17-01 Segment 1: Reengineering the ICU

All for the treatment of the patient


With monitors surveying every part of patients’ bodies, hospital intensive care units appear to be a model of high tech. But systems engineers say ICU’s are actually models of inefficiency because few of those high tech devices talk to each other. Experts discuss how ICU’s could be improved to save lives.

Stay in the loop! Follow us on Twitter and like us on Facebook!

Subscribe and review on iTunes!



  • Dr. Peter Pronovost, Senior Vice President for Patient Safety and Quality, Johns Hopkins Medicine and Director, Armstrong Institute for Patient Safety and Quality
  • Dr. Brian Pickering, intensive care anesthesiologist, Mayo Clinic, Rochester, MN

Links for more information:

ICU Inefficiency

Reed Pence: The intensive care unit for a hospital is full of distractions, it’s noisy and action packed – no place to be if what a patient needs most is rest. Lives are definitely saved there but a surprisingly high proportion of ICU patients later suffer from PTSD as a result of their recovery experience, and all of the disruptive technology with lights flashing and alarms beeping doesn’t have as much of an upside as you’d think.

Dr. Peter Pronovost: To walk into an ICU – it’s almost overwhelming, the apparent technology, the alarms beeping and data flowing but the reality is, the ICU’s today don’t look any different that they did 80 years ago. They weren’t designed to achieve a purpose, they don’t connect and talk to each other so as a result, they’re chock full of new technologies but they’re all isolated – none of them can connect with each other and each monitor tries to make their alarm the most annoying even if it’s the least important. So our nurses answer a false alarm every 90 seconds. Now that’s no wonder why we spend enormously on healthcare and have negative productivity; because all that is, is waste.

Reed: Dr. Peter Pronovost is Senior Vice President for Patience Safety and Quality at Johns Hopkins Medicine and Director of the Armstrong Institute for Patient Safety and Quality.

Dr. Pronovost: It’s not just a productivity issue, it’s a safety issue is the far bigger concern. You see, when you have an alarm every 90 seconds and most of them are false, you just predictably become desensitized to it – you don’t need to go in, you’re drawn often to the least important alarm so it distracts you and you might miss a more critical alarm or some other task that you need to do that’s critically important. In no other industry do we do it that way. What we do in every other industry is we would have an infusion pump or a ventilator just be a dumb pump – it gives you a breath and the smarts would be centralized. Right now we spend an awful lot of money for a whole lot of redundant software and alarms that are on each individual monitor rather than just having them connected on one control board like you would in a cockpit, to say “ok what are we trying to do here and are we meeting our goals?”

Dr. Brian Pickering: The issue of alert fatigue and informational overload is one that increasingly being recognized as contributing potentially to errors and how we look out of patients. So, a strategy for managing informational overload and a strategy for managing alarm fatigue, I think is a critical piece of the environment that we need to look at.

Reed: That’s Dr. Brian Pickering and intensive care anesthesiologist at the Mayo Clinic in Rochester Minnesota. He says the issue of false alarms has its roots in the best intentions.

Dr. Pickering: When you think of the voice of manufacturers requirements, they don’t want to miss adverse events they don’t want to miss someone’s heart rate becoming high or somebody’s saturation becoming very low, so they make their devices very sensitive for picking up those really life threatening problems. The downside of sensitive devices is that they’re not very specific. Sometimes they pick up false alarms and they alert them in a loud and irritating annoying ways. So there are many studies out there that show that the rate of true-false alert is unacceptable. So we have way too much noise in the system and not enough signal.

Reed: A few of the problems with monitors and the fixes those problems suggest are incredibly simple.

Dr. Pronovost: What we found was that there was complete variation in what alarms people set. Many people didn’t even know what parameters they set the alarms on for. In some case the alarms were just turned off because they were so distracting and annoying that they weren’t providing any good information which loses all the protection that you have from them and the message is – not to be slaves to these alarms but leads to design something that serves the needs of the patients and clinicians.

Reed: But it’s far from just monitors that are problematic in Intensive Care Units, they’re just the noisiest part of it. Pronovost has plenty more examples.

Dr. Pronovost: They way we check to make sure the dose is right for pain medicines and about a dozen other medicines is one nurse grabs another nurse and they go check the infusion pump. All that data is electronic, why can’t we make an electronic double check where the medical record connects to the infusion pump – it would be safer and much more productive. I’ll give you some other examples of waste; our clinicians spend over half their time documenting medical record. Think about that for a second – that adds no value other than checking some boxes you could submit, it doesn’t help patients get well but it’s how doctors and nurses are spending half their time now with these new systems. It’s because they were never designed to meet the needs of the patients and families and clinicians. They were designed largely absent of significant clinical input.

Dr. Pickering: I’ve never thought of ICU’s as being efficient environment, I do think they’re very technological in nature – and sometimes very inhuman in nature. But what all of those monitors do is obviously keep a very close eye on what’s happening to their physiology and other components of the patient. What they don’t do very well is talk to one another and give the clinicians and the family even this holistic view of what’s really going on with the patient.

Reed: Dr. Pronovost says that efficiency experts agree – ICU’s are a mess.

