CalmWave’s Hospital Safety and Operations Platform unifies disparate signals

How CalmWave Solved ICU Alarm Fatigue by Treating Healthcare Like Enterprise IT Operations

Patients in intensive care units generate up to 1,600 alarms per day. Every 30 seconds, a monitor beeps while someone desperately needs rest to heal. Device manufacturers have attempted solutions for decades. All failed.

In a recent episode of BUILDERS, Ophir Ronen, CEO of CalmWave, explained why—and how his team finally solved it by recognizing that healthcare’s alarm problem follows the exact same pattern as IT operations alert fatigue.

The Integration Problem Disguised as a Workflow Problem

ICU monitors track vital signs with alarm thresholds for each parameter. Default settings might trigger alarms for heart rates below 50 or above 120 beats per minute. The problem isn’t the devices—it’s that clinicians receive no data-driven guidance for adjusting these thresholds.

“The problem is because of the fact that there’s no guidance given as to what those limits should be set to,” Ophir explained. “Nurses make changes based on their intuition. And sometimes intuition is wrong, or you’ve had a really hard shift.”

Device manufacturers see only vital signs. EMR systems see only medications and interventions. Neither possesses sufficient data to safely recommend threshold adjustments.

“You can’t go in and say safely, hey, you should make this change, but not know what medications have been given to that patient that potentially could be impacting the vital sign that you’re asking them to change the alarm limits to,” Ophir said.

This is fundamentally a data fusion problem. No one solved it because solving it required building the integration layer first—not the application.

Why Middleware Interoperability Became the Moat

Ophir’s background includes selling his previous company to PagerDuty and taking Internap public in 1999 for $16 billion as an early commercial internet backbone. His enterprise IT operations experience revealed what healthcare companies miss: interoperability isn’t a feature you add later.

“In our context, interoperability is not a feature, it’s a prerequisite for safety,” Ophir said.

CalmWave built bidirectional integration with both Philips InteliBridge (high-frequency vitals and alarms) and Epic EMR (medications, interventions, labs) before addressing the clinical problem. The vital signs data represents 10x the volume of EMR data.

Everything flows into their common signal format—structured specifically for data science. “Anything that passes our integration layer gets normalized,” Ophir explained. “So our data scientists are in heaven because they don’t have to deal with all the crap data. It’s all pure by design, generated by the platform itself.”

One hospital generates 7 million data points daily. Their largest system now processes 32 million data points per day with 7 billion in storage—a 10 billion annual run rate from a single 14-hospital system.

The integration layer itself became defensible IP. Ophir noted that high-frequency vitals data is “erased on a rolling 30-day basis” at most hospitals. By fusing and retaining this data with EMR context, CalmWave created a genuinely novel dataset that enables algorithms no one else can build.

Math and Statistics Beat Black Box AI

CalmWave’s market entry came through an innovation investment arm at a major health system. After three months of due diligence, they presented to 30 senior leaders including the CMO, CMIO, CNO, and their staff.

The team had been warned about one leader: “Wicked smart, but also Dr. No,” Ophir recalled. “So if he didn’t like what you were saying, then that’s it.”

CalmWave walked through their algorithms—the actual mathematical basis for calculating safe alarm threshold adjustments. Nine minutes in, Dr. No stood up.

“My heart’s dropping in my chest at this point,” Ophir said.

Then: “You guys shouldn’t even call yourselves AI. This is math and statistics. I understand exactly what you’re doing. Well done. This is truly innovative.”

Two weeks later, they signed a comprehensive system-wide agreement.

This validates a specific positioning approach: in risk-averse environments where liability falls on the end user, explainability needs to be architectural, not cosmetic.

“Everybody likes talking about AI, but nobody wants black box AI,” Ophir explained. “So that’s what we have, what we describe as transparent AI, where you as the clinician can understand why something is happening and why this recommendation is being provided.”

