So just before I go in to talk a little bit about a vision for the ICU of the future there’s a couple of things I want to address first of all. To start off with a huge thank you to John and EBME Expo for inviting me here, giving me this opportunity, but particularly I’d like to really thank you guys for your confidence in putting on what is to me the first face-to-face conference that we’ve had. That confidence to embrace that challenge of logistics within COVID, you have my gratitude for inviting me to this event. And it also proves that we’re not avatars. People do exist, we’re still real.

The other thing that I’d like to say, John introduced me, I have a critical care background of some 20 plus years in the NHS before I joined Phillips. And when the COVID pandemic became a reality I had a real crisis of confidence, and my conscience was troubling me. What do I do? I work for this healthcare provider, but I’m a fully trained intensive care nurse, do I go back or not? And because of the products that my company sells, all of the ventilators and the monitors, I decided I could make a better impact by staying where I was. But I would like to make a really key point. I’ve had family and friends directly hospitalised and affected by COVID, so you guys on the front line, you have my eternal respect and my eternal gratitude. So thank you very much for all your efforts over the last couple of years, much appreciated.

Anyway enough of the soppy stuff, let’s go on and talk about something a little bit more perhaps interesting. So what I want to talk about is intensive care units are always developing. People talk about them being at the forefront of technology, they’re early adopters, they embrace technology, but there are a number of issues in and around intensive care that have been identified, highlighted over previous years. But COVID-19 and those unique pressures and challenges, the biggest peacetime challenge that our health service has faced, have brought some of these to the fore. So I’m just going to take you through and get you to have a look at them.

So a couple of ways we want to look at it is the environment that the intensive cares are running, just by being an intensive care patients are at risk, irrespective of what they’re in there for. Those iatrogenic risks are always a worry to our patients. The processes that we put in place in intensive care, there’s lots of end of life decisions that have to be made, patient surveillance, not always risk associated, everybody gets the same sort of input. We know unfortunately that due to the pressure that the staff and that system is under, that there are errors and clinical variations that happen across the UK. Some of our outcomes are adversely affected, hospital acquired infection rates are variable, some are climbing. Sepsis is still an unresolved issue with intensive care, it’s a persistent issue. Deterioration even in intensive care is often missed for multiple reasons as we’ll come along to. And staff facing huge challenges from a resource perspective, from an education and training perspective, but also from some of the unique challenges that that technology environment provides for them.

So if we carry on as we are with these challenges that are coming in, and people talk about another pandemic and another pandemic and what will the future bring in terms of that risk to us as a population, if we continue as we are, those working conditions for clinical staff are going to continue to deteriorate, putting them under more and more pressure, and the previous speakers have alluded to that, that impending staff shortage. Value outcome based financing is becoming a model that everyone’s talking about, and they want us to make sure that we’re improving outcomes before we get the resources. And I don’t know what you guys think, there always seem to be a bit of a mismatch there, the moneys comes after you get the outcomes rather than beforehand, so we need to bear that in mind and address that. Acquisition and retention of qualified staff, they are a rarity, you know, highly qualified staff are a real rarity. So some of the solution criteria we need to look at, we’ll look at operational efficiency, those working conditions for staff, those quadruple aims that we want to look at.

So, a hospital of the future, I don’t know if there’s any techies in the audience or trekkies, you might recognise this picture. So this is perhaps a utopic view of a Starship Enterprise type ICU environment. But it’s very clear if you’re reading some of the literature, I mean this quote from the Society of Critical Care Medicine, some years ago actually said we don’t actually need any more technology. We don’t need those breakthroughs to change what we’re doing. What we need to do is look at technologies that have been successfully utilised in other areas, other fields, a more scientific approach, perhaps some more safety focused approach like the aviation industry. Finance models that come from the finance industry, not from the health service. And then the biggest issue that I feel that we’re facing in intensive care today is that availability of data from medical devices, how we capture that data and how we use that, and that’s the biggest barrier to progressing our ITUs forwards.

Key problems in the intensive care that we need to address, so information overload and decision making, the technology that we’ve got in at the moment produces a wealth of data for our clinicians to look at. And some of the studies that are around there now, so 70 ICUs, over 1,300 patients surveyed, and medical errors occurred in 25% plus of those patients. And about 15-16% of those adverse events increased the mortality by threefold. 554 errors, and over 200 serious errors in a single ITU, simple down to information overload. There’s too much information coming through. And when they did a simulation study, there was about 40% of those errors were actually spotted, and the reason they were not spotted was not because people were being negligent, it’s just they had too much information. They couldn’t sift through that data.