Dr. Pronovost: We started partners with some system engineers – these kind of guys who do the Pluto Mission, who build submarines, who build satellites – they’re at the Johns Hopkins applied physics lab that’s part of John Hopkins that they do work for the Department of Defense and started rounding with us to sort of see how could be design an integrated ICU just like they would design an integrated submarine. And when they walked through, they were appalled – their statements were, “This is the most grossly under engineered thing I’ve ever seen in my life. Why on earth would nurses take data out of one computer system, write it down and then type it into another? Can’t you make them talk to each other?”

Reed: Instead, each monitor spits its own information creating tens of thousands of data points. There may be so much data that Pickering says doctors can’t use it – they’re overwhelmed instead.

Dr. Pickering: As human beings we’re good at picking up patterns in the environment, but when we’ve got so much data, pattern recognition becomes difficult and sometimes, you know, the basis upon which we make our diagnostic and treatment decision and if we have too much data – we can’t recognize the pattern with all of the noise. We embarked on the wrong pathway and in can cause diagnostic delay or diagnostic error or treatment delay or treatment error associated with high mortality and morbidity.

Reed: Pickerings team at the Mayo clinic spent 3 years building a system he calls “AWARE” that whittles down the thousands of data points to a few hundred – the ones that are important.

Dr. Pickering: We integrated into a database and we run multiple rules, about a thousand rules, and all of that data and the output of that analysis is presented in a dashboard. What we’ve demonstrated when you pull data together like this is that the cognitive load or the difficulty of decision making is dramatically reduced. The quality of the decisions are increased and that all comes out at an efficiency level that is twice what we see with a disorganized environment. But the impact of better decision making is that treatment are delivered more quickly and patient outcomes improve and they spent less time in the Intensive Care Units and less time in the hospital. And all of the consequence of simply managing data that was better and having processes of care that can be tracked and followed really closely.

Reed: Similarly, Pronovosts’ team at Johns Hopkins too a systems approach to tasks done routinely in the ICU.

Pronovost: We had done some work in making checklists and preventing catheter infections by making sure clinicians rigorously adhered to best practices as demonstrated by this checklist. Well, the infections that we were trying to prevent is just one type of harm – patients in the ICU are at risk for maybe a dozen harms, maybe 20 harms, every one of those harms has a check list. Every checklist may have 5 or 10 items to it and every one of those items may need to be done 3 or 4 times a day. So you add it up and you’re expected, largely from memory, 150 things every day to keep them safe and right now not a single electronic medical record that you can buy, despite spending an awful lot of money for them, gives any visual display if you’re doing those lifesaving 150 things. So you get caught up in screens and clicks and trying to find information that’s mostly geared towards billing but what’s lost is what’s most important. That is, “am I giving patients these life saving things?”

Reed: Pronovost and his team created a master check list app that he says makes it 40-50% most likely that patients get every therapy they’re supposed to and he says that kind of improvement could be done throughout systems in the ICU.
Dr. Pronovost: Our engineers just walking through said, you know, “Peter there’s so much waste here, we think we can get 45% productivity gains by designing systems to meet patients needs an radically improve patients safety.”

Reed: 40%? Pronovost says just a 4% improvement in productivity would go most of the way towards solving medicines cost problems. So why hasn’t that already happened? Well part of it is the medical culture – doctors have only started to look at their world like an engineer would. Device workers have followed in the same way.

Dr. Pronovost: The business models have evolved around each manufacturer making their own proprietary widget and not talking to each other so you have to buy their widget to get that alarm and we’ve not seen a system integrator. Think like a Boeing or a Lockheed-Martin come into the market and I think the reason for that is, they can’t do it alone. So if you’re making a new plane – that system flying the plane is controlled enough that Boeing of Lockheed could, and do, hire the pilot, they get a test pilot they build it, they go fly it, they watch what the pilot needs, they change it, and then they make it commercial. Well in healthcare, no company is going to build a mock ICU that puts patients in it or build a mock hospital – it would be too expensive and too risky. But what they need to do, what we haven’t done yet, is partner between – in this case academic health systems with engineering companies – that could really bring their skills together to design the kind of technology that will help heal patients.

Reed: Pronovost says that finally, industry and healthcare providers are starting to team up to integrate the ICU and there’s intense interest on both sides.

Dr. Pronovost: The time has come for this idea, because ultimately what we’d like is, what we can do for the ICU is what Steve Jobs did with the iPhone – that is link hardware, software and content. Right now, you can’t buy an integrated ICU in the world, even if you wanted to. What we should be able to do is be able to integrate these technologies and alarms and protocols so that you can get care anywhere in the world that’s equivalent to the worlds best ICU care. You can be on a battleship out in the Persian Gulf, you can be in a rural area, you can be in a developing country and just like you can get the same music on your iTunes, you should be able to download the same protocols and the same information systems that we could at Johns Hopkins or at other places. And that vision is completely doable.

Reed: You can find out more about all of our guests on our website, RadioHealthJournal.net. You can also find archives of our segments as well as on iTunes and Stitcher. I’m Reed Pence.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s