You note that, "What they need is the same level of sick leave time that many other workers in the…
Michael McNeal takes messaging seriously. Maybe you think of messaging as the mundane transmission of old fashioned analog signals from medical equipment, the boring stuff that only a screwdriver-wielding engineer could love. Michael and his company Emergin see messaging as strategic and mission-critical for hospitals. As a result, they’ve incorporated some cutting edge architecture and deployment ideas into their services for hospitals. It’s another formerly separate discipline that’s starting to look a lot like IT.
Hospitals buy a wide variety of message-capable equipment, often with little planning. That’s like cherry-picking favorite individual department IT systems with proprietary interfaces, with little regard to workflow changes, integration, impact on patient care, and optimization. That’s what Emergin addresses: architecture, deployment planning, and tying event notification and alarm management into strategic goals. Quite successfully, as it turns out, as the company is growing at a torrid pace after pretty much defining their own market segment.
HIStalk reader Art Vandelay mentioned Emergin in an article comment right after HIMSS, which is what put them on my radar in the first place. His description of their market is a lot better than I’d come up with, so here’s what Art said: “Emergin has a dynamic CEO and a great vision. If they’d only build a better biomedical data integration mousetrap – a la CapsuleTechnologie, they could set the world on-fire. Everyone in IT, the tsunami on the horizon is integration with biomedical devices and intelligent building technology (as opposed to letting these technologies persist on their own, poorly secured and sometimes poorly supported and utilized, with a few bridge-points spanning networks). It is time to hop into the boats and meet the wave rather than get over-run. We all want discrete data in our medical records systems, these devices generate more than you can imagine and it is a huge win with people who are in your hospitals more than the physicians. Emergin has a well-thought-out profile / use case for biomedical alert integration that should be in every mobile telecommunication and device request for proposal you put out. They also share their information in a brilliant guerilla marketing technique to entrench their product in the industry.”
Good intro, Art. Thanks for Michael for taking the time to chat and to Kathy McCall of Emergin for making the arrangements.
Tell me about the company and yourself.
Emergin was founded in 1995. The name derives from emergency and innovation. Since inception, our focus has been on emergency, real-time information as well as innovation, breaking new ground in new markets with a heavy emphasis on healthcare.
After graduating from the University of Florida as a computer scientist, I worked at IBM on operating system design. Above getting a comprehensive OS background, it helped me recognize the importance of governance structures in organizations, working on a team of up to 4,000 people architecting components into a single OS.
Hospital alarms used to be hard-wired, proprietary, standalone, and connected only to a desktop system or audio alert. How has that changed?
It really hasn’t changed. We’re hopeful that Emergin will be the driver of that change in healthcare. Most hospitals still buy monolithic, discrete systems. Phase II or III is workflow and optimization. We’re trying to get hospitals to think more strategically about planning and design. Healthcare hasn’t really changed that frequently.
We flew around the country over three years to really educate the hospitals on SOA and IT architecture in general, how they can be thinking differently about controlling their own destiny, focusing on work process design, human factors, and performance management.
We’re working strategically with customers, thought leaders like Texas Children’s, University of Chicago, Cedars Sinai, Memorial Sloan-Kettering, William Beaumont, New York Presbyterian, and Boston Medical Center. We need to work with them because it’s not one size fits all in the early innovation stages. We’re looking at alarm patterns, message discrimination, and improved patient safety.
The only problem is that there’s no category for Emergin [laughs]. We’re creating a market around of a category of alarm management that doesn’t exist in the US market.
The company has promoted and sold into a number of industries. Manufacturing, with Six Sigma and Lean. Aviation and the FAA. Government agencies like E911 for broadcasting of critical events. It’s the same business models being applied to healthcare market.
We researched the market six years when we decided to focus on healthcare. We realized that there were many discrete systems but nothing to orchestrate them all. Add on top of that the ability to create governance structures or committees or how to orchestrate nursing, IT, biomedical engineering.