Clinicians bear personal liability for decisions. “It’s not the system that’s going to be blamed if there’s a fault or a problem, it’s you,” Ophir said. Building trust required exposing the complete statistical reasoning behind each recommendation—not accuracy metrics, but the actual math.

The Innovation Arm Pathway

CalmWave’s first major deal bypassed traditional hospital procurement entirely. A gentleman from an innovation investment arm reached out after seeing their work on alarm fatigue. Three months of due diligence followed, then direct placement in front of clinicians and C-suite leadership.

This pathway matters for founders targeting large health systems with IT budgets in the hundreds of millions. Innovation arms are measured on finding novel solutions, not minimizing vendor risk. They act as internal champions, pre-validating startups before exposing them to decision-makers.

The quid pro quo: the problem must be urgent enough to justify the organizational friction of working with an unproven vendor. Alarm fatigue qualified because it carries regulatory scrutiny, patient safety implications, and direct links to nursing burnout—consequences visible at the executive level.

Pattern Recognition as Competitive Advantage

The alarm fatigue solution emerged from recognizing structural similarity to a solved problem. At PagerDuty, Ophir’s team built platforms for processing massive streaming data, correlating events, and reducing alert noise in IT operations.

ICU alarm fatigue follows an identical pattern: too many alerts without context, forcing humans to mentally correlate disparate signals under pressure.

CalmWave applied the proven architecture: signals generate alarms, alarms cluster into incidents, incidents group into events. “The definition of alarm fatigue is anytime you see a one-to-one association between alarms and alerts,” Ophir explained. “There’s always more alarms than there should be alerts.”

This pattern recognition created non-obvious competitive advantages. Healthcare incumbents lacked the specific systems thinking required to process 7 million daily data points from a single hospital. CalmWave’s team already knew how to build that infrastructure at scale.

“We as CalmWave are uniquely suited for this because of our enterprise IT background,” Ophir said. “We’re used to enormous data sets. We’re used to big SaaS platforms.”

From Beachhead to Adjacent Domains

Medical device manufacturers are now pursuing distribution partnerships. The alignment is natural: their devices generate the alarms, CalmWave’s platform manages them while building broader operational value.

“That’s the next big inflection point that I think is going to happen next year,” Ophir noted.

The company holds 51 patents with 20 more pending—an aggressive strategy driven by competition with multinational incumbents. “We knew, sort of going into this, that we are going head to head into this space,” Ophir said. “And you know, these folks have a lot of patents and rightfully so. They’ve been building over decades.”

His framework for patent strategy: evaluate who else operates in your domain and how aggressively you need to protect hard-won knowledge. “It’s also very expensive,” he cautioned. “So you got to keep that in mind too.”

Beyond healthcare, Ophir sees the same pattern in energy SCADA systems, defense infrastructure telemetry, and manufacturing PLC sensors. “Any sensor-heavy domain will have the problem of alarm fatigue,” he said.

“Alarm fatigue is the beachhead into any of these domains where in order to solve that foundational problem, we first need to fuse all of the data together, enrich it, structure it, normalize it, and then use that to build more and more value specific to that domain.”

The Extractable Lesson

CalmWave’s success demonstrates that the hardest enterprise problems often require solving infrastructure prerequisites before building applications. Device manufacturers and EMR vendors both attempted alarm fatigue solutions with their existing data sets. Both failed because neither could safely recommend threshold changes without the other’s context.

The integration layer—normalizing and fusing disparate middleware systems in real time—became the actual product innovation. The alarm management application proved it worked.

This mirrors enterprise IT’s evolution. Middleware and interoperability came first, enabling valuable applications built on unified data. Healthcare never had that foundation. CalmWave built it by recognizing the pattern from a different domain.

For B2B founders, the question isn’t whether your market needs better integration. The question is whether integration is a prerequisite for your solution’s safety or efficacy—and if so, whether building that capability first creates defensible differentiation rather than undifferentiated infrastructure work.

This post originally appeared on Front Lines Podcast 1/26/26