Contributing factors, as I’ve said, huge processing overload, if I give you 500 bits of information and ask you to pick out the most salient, it’s going to take you time to process that. And then you’ve got another 30 patients coming in the door, and another 20 patients have died today, and another 30 patients are coming in the door, there’s huge processing overload. The data comes from everywhere. Multi-vendor systems within an intensive care unit, devices that don’t talk to each other, that data’s got to be processed, all resulting in information overload. So, other contributing factors, we started to look at and talk about the human factors within that environment. So patient monitors, nursing data that’s coming in, IV pumps, ventilators, medication orders, none of these are to do with the clinical problem that the patient’s got, this is just additional information that people with different levels of experience have to try and process.

So quick question for you all, you don’t have to answer, but just think about it. What’s wrong with that patient? Typical clinical information system, observation charts, loads of data, every device in the department chucking it all in. How on earth do you know whether the patient needs invasive catheters and drugs or just needs to go to the toilet? It’s really difficult to tell. And we’ve got 20/30 plus patients in our ICU that we’re expecting our clinicians to make sense of this. So we’re not quite there yet with the data.

So, another question for you, what’s wrong with this patient, what’s the most relevant impact on that patient, where are you going to look? I would hazard a guess to say that red is the colour that’s going to draw your eye. So on this patient, we know that their heart’s struggling and their lungs are struggling. So this is a dashboard generated by exactly the same data that we’ve just looked at, but it goes through an algorithm. It goes through some AI process. Not to make diagnosis or make decisions, but just to show us these are the salient factors. So I now know that this patient’s got a problem with their heart and a problem with their lungs. So I get my consultant to come in and say don’t worry about their feet or them going to the toilet, just check their heart and lungs out. So we dig deeper into a system and we’ve got more information. So, we’re going down to the narrow part of the funnel, instead of coming up and getting too much information. This is what ITUs in the future are going to need if we want them to do the job that we’re asking for them. They’re going to need that synthesis done for them.

So if we do that this is the difference. So reduced cognitive load for our physicians, happy doctors, happy doctors make better decisions. And if we look at the standard, what we’ve usually got, against that novel survey and that research that’s been done, it shows that we can get quite a significant reduction in that cognitive load. Because they’re not overloaded with information, they can concentrate and focus, and then we get reduced errors, which gives us happy patients and happy lawyers as well to that effect.

So with that novel-based presentation of data, just focusing our mind a little bit, the research that we’ve got here shows that we get, we’re better at getting the information that’s needed in a timely way.  I don’t want to know when my patient’s had a cardiac arrest. I want to know six hours before so I can stop the cardiac arrest. We’re brilliant at responding to cardiac arrests. We’ve got medical emergency response teams, crash teams, code teams, you know, whatever you want to call them, emergency teams. And it’s a bit of a buzz as a clinician to be on that team. To me you’d get more of a buzz if we actually stopped cardiac arrests happening. It’s in our grasp to do that. Doctors report that they get better at presenting information and understanding it. It’s given to them in a more useful way. And it actually hopefully will then make that data gathering less difficult, less mentally demanding.

So another type, because I’m trying to be vendor neutral here, so this is a different one. So what’s wrong with this patient? Again we look at the red. Clearly right in the middle of that storm scenario we’ve got a hurricane in the middle haven’t we? We’ve got sepsis. So that’s the primary problem with that patient. What’s the primary driver of that sepsis on this picture, that big red blob up by fever, so we know that they’ve got a temperature. We also, if we focus, we know they’ve got a low sodium. They’ve got a high heart rate. We don’t have to sift through Excel spreadsheet type presentations, we just get this and then we can respond to it straightaway.

So, other issues for our clinicians. Non-actionable alarms endanger patient safety and staff health. Sometimes I would rather the alarm didn’t go off. You guys have all been into the HDU and the ITU recently haven’t you, there’s a cacophony of noise, how people can think when they’re working in that environment is really difficult. So some people say that alarms are a good thing, but we have to put them into context. First example that we’ve got up there is without any negative consequences the conductors on the train when they got a proximity alarm just shut it up, because there was a fault on the system. And over months and months and months that was the standard response, proximity alarm, there’s another train, one day there was.

The next one there, asystole alarm silenced. And it kept repeating, silenced, kept repeating, silenced, with no intervention, that information and that alarm overload. We really need to address this in terms of our monitors. More alarms is not better. Alarm fatigue in the ITU is a real phenomenon. You probably all, to my view I’d say you all have the ability sitting in your department on your bench with a really busy time sensitive task we can zone out the telephone. It doesn’t matter, we’ll get that later. That’s what happens on the ICU, they zone out the alarms, because they’re used to them going off all the time. So they’re flooded with 150 to 350 alarms per patients per day. That’s where that cacophony comes from. So we need to address that.

