The colorful, dancing ballerina pig statues on the first level of the Texas Children’s Hospital‘s Abercrombie building provide an odd sense of visual harmony. Their floral tutus add extra shine, which is hard to come by in a building filled with sick – some of them really sick – children.
On a sweltering Tuesday afternoon, Samantha Jacques, Director of Biomedical Engineering at the Texas Children’s Hospital, took me to the Progressive Care Unit, to show me the importance of alarm management.
As soon as we entered the hallway, I could hear numerous beeps in the PCU. Overlapping and loud, I found it difficult to imagine how even seasoned nurses handle them.
But alarms in a hospital are necessary. They tell healthcare professionals about changes in a patient’s diagnostics, such as heart rate and blood sugar, that can end up being life-saving. But the sheer number of bells and whistles makes it hard for the nurses to decide which alarm needs to be given priority.
“The alarm is not what it was meant to be when it was designed,” Jacques said, adding stress for healthcare professionals and patients alike – for instance, patients cannot sleep as well, which can hinder their healing process.
That’s where Houston-based clinical decision support device startup, Medical Informatics, comes in.
Medical Informatics CEO Emma Fauss said alarm management has been a well-documented problem for many years.
“In terms of alarm management, it’s something that has been a problem for many years. It’s been well-documented. There’s things called alarm fatigue. It’s where people are overwhelmed with alarms and become desensitized to those alarms. If you had constant noise in your work environment, it just phases out.
To give you an example, in a 500 bed hospital, you’ll have 70,000 alarms per day, and they have to be addressed.”
Generally, Fauss said, hospitals have policies that require addressing all alarms.
“But you can’t physically address all the alarms,” Fauss said. “When you look at alarm management as a problem, and how you start to address it, it becomes very challenging, because the reason why there are so many alarms, is not one problem. It’s not because of one medical device. It’s a combination of patient mix, the workflow that’s been put in place, the policies that have been put in place, the management of how those alarms are set, the communication between doctors and nurses and expectations on who’s addressing the alarms and who’s setting parameters. It can involve the technology, like what sensors have you chosen to use, are those the sensors that are best for your patient population? There can be other sensors that reduce nuisance alarms.”
Fauss said that of the 70,000 alarms per day, 95% – 99% of them are nuisance alarms.
“That means, non-actionable alarms,” Fauss said. “So when we look at addressing the problems that the nurses are facing on the floor, those are the people who care the most, it impacts their every day, it impacts the local care, it impacts the patients, who can’t sleep because the alarms are going off so much.”
“If I said to a hospital, I can reduce the alarms in your unit by 50% by using our product and going through our product. By working with you – it’s not just installing a technology. Technology only enables decisions that are made. It can only fix a portion. What you really need to be successful is working with a team that understands there’s different elements, and entrusting and targeting those different elements. The nurses have known this is a problem for a long time.”
The reason why the issue of alarm management is being taken care of now is because Joint Commission, which accredits a lot of United States hospitals, has said that the issue will affect accreditation.
“He had just finished his PhD and was hired into medical school,” Fauss said. “He was hired into a lab that was developing predictive algorithms for disease – so looking at a patient’s physiological data, and looking at that mathematically and doing calculations on it to figure out what are the precursor to illness, can we alert physicians beforehand so that they can take appropriate actions to mitigate the risks. He came into this world, and they had carts where they’d roll up and collect 2 hours of data here, 2 hours of data there. They had systems that deleted data.”
As engineers, Fauss said her and Rusin knew the healthcare industry could do better than this, including processing data in Excel.
“So, he helped install a platform that allowed them to collect, store, and process data coming off of the bedside monitors. This was a leap in terms of being able to rapidly prototype these algorithms.”
Rusin started to get interest from different research groups around the country, and people started inviting him and Fauss to talks to teach about what he was doing. At that point, Fauss suggested starting a company to Rusin, and the startup began.
After they did a competition in South Carolina, they caught the attention of the Texas Children’s Hospital here in Houston.
“They were interested in revolutionizing healthcare, and they wanted not only to bring on [Rusin] as a research professor and build a group around developing predictive algorithms, but they wanted to install our system’s pilot.”
Seeing the opportunity that the Texas Medical Center provides, the two decided to bring the company to Houston.
Fauss said that from a technical deployment standpoint, people can overcome technical problems if they throw enough time and resources at them. However, what’s hard to overcome and replicate are understanding the players within a hospital, the politics, the people aspect of business.
“There’s a lot of knowledge we’ve picked up from who we’ve worked with since we started in 2009,” Fauss said. “Everything from understanding how you build predictive algorithms, to how you actually deploy those predictive algorithms. I think that a lot of people get caught up in – in oil and gas, they’ve done alarm management for years, why can’t we just take one of those companies and do it in healthcare? There are so many answers to that question, and the least problem is teht technology, even to the technology itself is very hard and complicated.”
Furthermore, there is a lot of integration involved with Medical Informatics – from integrating with the medical devices, the network of the facilities, a hospital’s biomedical department, HL7 (the standard communication protocol for different internal devices), and a hospital’s authentication and security protocols.
“Every time you interact with one of those groups, you’re dealing with different set of stake holders,” Fauss said.
Fauss said she thinks as a startup matures, there’s a balance of supporting your products versus doing new products.
“Right now, we shave a set of core products, but there’s a lot of pivoting going on,” Fauss said. “We’re finding our place in the market, our voice in the market. And building the company culture is def a process too. At this point, every time we bring somebody onto the team, they add a new personality to the team, they add a new dimension. I think at some point, the company culture will be strong enough where people come into the company and fall into the culture. I think they’ll always add to it, but as we get larger, things change. But still, at 7 people, we are definitely still a startup.”
As Medical Informatics works with the PCU at Texas Children’s Hospital to collect and analyze data, a better experience is becoming a reality for healthcare professionals and patients alike.