HyperWellbeing 2016 conference notes

A new technology industry is being born, Wellness as a Service. It represents one of the greatest economic and ethical opportunities of the 21st century. It is also a redefinition of the nature and purpose of computer science.

Introducing the Wellness as a Service industry

This week has seen the highly successful debut HyperWellbeing conferencetake place in Silicon Valley. The event was convened by Lee S Dryburgh, a long-time friend and iconoclastic convenor of radical change agents. It was hosted at the Computer History Museum, a symbolic location chosen to represent the evolution of computing science towards more human concerns. The many speakers were all founder-level, and the enthusiastic audience was equally senior and sophisticated.

Whilst the permanent exhibition downstairs showcased past milestones of the computing industry, like the first prototype mouse, the activity upstairs was making history as the new Wellness as a Service industry had its maiden gathering. I co-chaired the event and below is a synthesis of my notes. For a second opinion, you may also wish to refer to the slide deck put together by Dr Heidi Forbes Öste.

If you would like to stay informed of developments in the Wellness as a Service space, then you can follow @HyperWellbeing or sign up to the email list at the foot of the page here.

The spectacular failure of the ‘sick care’ industry

Our long-term health is a product of our fitness, genetics, and microbiome. The level of wellbeing we experience is roughly attributable 30% to our genes, 60% to our environment, and only 10% to the healthcare system.

In the US, healthcare is 18% of GDP, much of which is expended on prolonging the agonies of death in the final month of life. When you are only addressing 10% of the problem, that is a truly insane amount to spend. Furthermore, 97% of that is spent on illness, with only 3% on prevention, which makes it morally indefensible, too.

What we call the “healthcare” industry is really a disease industry, dependent on an endless supply of distressed customers. The American healthcare business has devolved into an accountancy system based on reimbursement for medical procedures. These (ultimately unwanted) procedures are often commissioned based on inaccurate episodic data. Your mere existence as a sick ‘patient’ is the result of its failure to engage with you as a citizen before you became ill.

The problem of endemic inefficiency and ineffectiveness is not confined to the ‘sick care’ business. The emerging ‘wellness’ industry is highly fragmented. Around 60% of the $100bn/year spent on exercise, weight loss, yoga, etc. is also wasted. For instance, 60-90% of diets don’t work, since the same diet has different impacts on different individuals. Their one-size-fits-all approach and advice is unsuited to the diversity of humans and their needs.

The birth of a new wellness industry

We are at the beginning of a period of enlightenment as a new academic discipline emerges, Scientific Wellness. This quantifies wellness, and brings fresh insight into technology systems.

There is a fundamental transition as the highest aspiration of ‘well care’ takes us from being merely ‘normal’ to aiming for ‘optimal’. After all, the ‘norm’ is actually pretty grim in terms of the impairment we experience, or have impending within us, such as being diabetic or pre-diabetic.

The transition from ‘sick care’ to ‘well care’ is initially brought to us by ‘P4 Medicine’: predictive, preventative, personalised, and participatory. This applies individualised body and lifestyle insight, based on robust clinical and behavioural research.

The end goal (as hypothesised by ‘hyperwellbeing’) is when the aspiration is raised even further, towards fully optimal living from birth to death. This extends preventative medicine to include factors such as our social and environmental context, and wellness at the community and ecosystem level.

To achieve this goal of total wellbeing we must move from discrete (institutionalised) measurements and interventions into a system of continuous (personal) monitoring and lifestyle management. Only as we move towards a denser data set can we resolve the fundamental issue of false positives that has stymied previous preventative medicine approaches.

This is a process of improved fidelity of data that results in ever earlier intervention. The current products and services for personal monitoring are generally attached to the ‘after’ condition of sickness to manage chronic illness. This will move towards a hybrid management model of ‘during’ illness and its onset, and ultimately to an end state with most all the effort and spending being focused on ‘before’ the illness ever occurs.

The achievement of this transformative outcome requires a convergence of disciplines, including fashion, workplace wellness, pharmaceuticals, the current health industry, and the insurance business; as well as consumer electronics, silicon and nanotech, data science and behavioural psychology. By working together these create the new Wellness as a Service industry.

A data-driven health revolution

Historically, the cadence of biomarker sampling has been too low for intelligent lifestyle change. The cumbersome kinds of sensors used for things like sleep monitoring have involved a slow process to physically apply them to the patient, requiring professional technician skills. This is changing as the monitoring systems become miniaturised and smart, and move from the clinical setting into the world at large.

As we acquire a dense, dynamic, personal data set we create a “Facebook for our health” that tracks far more than the 98 things that Mark Zuckerberg knows about us. This provides a synthesis of the “physical me” and “digital me”; both are “me” and “I am the digital version of me”.