How do we start addressing these things then? Three ways, it’s not a technology answer. It’s technology, it’s people and the process. And manufacturers of medical devices need to interact with hospitals and customers and clients to partner up with you. Because we’ve got the technology, we’ve got perhaps the insight to deal with this, but we need your input and your buy-in to make these systems work.

Existing technology is there, OK, so rather than just showing numbers on a monitor, if you can see on the bottom there there’s some graphs up and down. If it’s above the line it’s pretty good, if it’s below the line it’s not so good, really simple representation. Majority of manufacturers’ monitors have this facility on them, but we don’t utilise them. We go for bog standard this is what we’ve always had and this is the way we always do it. It brings out something called the Semmelweis reflex. Mr Ignatius Semmelweis in 1833, if I remember rightly, put forward a proposal to the medical fraternity that would improve patient outcomes. They disagreed with him fundamentally, put him under so much pressure they ended up locking him up in an asylum. Do you know what his suggestion was? If we wash our hands we’ll cut infection rates. The Semmelweis reflex, if we keep doing what we’re doing we won’t cause any more harm - not necessarily true.

An example, St Antonia’s Hospital in the Netherlands, by putting an alarm management programme in to cut those alarms, we got a 40% reduction in non-actionable alarms, 54% reduction in clinically non-actionable SPO2 alarms. That patient sat there wiggling their fingers. And we reduced the alarms per bed per day from 344 to 204. Still really high, but hopefully those are now important alarms that people should be responding to and reacting to.

Environmental conditions, you’ve all been into intensive care units, I lived in them for nearly, well nearly 30 years actually. They’re not good places to be, they’re not conducive to getting better. ICU delirium is a real phenomenon. And we’ve seen it more and more with COVID: patients coming out or being agitated, being disturbed just from being in that environment and being sick. So we have a normal circadian rhythm: melatonin, serotonin, light, night-time, daytime. ITU interrupts that: lot of bright lights and lots of noise going on. If we want to prevent ICU delirium, one of the strongest factors we can do is maintain that circadian rhythm, and we have to work with the environment to try and do that.

Patient five on the bottom, awake, asleep, awake, asleep, awake, asleep, and these are different examples of the different types of sleep patterns that you’ll see in an ICU - none of them conducive to good circadian rhythms. If we go back to the previous slide you see that nice cadence, that’s what we want to see. We don’t see it in ITUs, because we intervene with them all the time.

So how can we help with ICU delirium? Midazolam, cor it’s a great drug, there you go, off the sleep, nighty night. Propofol, even better, as long as you’ve got an ET tube give them some Propofol, they haven’t got delirium. There’s hyper delirium. That’s the patient you’ll see the staff and the doctors physically fighting with, and then there’s hypo delirium, and that’s the shutdown switched off quiet patient in the corner. No less confused, no less distressed. So what we can do is something really simple. We can make daytime and night-time a reality for these patients by the simple use of lighting solutions. Try to restore that circadian rhythm. And we’ve shown with research that you can repair it within a couple of days. So that recovery phase, the last few days in ITU, if we had some bright light for them on a specific schedule, we could fix that. You can see that different colour lights, different temperature of lights can make a difference. The research is there.

So what do we do with our ITUs? We turn the lights on, we keep them on, and we get that. No wonder patients come out a bit crazy. To work in that environment, and you guys are probably looking, I recognise that one, or we don’t use them anymore, but we’ve got multiple devices in there. Somebody once said to me when I started in ITU, if the number of invasive lines going into your body is higher than your haemoglobin, you’re not coming home. Average haemoglobin is around about 10 to 12, you look at all the pumps that we’ve got on there.

So what can we do about that? Wireless technology well-lit rooms, calm environment. And if you look closely at that, there’s no less equipment in that room, but it just changes the ambience. They’re subtly hidden. We use wireless technology for the monitoring. All of this designed to help with ICU delirium. We have a screen over the top. This technology’s out there, it’s not just Phillips, there’s lots of vendors out there trying to put this in, but it’s not seen as valued. It’s let’s get some more syringe drivers in or a new ventilator, but let’s not worry about the lighting. ICU delirium accounts for nearly 35% of excessive length of stay in ITU at around about three and half, four grand a day.

