Artificial Intelligence can help to break down silos, aggregate data, integrate our health journeys and reveal patterns to initiate beneficial action
Too much data, not enough resources
On its current trajectory, the world will not have enough human and financial resources to deliver quality care to rapidly growing and aging populations using existing care models. Fortunately, the digitization of healthcare and the development of innovative new data-driven models for healthcare delivery are already beginning to offer solutions. It’s no longer the case that we don’t have enough data. The problem is we already have too much of it. More than we can currently handle and put to good use. Cloud-based Artificial Intelligence (AI) and Machine Learning operating on vast amounts of data from multiple different sources could provide the answer. As scientists look for better ways to diagnose and cure disease, and hospitals look to lower costs and increase efficiency, advances in healthcare AI are about to make their presence felt.
Here are a few ways AI could help to save your life.
Identifying risk – population health management
Neither you nor your doctor might be aware that you have a particular disease, but looking at your medical record it may just be that you exhibit similar indicators to a group of people who do. Pattern matching – in this case working out which indicator combinations correlate with which diseases – is something that machine learning algorithms are brilliant at doing. Given enough data, they can provide clinicians with the statistical data they need to cluster patients into ‘at-risk’ groups that can be closely monitored and treated or given advice on lifestyle changes to prevent a specific disease from developing. Machine learning algorithms are also highly efficient at picking up subtle changes in medical images that are extremely difficult, if not impossible, for human beings to spot. For example, they’re capable of spotting the minute differences between naturally aging brains and the brains of people in the very early stages of dementia, offering the prospect of screening programs and early treatment to ward off its worst effects. To learn how to do it, they need to look at hundreds of thousands of images, so we’re helping researchers in the Netherlands to build huge databases of longitudinal images and patient profiles on which they can carry out the necessary analysis.
Identifying your personal risk – genomics and epigenetics
One question that doctors often get asked by their patients is “I’m not interested in knowing a particular group’s risk of getting a disease, I want to know my risk of getting it.” With AI’s ability to sift and analyze the 100 Gbytes of data that represents your DNA – what makes you unique rather than one of a crowd – being able to identify your personal risk is moving a whole lot closer. For a very small number of diseases it’s a binary decision. If you have the defective gene, you will get the disease. However, for the vast number of diseases it’s a massively more complex problem that AI holds out the promise of solving. But it’s not only your genes that determine your risk. It’s also how those genes switch on and off in response to the environment – your so-called epigenetics. At Philips we’re working on solutions to blend genomic data with other clinical and lifestyle data to uncover patterns and trends that will help to prevent the onset of disease and help researchers find new cures.
Catching you before you fall – predictive analytics
After you reach retirement age, simply falling over can be fatal. Falls are one of the leading causes of fatal injury in people over 65. For several years you’ve been able to wear an increasingly intelligent fall-detecting pendant – one that can reliably sense if you’ve fallen over, and automatically call for help if you don’t stand up again within a certain time. Philips is now taking that a step further, using a predictive analytics engine that combines historical data from millions of patients and real-time information from your pendant to work out whether tiny changes in your activity or gait mean you are likely to fall up to 30 days in the future. Via the use of smart connected body sensors, we’re also helping researchers to develop similar predictive algorithms for other life-threatening events, such as cardiac arrest and sepsis. Alerting care teams before these emergency situations occur reduces hospital admissions and healthcare costs. Most importantly of all, it improves people’s independence and quality of life.
Weaving healthcare into the fabric of your life
So where is all this heading? The answer is the revolution in health data and AI has the capability to create your own unique ‘digital twin’ – the data equivalent of a caring brother or sister who knows where you’ve come from, where you may be going, and how you can be helped to get where you want to go. However, keeping that digital twin up-to-the-minute about your health must be totally unobtrusive, not a burden on your daily life. Leveraging new technologies such as the Internet-of-Things (IoT) and 5G communication networks, future solutions will automatically sense your vital body signs, mood and state-of-mind, responding to them in intelligent ways that reflect your personal preferences and needs.
These are just a few of the healthcare challenges we’re addressing with AI at Philips. By overcoming the barrier of ‘too much data, not enough resources’, we’re seizing a huge opportunity. AI can help to break down silos, aggregate data, integrate our health journeys and reveal patterns to initiate beneficial action. It can build up a complete picture of our health, our lifestyle, and our emotional state – everything. AI's ability to sift through, remember and learn from the vast amounts of data available will get us closer to achieving Philips’ mission to improve the lives of 3 billion people a year by 2030, allowing more of us to enjoy a healthy and enjoyable life.