Thank you. I understand that the King’s Fund is 120 years old so that’s a respectable age. I represent a company that’s 126 years old so we’re slightly older. Also it’s interesting we’re a company so we have to reinvent ourselves time and again, time and again, and we actually did that recently. We started out as a lighting company 126 years ago because two brothers came to the conclusion that with the emergence of the electricity grid, light would change the way people live and work, then early last century Runche invented the Runche machine and I thought, okay, this is going to change the way medicine is being practiced and diseases being diagnosed. Then a couple of years later they jumped on the radio as a completely new way to communicate to people and this has had a big impact on the political landscape and the social landscape of Europe and then the television making it even more immersive. Interestingly all the same core technology, tubes, and now we’re in a very different period and, for me, actually yesterday I did a lecture at the university where I graduated back in Amsterdam and I said, “Okay I sat here 35 years ago and actually literally in this room studying computer science and I was being taught the models of if-then-else, creating models that you bring to the logical conclusion and so we thought about administrative systems and we had some ideas about the future, but never could I have imagined where we are today 35 years later,” and I’ll talk about it and it’s extremely exciting time for not just technology but the application of technology in our own world, our personal lives and the lives of our kids.
I want to start with an organisation that’s 70 years old, the NHS. I think you’re rightly proud of what the NHS has done for the population here in the UK. There’s been a whole lot of progress in the way we apply medical science, but essentially the system hasn’t really changed. Essentially, and it’s not just in the UK, essentially we’re still having a model that’s organised around sick care, you go to a GP when you’re sick or when you feel sick, they do an initial diagnosis, they refer you if something is bad to a hospital, you get another consult, you do some tests and you may undergo a procedure, you get medication, you have to go back, it’s based on appointments, it’s based on brick and mortar, it’s being reimbursed largely for activity. If you now look at the burden of care and you look at the real needs of people it’s different, it’s chronic disease. 80% is spent on chronic disease.
You read stories, at least I read them even in the Netherlands, about emergency rooms that overflow on the weekends, GPs that are getting burn out because of the pressure of their jobs and it makes you think. It makes you think that with the technology today and our understanding of what drives people into care is there not more we can do? Is there not more that we can do to anticipate the needs of those people to help people deal with true chronic disease? And chronic disease is actually 24/7. I know that personally my daughter was diagnosed with type one diabetes when she was twelve years old, she’s been suffering from a lot of complications, she’s been suffering from anxiety and I’ve seen up close how the care system in New York as well as in Amsterdam, and both of them generally regarded as the best in the world, are not equipped to deal with the requirements of the chronically ill, with the complications that come with the disease, with the way we can impact that disease. If you look at the demographics of both the UK, the Netherlands and the US you come to the conclusion that we actually have to pivot the system and we can do that because a lot of the technology is here today but it will take time to adjust the system.
So what is there today, and this is a busy slide but that’s because so much stuff is happening and it’s happening now. We have cloud technology that allows us to have tremendous compute power and storage going very rapidly to zero cost. We have sensors everywhere, everywhere, sensors for waves, sensors for our flow, for light, for … we put sensors everywhere so we can sense anything around it. We have micro and nano cameras that we can put anywhere and then we have the tools that we can actually use to make sense of all that data that’s flowing. There’s no longer structure like I learnt when I was studying computer science where you apply if-then-else on. No, it’s this data that we can find patterns in that we can apply statistical methods to give us the highest probability of an outcome. So we can use this. We can use new interfaces, conversational interfaces, everybody is using them today on Siri, on Alexa, on Google Home. These become very, very, very sophisticated and there’s this concept that we embraced as Philips 20 years ago called ambient intelligence, what does it mean? It means that technology disappears in the background. It’s just there. It knows who you are, it adapts to you and it provides information about you. So that’s the theme I come back to because it’s directly applicable to health care. Robotics, autonomous systems, all of this is relevant to health care.