Wireless sensors, there is a, this is the only Phillips bit really, there’s a new iteration out there on our stand of the wireless sensor, the biosensor. Let’s drive down that route for the future, that vision of the future in ITU is very much that Star Trek module, let’s stick something on that transmits. We’ve all smart technology in our phone, why is that focused on making a new app to get us onto TikTok or something like that, when that technology should be focused to this environment.

Contactless measurement, let’s use audio visual. We can actually monitor heart rate. We can monitor saturations, blood flow, just with a camera, the technology’s out there. We need to do that in ITU. ITU delirium, there’s a great trick, there’s a great product out there at the moment that looks at the way a patient moves in bed. Anybody had a sleepless night recently? Tossing, turning, twisting, changing the pillows, throwing them out, too hot, too cold. Our patients in hospital do that. And if we track them with a camera we can get a prediction, an idea of what their level of delirium is, and then we can institute treatment earlier.

So how does it all fit together? I’ve raced through this. What I wanted to give you a sense of is that COVID-19 in terms of meetings, the way we interact with people, has given us a huge leap forward in the way we interact, probably by about five years current estimate says in the way we train virtually, we interact virtually. We’re looking at virtual technology to support each other. It’s forced us into that two, three, maybe even five year leap forward. It’s done the same to medical technology, to ITU technology, and COVID has put such a challenge on. There isn’t enough nurses and doctors out there. So we’ve got a COVID pandemic, let’s take some nurses, because they’re nurses, from the outpatient department and put them in the ITU and give them 20 minutes of training. No wonder our errors and our stress goes up, because they’re not suited for it.

So we need to have a bigger solution and embrace the fact. This is my message to you guys, EBME as scientist, is try and step out of that Semmelweis reflex of this is what we’ve always looked at, this is what we’ve always ordered. What we need to be doing is embracing that change, that technology leap that’s been forced upon us and see if we can fit it into our most stressed and challenged environments.

So solution criteria, when we’re looking at equipping ITUs and building ITUs, we should be looking to improve staff working conditions. And I’ve deliberately put that at the top, because without that it doesn’t matter what technology you’ve got in there. If you haven’t got the clinicians working well, working in a comfortable supportive environment, then we’re not going to get good outcomes.

Operational efficiency, let’s do more with less. Let’s use the technology we’ve got in there, link it together. Let’s put it on dashboards that show the real picture, rather than another server and another bit of software. Let’s use what we’ve got. Hopefully then we can improve outcomes and we should be focused on that. Maybe if a manufacturer comes to you and says buy these monitors, challenge them, say what’s it going to do to my outcomes? If it doesn’t improve my outcomes, I’m only going to pay you 50%, I’m not going to pay you 100%. Try it, see what they say. Let’s use the technology we’ve got.

We need to enable better decision making, and that sense of detection, and that’s where AI comes in. I think we’ve been taken on a journey now in terms of technology with COVID that we can no longer afford to be as wary as we used to be about AI. We need to embrace that. So a really good system we’ve come across recently, everybody that goes into ITU as a population, 30% will get an acute kidney injury. That’s fact, just by being in ITU, those are the figures. We know that there’s about 30 clinical indicators of acute kidney injury. So let’s programme our information system to constantly look for us for those 30 indicators. And when it finds them wave a red flag, put it on a dashboard and tell us. But then the big challenge for us to do then, to actually improve those outcomes is make sure that we support the clinicians and the medical teams and the biomedical teams to actually put processes in place to respond. Because it’s the intervention that makes a difference; it’s not actually getting the data, it’s what we do with it. That technology again is coming to the fore and we need to embrace it. And we need to improve the environmental conditions.

I went into a hospital probably about eight months ago in Bristol, and I’m an ITU nurse for many years. I was a top cardiac specialist, trained to do resternotomies on my own without the doctors present. Reopen their chest, really used to working in that environment. I lasted five minutes, because the alarms were too loud and it was too busy. I just be in, it was too stressful, and I’m an experienced nurse. So we need to make sure that that environment is conducive to that concentration and that focus. So that holistic approach to light, noise and disturbance fits in with our solution criteria to reduce those errors and complications. It’s something that we forget, you know, that we need to replace normality if we want people to thrive and to get better and to improve. We need to really challenge ourselves. Rather than fostering and generating this unbelievably alien environment that ITUs are, we need to try and make it more like the norm, and only then can we get a real progression in ITU, other Mr Semmelweis will just keep driving us forward, and we’ll do the same thing, the same kit, the same motivation.