These data-driven systems will guide us by giving us real-time feedback on the choices we face and make. How much more (or less) time do I have left on the planet if I eat a burger versus a bowl of prawn fettuccine?

Rather like how we endlessly scroll through social media for gossip, we will face a cognitive dependence on our machinery augmented decision making to help us to live our lives. This might be seen as threatening, but possibly no more than how we are dependent on combine harvesters to keep us fed.

The underlying data set will have a few ubiquitous key personal variables, like heartrate variability, as well as a growing set of other biomarkers. This will be complemented by vast amounts of “library” information on our collective human wellbeing and its relationship to genomics and the microbiome.

These endeavours to create “wellness wisdom” involve processing vast amounts of information. For instance, the current “Hubble of genomics” has 24Pb of data stored, and pushes the limits of Amazon’s web services to deliver enough compute power.

The longitudinal nature of the individual data sets is new, and potentially transforms the incentives for adoption of the measurement technology, and also adherence to the behavioural change recommendations it offers. The quality of the data from consumer devices is rapidly reaching (and exceeding) clinical grade, becoming accurate and standardised. This makes it relevant and actionable.

In this data development process, existing medical records will likely play only a small part. “It’s hard to train a neural network when the data is on paper”.

Capturing the raw data using sensors

The defect rate in medical data is often high, as much as 30%+, which undermines its utility. The accuracy of miniaturised sensors is going up rapidly (with improvement factors as high as 200x reported). As they become smaller, thinner and lighter they can be worn closer to the body and more continuously.

The gold standard is from blood samples. A simple and cheap device can take a microliter of blood, and use algorithms evolved from facial recognition to look at spectroscopy data. This works out how much of key indicators like glucose or triglycerides are present. New technology that prevents the formation of scar tissue allows sensors to become embedded under the skin.

Developments in machine learning also allow highly accurate diagnostic data to be taken from less accurate sensors or readings. For example, machine learning allows us to see changes in self-administered breast ultrasounds over time. Since breast cancer has a 99% survival rate if detected early enough, this is transformational to the treatment of this disease.

A human has 30 trillion of our “own” cells plus another 30 trillion “payload” cells in our microbiome. We are at the point where we can sequence one human genome every 15 minutes, and can construct a “picture” of the human from that data. For instance, 500 to 600 points on the genome are collectively enough to accurately predict a person’s height.

The analysis of the human microbiome is at a less developed stage, but the results thus far are described as “provocative”.

Big data needs big theory

There is a process by which unrefined data is turned into meaningful information, and from this it becomes possible to extract actionable insight for the individual. It is easy to confuse data with refined information: what is today called ‘fitness tracker’ is really only capturing activity.

The challenge, as in all scientific disciplines, is to understand cause and effect relationships, even if they are transient or contextual in nature. This means that Big Data needs to be paired with Big Theory.

The first stage of this is to capture the set of meaningful things to measure. For illness we have an ontology derived from billing, since disease is what the system pays for. One of the present barriers to progress is that we lack an ontology for wellness, a language for describing how ‘at risk’ we are, or a system for solving wellness problems.

We are only just beginning to explore the philosophical and practical structure of this problem domain. We even lack a standard vocabulary to describe food. Machine learning is being used to identify and define the biomarkers that we need to pay attention to.

Over time we will develop standardised health and wellness scores that will be updated in real time. These will cover aspects of our lifestyle, body, and feelings. For instance, we will have a rolling ‘probability to live another 10 years’, doing for health what a FICO score does for one’s creditworthiness.

These scores will initially develop as pockets of ‘vertical intelligence’ to address specific needs, becoming more general as our ‘alien intelligence’ grows beyond the level of wisdom and understanding we currently conceive as possible.

Our current wellness ‘meme pool’ is based on millions of years of interaction of our gene pool with the environment. The rapid change over the last few centuries means we need to upgrade our intuitive wisdom. For instance, if the contribution of diet is roughly 10x that of exercise, today’s wellness regimes have an inappropriate focus on the ‘sweat’ of exercise.

Enabling the ‘upgrade’ will be a sophisticated multivariate analysis of the root cause of lifestyle problems that are rooted in genetics and our microbiome. Supervised learning will go away, with the learning process becoming fully automated. The “big 9” technology companies have already open-sourced their machine learning and inference engines.

Over time every human will have the opportunity to develop in the way a Formula 1 car is designed. This is a continuous field-based trial of living, with the ability to simulate the effect of lifestyle changes, and adjust and evolve to developing needs. Everything will be centred on the ‘driver’, helping to evolve his or her goals, create custom risk/reward profiles, and align long-term goals vs short-term decision making.