So how is it relevant? Number one, we can make and we make our devices smart. It used to be that if somebody in an intensive care unit gets put on a patient monitor that monitor has to be configured based on the patient. The MR machine has to be configured to do the diagnosis for this specific patient. What if that machine knows who you are? What if that machine knows about your clinical history, it knows about your physiology and it automatically will configure to you? It will glean information from other devices that’s surrounded by you. That’s what we can do today. These are smart devices. These devices know who you are and they support you. We create this ambient intelligence around us not just one device it’s ten/20/50 devices. It’s the smart work flows that make sure that the nurse gets to that patient that needs an intervention now and whether that patient is here in an intensive care unit or it’s at home or somebody somewhere that has an accident.
So smart work flows and digital ecosystems that allow different people, different organisations to participate in the single smart work flow. So that means that we’re going to create continuous health tracking. We’re moving monitoring out of the intensive care unit - the only place actually today where we do continuous monitoring and the only place where we can do it with very high accuracy because people are stationary, but now we can do it in the general ward. We can do it wherever you are because we created a patch with a set of biosensors that you can put on your heart that will allow medical great monitoring anywhere. We link that with predictive algorithms that can help us see deterioration very early on and we can actually start predicting cardiac arrest 48 hours in advance because we suddenly see an increase in probability of that event happening. An increase of probability. We’re never 100% sure but these big changes tell us something about what kind of intervention we can do.
What’s really exciting is what we’re doing with imaging. So talking about radiology. Radiologists for 100 years have done the same thing, they’ve been looking at an image and that image increasingly became more accurate, more higher quality, but they’ve been transcribing, they’ve been explaining what they see on that image in a report. Pathologists do the same, for 125 years they’ve been taking a tissue looking at it under a microscope and describing what they see on the tissue, but now we’re applying deep learning on those images. We can start making a distinction between what’s normal and what’s not normal. We can actually measure things, quantify lesions, we can quantify brain atrophy. So if we can quantify that between two studies we can measure how much the brain shrinks but we compare that to normal and if that’s faster than normal it may be an indicator of a neural disease. If we can see tiny spots that a radiologist can never see with his naked eye that can be an indication, combined with atrophy, of onset of Alzheimer and we may see it ten years before we see the symptoms and actually there are some therapies that can apply to Alzheimer’s but only if you catch it before you get the symptoms, but this can be applied to many other places where more precision will allow us a better diagnosis, a more quantified diagnosis which will allow us to compare to other things and to select the right therapy with it. So there are tremendous opportunities to quantify, to compare, to measure and to find patterns in complex information.
One of the most interesting things that you must have heard that before is what’s happening with genomics. So actually last week somebody showed me a prototype of a tiny little device that you can literally spit in, it will make a genome sequence from your spit and it actually will spit out a genome sequence. Now obviously you can program that to look for certain disease biomarkers which is way easier than just doing a whole genome sequence, but we’re looking at hand held devices that you can carry anywhere that almost in real time will generate a genome code. If we combine that with what we see on those images, what we see in the digital pathology then we can become amazingly precise in what we see and then we can find the therapies that best fit with the diagnosis but not just with this diagnosis also with the personality of the patient, because I always believe there’s a difference between precision diagnosis and personal therapy because personal therapy is by definition personal, it has to do with your outlook on life, the way you look at risk, the discipline you can apply to the therapy and a whole bunch of other factors, but that’s a big change from the way health care is practiced today.
The current model is that we spent 88% on medical services, only 4% on helping people deal with healthy behaviours. If you look at what the true health drivers are these are not again exact science because they’re approximations, but there are many studies that say that healthy behaviours are bigger drivers of health than access to clinical care and of course healthy behaviours we all know what that is, the way we eat, sleep, drink, drugs and the environment in which we operate and genetics, yes, are very important but clearly they tell us that we’re disposed to a disease but something still needs to trigger the disease.