How do you pull it all together into a master technology plan? You start out with a clean sheet of paper, a governance structure such as committees, and then build an architecture. Get representation from all departments to get everyone on the same page. Then, build around the hospital’s key strategies, like patient safety or financial performance. When departments and budgets are decentralized, departments choose their own technologies, which create fiefdoms. Then it becomes a battle to see who has the most power.
We’re taking it to next level, building IT architecture, mapping traditional alarm systems into medical architecture with a focus on workflow and performance management. I just received the initial alarm reports from Texas Children’s today. We’re assembling them to look at the rate of alarms on physiologic monitoring alarms, nurse call systems, IV pumps, and how effectively that info is being communicated to the nursing staff.
Do you have an example of what you do?
A example is William Beaumont. We visited their facility, then did an inventory of systems. We asked questions of IT about how they perform certain workflows. We put together a bunch of questions to ask care units to determine whether there was consistency in workflows. Often you’ll find that IT organization will claim that workflows are consistent until they do the research. We designed tools for them to collect the information and then give us the evidence of what’s happening. Hospitals also do research of individual systems, common nomenclature for bed labels, or staff classification or alarm categories.
What that uncovers is that they don’t really have a plan to drive the standards. They’re governed individually. Emergin created templates and tools to set standards and architecture to drive conformance over time. It’s easiest to go into new construction or renovation, but in many cases hospitals are retrofitting existing systems. When they realize how painful that is when there aren’t standards, they’re more apt to do it at the beginning of any new deployment.
Much of that isn’t tools, it’s a workshop. We meet for half a day with nursing and talk about workflows, care models, and response to alarms. Then, half a day with IT for education on service-oriented architecture An issue that IT has is that they’re typically responding to capital equipment already purchased. Someone’s already bought physiologic monitoring and needs to get IT involved. When IT isn’t involved upfront, workflow changes aren’t uncovered until implementation. They may need network changes, interfaces, or workflow changes.
We’ve gone through 1,000 deployments with 200 best-in-class systems. Emergin has learned in six years how to do technology lifecycle management and change management. We understand that when you’re planning or designing your architecture and third party systems, you have lots of things to consider. Most hospitals won’t consider process flows from the beginning.
Maybe you’re sending an RFP for a nurse call system or IV pump system. Each monolithic system is being deployed. There’s no real consideration of workflow until afterward. All of a sudden, nurses are carrying four or five devices or entering the same data into three or four systems. That creates workflow challenges for nursing and frustrates them because they want to spend time at the bedside and technology makes it more difficult.
The next phase of electronic medical records seems to be the incorporation of digital device information. How does that fit in with your focus on alarms and events?
Most of the technologies are used to import information from data from medical equipment into EMRs for charting. Emergin’s value proposition is around messaging and notification, within five seconds for an emergency. Our core technology is the Emergin Enterprise Service Bus. For the mobile worker, which is 90% of caregivers, getting information is critical for patient care and safety. The EMR will collect info to record it as part of the patient record, but it’s traditionally a data repository for that information. Some EMRs can trigger rules or events, but not usually real time in five to 10 seconds.
What do you think of the work of IHE?
They’ve just created a medical device subcommittee of IHE. Some of our customers are part of it. They’re trying to drive some standards around how medical devices interface alarms and events. We were invited last week to help work on those standards.
IHE started on cardiology, pharmacy, and lab and is now expanding into the medical device domain, which we consider our core competence. Because we’ve integrated a lot of the medical equipment, we’ve created a standard integrated profile that we’d like to make publicly available. Some customers are including those specs in their RFPs. We’ve found that interoperability, not to overuse the term, is normally not considered until Phase II, III, or IV of a deployment cycle. We’re trying to coach hospitals to include it in the initial acquisition. Maybe not in the Phase I deployment, but interoperable by design.
Failure to rescue is an often overlooked phenomenon.