As you can probably tell I’m quite passionate about this subject. So we need to change our viewpoint and our process, if we want to truly embrace ITU and that vision for the future that I’ve alluded. I’m not going to be so arrogant to stand here and say I know what the ITU of the future looks like, but it’s that process of change that we need to embrace. An example that I’m going to throw out to you as my last point is a little bit controversial. I do get shot down for it. But every single one of you in this room, I guess, and I’m making an assumption, but it’s usually pretty true, have had at least one training session on basic life support. You’re healthcare professionals, you work in that environment. So hospitals spend millions and millions and millions of pounds every year training all of their clinical teams to respond to cardiac arrest, basic life support as a minimum. The data is there from UK cardiac arrest audit, and all of the ITNARC data, 50% of staff that work in the health service never see a cardiac arrest in their whole career, full stop. Of the other 50%, only 30% of them actually see a cardiac arrest, and even those people that are present on high level cardiac arrest response teams, the median is they see one a year. Can we not take that money and invest it in embracing change, in training people and putting equipment and processes in there to help our clinicians spot that deterioration over here before they fell off the cliff and we have to rescue them.

That’s the ICU of the future that I see. One that’s got the technology, the support and the guidance to be able to say hang on a minute, stop, intervene, stop the falling off the cliff. If you have a cardiac arrest in hospital, quality adjusted life year, which is measured at four years, it’s about 14 to 16% survival rate, and there’s only about 80 cardiac arrests per district hospital per year. Those are the UK figures, yet we invest millions and millions of pounds. So I’m suggesting that we should stop training everybody in basic life support in hospitals, not outside because that’s completely different. If I fall over with a cardiac arrest on my way home today, I want somebody to resuscitate me. But in hospital we’ve got facilities. It’s just changing that paradigm. But that’s, I put that forward as a bit of a controversial challenge, but that’s the way we have to think in my view. We have to change our tack, and look at making sure that that ITU is fit for purpose but in a different way. And on that note thank you very much for listening.

APPLAUSE

If anybody’s got any questions or any challenges, or tell me I’m talking a load of rubbish, then feel free. I will be on the Phillips stand later on in the day if anybody wants to come and have a chat, but thank you very much for your time.

DR JOHN SANDHAM

Any questions? I’ve got a question before you leave the stage. So the ITU of the future, with regards to the data, do you ever foresee a time, a bit like Star Trek, where somebody might be sitting in the equivalent of the bridge and looking at information that’s coming from the ITU, and allowing the AI, the artificial intelligence to give more remote monitoring of ITU patients?

PAUL HINCHLEY

Definitely, and if we look at our colleagues over the pond in the US, there was an experiment done in Texas, a virtual ICU, and it was a bank of monitors and central stations, not on a hospital site, but linked to five different ICUs. All of that data came in, and they had clinicians just viewing and looking at it, and calling them up. They had cameras in the rooms, and they would ring up and say this patient’s bloods say this, this infection rate’s there, let’s start antibiotics, and they improved outcomes by about 20%. So that technology is there, that premise is there, it’s phenomenally expensive at the moment. But all it did is it looked at existing data, and gave it to people that that’s all they were doing. But one of the interesting facts that did come out of that is they started off with eight hour shifts, and they very quickly reduced them to four hour shifts, because they couldn’t concentrate, again that information overload.

So I think there’s a bit of a dual thing there. We’ve talked to some customers from a Phillips perspective recently whereby they’re looking at remote monitoring, particularly in COVID. So a small district general hospital with a full ITU can dial into a consultant at the teaching hospital just up the road, and they can overview those patients and give advice and support. But that’s more about communication rather than about that AI. So a bit of both really yeah, but I do see that, that we need to allow that AI to guide us.

DR JOHN SANDHAM

Have you got any personal experience of AI in use in an ITU environment?

PAUL HINCHLEY

I have. So it’s that acute injury one that we’ve looked at. It gives a synthesis of those. It can bring together 30 points of data and give you quite a sophisticated - I guess it’s like an early warning score system that we’re used to for vital signs. But what it does is it risk stratifies those individual points, and the algorithm is quite sophisticated in as much that if you’ve got a tick on number one and number four, they get one point. But if you’ve also got number five added in, those two get two points because there’s a bigger effect on that. And that’s where that processing can come in. But at the end of the day it just shows us that there is a problem and we have to react to it.

DR JOHN SANDHAM

That’s really good, well thank you very much.

PAUL HINCHLEY

No, thank you very much for your time.

Paul Hinchley's presentation at the EBME Expo : A Vision for Critical Care: The ITU of the Future.