So what we’re thinking is that with all this technology, with a different approach to clinical and social care you can move to a different model. A model that’s starts with understanding the needs of the patients and segmenting these patients accordingly and this is not new. I worked at City Bank 25 years ago, I ran their tech lab and the first thing we did was segmenting 30 million customers on their financial needs and we then organised our products according to those needs and we crafted the services around it. We can do the same. You can look at very sick patients with four or five chronic diseases and so seven medications, you can look at people that have early onset of type 2 diabetes, you can look at people with heart failure and diabetes, you can even look at people at risk for breast cancer and create a program around it. In other words, once you understand your patients and not just from a clinical perspective also from social economic and behavioural perspective, you can become very precise in the way you segment those patients and then you can become very accurate in the way you want to service and support and screen and manage those patient groups. That’s a completely different model, that’s a proactive model that starts with a deep understanding of the population and we can do this today.
What we can then do is home in on these key drivers of health and identify what we can address in terms of these drivers so that we can start creating better outcomes and better outcomes is obviously with it’s about. So we as Philips are a big proponent of value-based care, value-based care is an approach to care that starts with the patient and patient outcomes and today in the Netherlands or in the US and most probably in the UK as well there are around 2,000 key metrics that are being applied to the quality of care and they’re being used to make sure that the reimbursement that takes place is according to these quality standards. Only 7% of these 2,000 is related to outcomes, only 2% is patients reported. So I think that needs to flip and, yes, we applied all these quality measures because we still have a fee for service, an input based, an activity based reimbursement system, but if we flipped the model we can drive towards outcomes.
If every hospital, every clinical provider makes their outcomes transparent we can find where the variance is, we can find where the opportunity is and, for instance, there’s a hospital in Germany that I admire a lot it’s Martini-Klinik, they’ve been unasked publishing their outcomes on their treatment of prostate cancer, but the result is very ... have way better outcomes than any other hospital in the region. They entirely organised their hospital around value-based care. So what you’ll find then is that the people who do that, use that information to provide better care and therefore create better outcomes.
So having access to data helps not just to report those outcomes, track those outcomes, but also to start creating this knowledge base around care flows, around segmentation, around stratification and around achieving those outcomes. So, for instance, what we’re doing here also in the UK is digital pathology. So we’re replacing the microscope with a machine that actually reads the tissue, creates the image of that tissue, identifies the cell structure but actually can also identify key biomarkers. So this can be used of course in diagnosis, this can be then combined with genomes, this can be combined with imaging and it can be brought together into a full view of a cancer patient that can then be used in a multidisciplinary team to identify the right therapy, but this does not have to be in the same location. This can be virtual. You can have a machine in one town you can have a pathologist in another town looking at the results. You can aggregate those results, you can analyse those results, you can link them to outcomes and you can derive conclusions from that. You don’t need to wait for chronic patients to make an appointment and come to the hospital, you can do virtual care remotely.
We have seen some great examples where we had the sickest of patients, COPD and heart failure and diabetes and early onset of Alzheimer’s, people who panic quickly and we gave them a tablet, we measured their blood pressure, we measured their weight, we measured their SpO2, we asked them some questions every morning and you can easily use a tool like Alexa to personalise the questions you want to ask, did you wake up this morning with any swellings? Did you sleep well? Actually we saw that you woke up 20 times last night to go to the rest room, maybe we should check your bladder. So you can glean a lot of information about these patients.
Actually we took it one step further, we now use video. So in the two way video we can actually measure respiratory rates, heart rates, we can measure temperature just from the video. So we use that in emergency rooms as well, we can in emergency rooms scan the patients and then identify the patients with the highest acuity levels so we can prioritise them that way. So the way we monitor is no longer attaching these sensors to the body, the way we monitor is increasingly going to be through this type of technology.
Emergency care, what we’ve done in Sweden is basically look at patients with the highest probability of stroke, so who are these patients? Well, these are the patients that had a stroke before and then how many of these people had complications with the stroke and actually which are the hospitals that can deal with that? So when an emergency happens we actually know who that person is, what their background is, what their problem is and then we can check with the hospital emergency care whether they have availability to deal with the stroke, we can take a CT scan on the way there, we can measure but also we found out that some of these patients have a complication that can only be treated in one hospital, so if we know that in advance we can route the ambulance to that hospital instead of the one that’s closest by.