That’s the core area we’re focused on. We’ve done tremendous amounts of FMEA studies and are involved in clinical trials with medical equipment manufacturers. We’ve found that failure to rescue is often because of communication errors. We’ve built the ability to trace back on patient activity when the alarm was generated, who it was assigned to, which system, which device, who read it, who acknowledged it, and when the caregiver responded. By integrating with systems, we can produce an RCA transcript that can re-create life cycle of a patient alarm.
In the years of workshops, we were often asked about how long does it takes to do an RCA. It’s six to eight weeks with a reliability rate of 40%. 60% of the data was anecdotal. Hospitals couldn’t recreate the incident to avoid having it happen again.
I co-presented with Darren Dworkin, CIO of Cedars-Sinai, at HIMSS. A network engineer stood up in the audience and said he didn’t really understand the clinical domain, but said if he couldn’t find network problem or security breach within 15 minutes, he’d be out of a job. It takes six to eight weeks to find the cause of a patient’s death? That’s a great question. It’s a very interesting analogy.
Is alert fatigue and false alarming a problem?
Part of the Texas Children’s study is measuring the amount of false alarms. A key part isn’t just technology or relaying of an alarm, but looking at its impact on vigilance, what the fatigue lines are.
That’s another common workshop question: what’s the fatigue line for a 12-hour shift? No one could answer. Texas Children’s is writing an AHRQ grant for further research over three years. They’ll study the rate of alarms on a per-care-unit, per-patient basis. How many alarms can be physically handled during a shift? What’s considered an adequate staffing level to provide care?
Texas Children’s is also studying by system and by vendor and looking at trends. They recognize that, in a population of 12 patients, two patients generate over 80% of alarms. When you do staffing levels and look at what causes delays in response, which is unmeasurable in most hospitals, what is their average response time per patient? If they have two simultaneous alarms, what impact did that have on the second patient? You have to measure failure to rescue to determine response.
Hospitals struggle with throughput. What systems can improve that?
A Stanford graduate wrote about complex event processing. How do you know when certain things hit thresholds, like ED at capacity or census at capacity, and then do just-in-time management? Emergin is studying this as well. We started out with life-critical alarms — V-tach, asystole, or apnea — and expanded beyond that to look at other critical events, such as critical lab results, getting them to caregiver, and having them call back receipt. As you’re looking at bed turnover and dirty beds, how do you streamline the overall operation to effectively respond?
When Emergin looks at an enterprise, we look at any input event and output event as discrete units, whether from physiologic monitors, nurse call or lab system, EMR rule, ED capacity alarm, or RFID alert if an infant leaves a certain zone, Emergin can take that data from discrete systems and take action on it, delivering it to whatever device is appropriate, such as a Cisco or Spectralink phone or Vocera badge. Some departments use LED signs in the hallways for real-time status updates. Doctors may have a RIM Blackberry, Treo, or cell phone that is their preferred device. How do you orchestrate these devices that carry events?
What are some creative uses of alarm systems in hospitals?
The pneumatic tube. If sample is sent and the tube gets stuck, it alerts the biomed department. Or, the Pyxis machine hits a threshold number of meds that need filled and only two pharmacists are working. We can alert the director of pharmacy to pull in more pharmacists to make sure they meet SLAs. Or the blood bank. If the temperature goes above or below thresholds, then notify maintenance. Customers generated these ideas.
How important is service-oriented architecture to your products?
We were hired by Motorola in 1996 to become the architect of their next generation platform called Heir Apparent. We’ve been SOA-driven since Day 1, a core messaging and notification engine with service adapters that became the interfaces to third party products. As the platform evolved, it enabled Emergin to continuously evolve and add more systems to its architecture without changing the foundational architecture over a 12-year period. We went from zero integrations to over 200 because we build a service model from Day 1.
There’s a lot of hype talking about SOA, but you’ll find that most companies that are doing it are in middleware, traditionally in logistics, financial services, and manufacturing. It hasn’t been applied well in healthcare. To build true SOA is to take a step back, study the inventory of systems acquired in a hospital over time, then build reference architecture, plugging systems into it, then looking for ways to increase operational efficiency.
