Skin cancer and virtual shrinks. Depression and Instagram. Where would we be without AI?
Skin cancer is one of the most common human ailments – most frequently in Australia and New Zealand, but also 5.4 million USA Americans are diagnosed with the disease every year. At a cost of over $8 billion. Fortunately, when diagnosed early enough the prognosis is good.
For some years now our smartphones have been very helpful in bringing about this happy ending. Selfies of skin disorders and lesions can be sent to an array of apps that analyze them from a distance, deploying AI neural networks. Photos are not the perfect way to identify malignant cancers of course – biopsy provides more essential information. But AI and algorithms already perform better than certified dermatologists looking at the same images in selecting what patients need biopsy and what patients only need reassurance that all is fine. (Research findings in Eric Topol in ‘Deep Medicine. Chapter: “Doctors and Patterns.”)
Recently Instagram did an interesting experiment with deep learning AI. From the images people upload, can it predict which of them are depression-prone? Yes, AI can – though not with 100% accuracy. However, the accuracy scores by AI are better than those of general practitioners, though not as high as amongst psychiatrists.
Apparently, analyzing Instagram images on what colors dominate, whether there are people in it, and the frequency and time of uploading have predictive power that goes beyond the perceptive powers of average humans. Facebook, in a similar drive to assess health risks, has developed AI algorithms that analyze parts of its users’ posts for risk of self-harm. The algorithms are quite successfully, the company claims – but former love-brand Facebook is unwilling to disclose the algorithmic details.
It seems safe to predict the arrival of the virtual shrink soon. Research shows that we disclose intimate details of our psychic lives easier to virtual shrinks than to the real thing, as we feel more reassured that the virtual ones will never judge us. That’s the advantage of smart empathic robots.
But the virtual shrink with whom we communicate via our keyboards in the comfort of our homes come with more advantages: they collect and analyze more signs of our psychological state than human psychotherapists can. The AI-empowered virtual shrink for instance, will register your speech by word pace, length of phrases, coherence, volume etcetera – and will do so more diligently than human professionals are capable of. Also, while conversing with your shrink, your actual keyboard behavior will be analyzed for reaction time, speed of typing, use of emoji’s etcetera – once again revealing more of the deep state of your mind than human shrinks can observe.
Add to this all the mini emotions betrayed by your face, your tics and eye movements that only face recognition AI can notice – and save as data. And what AI-empowered biometric sensors will inform your virtual shrink about heart rhythms and rates, about breathing movements, about sighs and voice modulations (‘honest signals’ these are called, because no patient can manipulate them). It all adds up to a not too-distant future in which AI shrinks will compete with, sometimes even outperform, old school ones. (Eric Topol in ‘Deep Medicine. Chapter: “Clinicians Without Patterns”)
From shallow to deep: Rise of Sensor Society
The examples above illustrate the future of disease management – and, more profoundly still, of health management as well. It is a revolution with ever-expanding implications and ubiquity.
Two key factors are behind it. One is the rise of Sensor Society. Consider the avalanche of searches we perform by clicking on our keyboards. Consider body implants and health (risk) apps measuring our physical functioning, from heart rhythms over glucose levels to breathing capacities – plus our mental functioning, from stress levels to online depression assessment. Consider the health-related selfies we make of our bodies, from skin spots to potential eye diseases to measles indications on our children. Consider the questions and requests we make of our virtual assistants. Taken together, it boils down to the Rise of Sensor Society in which vast amounts of health-related data are perpetually harvested by a host of sensor devices; all of them non-existent only a decade ago, and now fast-becoming universal.
The second factor behind the future of disease and health management is that we now have AI’s algorithms ready to analyze all the data that comes with living in Sensor Society, that comes with 24/7 360 degrees real-time monitoring. Together it will transform our current, ‘shallow’ health industry – based on isolated contacts with our doctors separated over extended stretches of time – towards a deep health industry, in which we are meticulously monitored, warned and advised on an almost continuous basis, with more scope, depth and quality than existed before. All supported by a profound abundance of real time data. Fully empowered and supervised by AI.
Livongo is a digital health platform. It focuses on help for diabetes patients. More recently also patients with hypertension, weight challenges and mental problems can find a benevolent place at the platform. The expansion will continue. Diabetes patients must make many micro health decisions daily, for which consulting doctors are not available. At Livongo patients can interact with each other, but even more importantly the platform continuously collects data about each patients’ blood sugar level.
