocol updates. All clinicians review case-specific protocols on personal smartphones while moving between rooms.
"The AI responds to natural language questions, such as, 'What is our protocol for managing hyperkalemia in renal patients?' and provides instant, accurate answers sourced exclusively from approved guidelines," he continued.
Engagement metrics show which protocols are accessed most frequently and highlight knowledge gaps.
"Most important, 89% of interactions occur on personal smartphones, confirming that mobile access is essential," he said. "The combination of AI, mobile accessibility and integration with our EHR ensures that clinicians receive the information they need quickly, improving decision making and patient safety."
RESULTS
Within five months, MetroHealth achieved 100% adoption across the department. Clinicians average 3.8 sessions per day – which is remarkable for healthcare technology, where 30-40% adoption is typically considered a success, Tollinche said. This metric demonstrates the platform is actively used and integrated into daily workflow, he added.
"Clinician satisfaction with accessing clinical resources increased from 22% to 84% reporting satisfied or very satisfied," he noted. "The percentage of clinicians finding it very easy to locate information rose from 15% to 73%.
"These results were measured using standardized pre- and post-implementation surveys of the same clinician cohort," he reported. "The AI's ability to deliver context-specific information quickly is the main driver behind these improvements."
Protocol adherence also improved significantly. Compliance with enhanced recovery after surgery protocols increased from approximately 65% to over 88%. Critical protocols, such as pre-operative potassium testing for renal patients, are now consistently followed because the system proactively surfaces them when relevant.
"These results show the AI platform directly addresses knowledge gaps and prevents the types of errors that previously compromised patient safety," Tollinche said.
ADVICE FOR OTHERS
Healthcare organizations should start by conducting an honest audit of their current information environment, Tollinche advised.
"Critical clinical knowledge is often scattered across far more systems than initially assumed, and clinicians may make decisions based on incomplete or outdated information," he said. "Understanding exactly where information lives and how it is accessed is essential.
"Mobile-first design is critical," he added. "Platforms that require desktop access or complex login procedures will not succeed in clinical environments, where professionals need instant access while moving between locations or working in sterile settings."
Seamless integration with existing workflows and EHR systems is essential, and large IT infrastructure changes should be avoided if possible, he noted.
"Finally, focus on measurable outcomes rather than features," he said. "Pilot programs in a single department allow organizations to track adoption, engagement and impact on protocol adherence before scaling.
"Demonstrating clear, quantifiable benefits to clinicians is essential, because healthcare professionals are often skeptical of technology that promises transformation but delivers frustration," he concluded. "Showing tangible improvements in workflow and patient safety ensures adoption and long-term success."