SOA is a foreign concept to most hospitals. We’d ask, “Who’s the most important person when you build a house?” The joke answer was always the wife or general contractor. Often, no one thinks about the architect as the most important person, even though they build the plan and design. That ensures that when you build a bedroom or bathroom or kitchen that you’re using the same plumbing and electricity for economy of scale.
Think about the blueprint from Day 1. If hospitals step back and build an architecture, figure out how systems fit into it to ensure that you’re not buying the same components over and over. You’re able to leverage some of the systems that you’ve already purchased. When you have a core engine like an enterprise service bus, there’s certain user functionality that should be part of the core engine and not the individual systems you attach to it or you’ll duplicate job functions.
What involvement should CIOs have in alarm systems?
When we first kicked off the workshops, we thought our target audience was the CIO. We found that getting a CIO to spend an entire day to learn about SOA won’t happen. Later, we changed it to IT directors, biomed directors, and nurses moving into informatics. We fund that the IT director was the right audience, the right-hand technical gurus to the CIO. The CIO is more business-minded, with a holistic, vanilla operational focus. They don’t need to know details, they just need a team who knows details.
We found that when training directors of IT, they can translate language to the CIO and demonstrate the importance. We thought the CIO would be the strategic eyesight of the domain and would have to pull it all together, but we found that you have to speak different languages. Nurses don’t understand techno talk, but when you talk about care models and workflow and patient complaints, that’s what they related to. With IT or IS, we talk about networks and servers and databases and how systems interconnect. For biomed, it’s medical equipment, maintenance, and how they’re deployed and used. For facilities, it’s fire alarms, security, blood banks, and pneumatic tubes.
How do you get all those audiences with domain expertise on the same sheet of music? The workshop accomplished getting the group together for a common vision. SOA explains everyone’s responsibility for the architecture as it relates to their job functions.
Emergin was just named to the Healthcare Informatics 100 and one of Deloitte’s fastest growing technology companies. What are your goals going forward?
To continue accelerating our growth rate. We’ve had 70% to 75% for past three or four years and we’re still shooting for a triple-digit growth rate. We will probably hit closer to 80% this year. Those are ambitious goals. We’ve worked hard to build a market. We focus on customers. Doing the right thing has really helped us not only sustain the company, but build a solid brand recognition to set us up for phenomenal growth in years to come.
Most companies are happy with a 30 or 35% market share as a market leaders. We’re taking the Cisco approach. We don’t just want to be market leader, we want to set the standards.
As far as competition, there’s not much out there. We’re finally working with certain reporting organizations that are creating alarm and event management as a category for the first time. Joint Commission had written the effectiveness of alarm systems into one of their National Patient Safety Goals in 2006. Because there weren’t any vendors that could achieve those goals for hospitals, they removed it. We would suspect that after publicity from Texas Children’s and Beaumont, it could be written back into the charter and will fuel demand.
Your a technical guy. Are the skills you need to run a fast-growing company a lot different?
I’ve learned quite a bit on the job and I’m an avid reader. I’ve read about great companies, business leaders, strategies, leadership, marketing, and sales.
I’ve taken an interesting path from computer scientist to product manager, looking at business models and market segments for what we want to build, then putting the pieces together. My passion is getting more intimate with customers and leading sales and marketing initiatives for the company. It’s a unique transition for someone technical to grow into the business side.
Through the transition, I’ve recognized that hiring the right leaders, administration, and the engineering and services group really freed up my time to focus on sales and marketing. It allows me to spend every single week with customers.
When you look at personality types and skills, I’m a visual mathematician. I can take complex patterns and figure out formulas.
With the complexity of healthcare and the vendors in the ecosystem, we can help hospitals put a strategy together. The more you listen to customers and have them drive your roadmaps, the more successful you’ll be as vendor.