So what we’ve done in Dubai is we actually put a GPS in every defibrillator but also we created an app that we connect everybody who can actually do… (resuscitation). Exactly, thank you. So if somebody falls down we can press the button, we can find the nearest person who can do that and we can find the nearest AED, at the same time a message goes out to the ambulance that comes here. So real time dynamic scheduling nothing that has been done before but now, based on the knowledge we have, an increasing knowledge as we rush to the hospital.
So what we’re looking at is actually a world that we believe can be enabled through technology, can be truly adaptive, the ambulance knows who you are, the CT scanner finds the clot, the personnel in the ambulance knows what to do, the personnel in the hospital is ready to receive the patient and ready to do surgery to remove the clot, but similarly if you’re going about the house you’re an elderly person you forget stuff, Alexa is there it will remind you, it will make sure that you take your medication on time because 50% of the people actually don’t take the medication on time. The environment can sense whether you need help, when and what the best way that help is being provided and nothing that I explain here doesn’t exist anywhere else. So all of this exists somewhere else in another industry we just haven’t applied it to health care yet. Dynamic scheduling is very well known in the logistics world, sensing environments you can see everywhere. Most people are using artificial intelligence today.
So what we’re looking for is actually bringing that intelligence, that smartness into the environment we are. What you see there on the top left is a completely adaptive hospital room where the bed, the lighting, the screen, the sensor are in the beds, everything knows about the patient. When the nurse comes in the room knows the nurse comes in and automatically on the screen all the vital information about that patient will be displayed. When the parents of that kid come in the room the screen changes to give them the relevant information about their kid. What you see at the bottom there is a tumour board where we actually bring together radiology information on the cancer, the quantification, metastasising of the cancer between studies, we compare that with the pathology, we compare that with genomics and we then have the system tee up the relevant therapies that would apply to this patient so that the oncologist, collectively with the patient, can select a right therapy and follow up on proper application of that therapy.
What you see as well here is a 3D model that we today generate from an ultrasound. Some of you may remember ultrasound as the grainy back and white pictures, now you take the ultrasound you put it over the heart and it will literally generate a 3D model of your heart. Now, our ultrasound we basically put on a tablet today, so basically anybody who knows how to operate an ultrasound can now take an image of that heart and in real time a cardiologist can look at that heart and here and now you can make the diagnosis and discuss the right therapy.
So what you see is that a lot of that technology allows us to bring the care to where the patient is and make that care really relevant to that patient. So what we’re working on is creating a true digital twin of yourself, a twin where we model the heart, we model the lungs, we model the brain, we look at the history as well, because it’s not just a snapshot that’s relevant it’s how you evolve or emerge, whatever you want to call it, that’s really relevant for your health. So all of that information brings together a view that allows you to better determine what the right therapies are, it creates earlier diagnosis and it can create obviously way better outcomes.
So what we see, what we hope that the 70th birthday of the NHS will prompt is a rethink of the model that has been successful for 70 years but that needs a current look based on where we are with technology. It needs a transition to value-based care, to measuring outcomes, reimbursing outcomes, and, yes, we can measure it now and, yes, we can compare it. Population health, so being organised around the needs of people with similar conditions, similar needs. Network care no longer just a hospital but a combination of hospitals, clinics, home care, a health care system that is truly connected. I come from the financial services industry 25 years ago we created the infrastructure that allows you to pay anywhere in the world with your card or your phone and that means that anywhere in the world they know who you are, they authorise you for that transaction and I’ve been in the weirdest places and it worked, yet in health care we don’t know anything.
So, yes, we need a health care system that is connected that has not just the bricks and mortar infrastructure but it has an information infrastructure, that has an environment that’s aware and truly adaptive to patient needs, where we apply technology but we want to push that technology in the background to make it non-obtrusive, make it natural, make it part of the environment, both the environment in which professionals operate like nurses and specialists and GPs, but definitely patients and understand that health care will be collaborative. It’s not just your heart valve that’s the issue it is a broader set of issues that drives heart disease, it’s a broader set of issues that combines heart disease with diabetes and other diseases. So a collaborative holistic approach and I think it’s possible. With the state of technology I think it’s possible to craft the models around this that truly give better outcomes for all of us. Thank you.