When things get dangerous the patient will be advised by a coach within 30 seconds what to do. Depending on the danger level the coach can be human or a bot. Depending on the level of urgency, both can communicate more, or less, intrusively. Livongo collects much more data than simply blood sugar levels. It knows about each patient’s disease history, medication over time, diet regimes and fitness exercises. Based on all these data – in amounts no real doctor can digest – Livongo does much more than broadcast warnings. It reminds you to take your pills. It anticipates your high and lows, encouraging you during the last ones, challenging you to improve during the first ones. Based on much deeper knowledge of who you are, what makes you tick, what motivates and irritates you and which are the best moments to engage you, Livongo can powerfully nudge you into healthier lifestyle choices.
Livongo is not the only digital disease/health management platform. There is Onduo: “To bring the best, most updated care to people with diabetes, anytime, anywhere.” There is Omada: “Empowering people across the chronic disease spectrum to set and reach their health goals”. All platforms have the same kind of DNA: utterly digital, getting smarter with each click, continuous monitoring by remote tech, functioning and delivering in real time, automated like nowhere else, yet at the same time performing deeply personalized services to all, in the meanwhile sorting out meticulously how to engage you into coping with your diseases and improving your health.
What to expect?
- Expect the solid worldwide rise of digital health platforms. Some will focus on helping you to cope adequately with a specific disease. Others will focus on prevention, e.g. improving your health. All will show the tendency to expand as, here and everywhere else, software eats everything it, so when a platform is strong in helping diabetes it will be inclined to expand to adjacent terrains of illness and health. Whether in the end, one winner-takes-all platform will dominate or whether we will see the blooming of a thousand flowers, is uncertain. In the digital world, the winner-takes-all dynamic is always urgent. But when it comes to intimate issues of sickness and health, we might prefer highly-specialized platforms.
- Expect the decline of the traditional hospital. It will be a careful, balanced decline as no one is encouraging the ‘move fast and break things’ techno-mentality when it comes to human lives. But why stay in a hospital when you can be monitored, supervised and consulted via remote tech in the comfort of your own ‘home-spital’?
- When you have caught a cold or feel worried by an arcane new body symptom, don’t expect the traditional family doctor to be your first point of contact. Increasingly it will be your platform coach. Bot or human. Coach Will Know You Better – thanks to the avalanche of tiny data exchange moments you share with him/her/it while hardly noticing. Your coach will be driven by the three P’s: Preventative, Precise, Personalized services.
- Expect a change in focus from curing diseases to improving health. Collected platform data can be equally analyzed for both. Imagine for instance collaboration between specified disease management platforms and general digital fitness platforms. We are going beyond the pill.
- Expect your platform coaches to become experts in nudging you towards less illness and more health. Always guided by AI data analytics. The platform and its coaches will excel in the art of nudging you with deeply personalized suggestions.
- Expect that very many people, both the ones suffering from diseases and the healthy (in the end almost everyone in the last category will surrender to the first category), will feel empowered by the digital platforms and its tools to realize more agency over their lives. For patients it will feel like more service and care, in quality and quantity, which means: more independence. For healthy people it will simple mean more joy in living.
The platforms represent the arrival of a powerful alternative in health care. It will invite us to collectively question the culture and habits of traditional health regimes and institutes. The platforms will educate us to be more demanding.
On a business level, digital platforms represent a new road into health care. The old road runs via family doctors to hospitals and their related health care companies within the existing system. Digital platforms disrupt that, with enough innovative power and data to substantially change the competition field. The platforms build new channels through which care can be delivered, deriving its power from a much more in-depth knowledge database – continuously trained and updated, refined and personalized, as only AI can.
As a consequence, expect changes in brand loyalty. Will you be faithful to your familiar doctor or will you learn to rely on your virtual coach more? When your digital platform recommends you a drug from a brand you don’t recognize, do you trust it enough to adopt it? Will this nudge you into letting go your insistence on Big Pharma logos on your medicine packages? Who you trust more, the digital platform you are intimately connected with, or Big Pharma?
Of course, expect intimate collaborations and mergers between Big Pharma and the digital platforms. There are major shake-ups on the way. What they will look like, and especially who will dominate, will depend on how we collectively approach the digital platforms – with trust or hesitation or both. And related to this, how we deal with Big Pharma – and them with us. The trust dimension will be key.