Behaviour change is hard, engineering it is even harder

As I type I am sat in a restaurant in Sacramento waiting for my cinnamon crepes to arrive. When I paid the $1.25 extra for ‘ice cream’, I didn’t expect it to be in addition to massive piles of whipped cream. The hazards of being an Englishman in California… Anyhow, as Oscar Wilde once said, “I can resist anything except temptation”.

It’s no secret that behaviour change is hard. We are not good at planning for the future, as the gap between today’s action and its felt effect is too long. Behavioural scientists have taken a long and close look at this phenomenon, as it kills us. For example, treatment adherence for diabetes has a 60% failure rate.

The answer is emphatically not merely more insight and data. In one study, when McDonalds advertise the number of calories on its menu, calorific consumption went up. It is quite possible some buyers were optimising for the number of calories per dollar spent! Appealing to willpower is unhelpful when our immediate desires point in a different direction to our goals. (That said, willpower can do amazing things, like let you vary your fingertip temperature merely through thought.)

People will respond to warnings of impending problems that they may have been unaware of. We need a “check engine light” for the human body and psyche.  Once armed with this knowledge, we can employ techniques from marketing and propaganda to engage in behaviour manipulation for own benefit.

If we want to get real and deep change, then we need new brain/computer interfaces. One important example is biofeedback using haptic interfaces. A use of this technology is “meditation with training wheels”, to greatly accelerate the rate at which newbies can acquire the skills of attention control of skilled meditators. An hour a day wearing the right headset for a year might give you the powers of a monk who has spent ten hard years working continuously on the same capability without technological help.

Another approach is to gamify our behaviour, for instance to help solve the engagement problem with biometric measurement.

What underpins the behavioural change are the neural “pathways” that we create that form our behavioural strategies. These form feedback loops between action and outcome, and are a place for intervention. Our strategiespredict our behaviour, and we can help to form and develop new strategies without having to wait for the desired behavioural outcome to fully arrive. If we can do cognitive behavioural therapy for worms, we can automate it for us mammals, and do it in a way that makes change more “thinkable”.

A journey towards lifetime optimisation

The process of data collection offering recommendations for lifestyle change is one with growing accuracy, and where prices are falling fast. We are now able to capture emotions, what food we eat, and understand our metabolism from body scan heat maps. The cumulative effect of these feedback loops at different timescales is what takes us to Scientific Wellness.

A critical component is the “coaching” relationship between the system of wellness and the individual. The kind of empathy and encouragement skills are those similar to smoking cessation. This today is being largely performed by humans, but it too will become automated by software. What Siri does for personal assistance, new vertical helpers will do for personal wellness.

A comprehensive service of data capture, analysis and coaching today costs around $4k/annum. This is forecast to drop to $500/person/year by 2020. Today’s service already has a 60% renewal rate.

These systems of wellness will help us to navigate our own “wellness possibility landscape”, with its “peaks” of flourishing and “pits” of languishing. We will have to identify appropriate and attainable goals, and design systems for motivation, so as to make harmful behaviours harder and beneficial ones easier.

This requires us to look at the full range of motivations, the availability of different behavioural paths, and the triggers that can take us down each one. We will need to engineer behavioural strategies. These in turn can be bottom-up (from behaviours) vs top-down (from values and goals) vs inside-out (recovering values from what you do).

Which is the “unregulated, over the top” organ?

We have now seen the grand vision, but where to begin? There is one organ that allows us to get through the regulatory barriers, and go “over the top” of the medical profession.

That organ is the brain.

If you want to interact with other organs like your liver or thyroid, you generally need to go through tediously slow and costly approval processes for medical devices. But we can “talk” to the brain directly! Nobody needed to get FDA approval to write an IQ test, for instance.

We are missing vital data about the brain, as basic data from a cognitive screen has traditionally cost around $12k, and comes from a profession of neuroscientists where there are 40k patients per care-giver. The software revolution of medicine is forecast to drop that cost to the $30-40 range, more like a blood pressure test. When anything drops several orders in magnitude in cost, it becomes a fundamentally new thing in terms of its potential.

The mind is the “pioneer” organ of this new wellness movement. For example, there was a single mindfulness study in 1980, whereas there are 80 per month being published today. The mind is ripe for a wellness revolution. We on average each spend 47% of the time with our mind wandering, unhelpfully ruminating on the past or future, rather than being in the present.

If we do nothing more than help individuals to gain mastery over the “noise” in their head, and to quiet their mind on demand, then it will be a transformative leap in wellness of society. Levels of stress and burnout in organisations would plummet.

Electroceuticals and wearapeutics are the new interaction paradigm

Users already face too many smartphone apps and notifications, and are not clamouring for more. Hence the brain/computer interface will evolve to meet these new wellness needs.

Virtual reality is a key enabling technology for engineering wellness. The level of development of VR in 2016 is similar to the Web in 1996. The potential is limitless, whilst the penetration level is low. The applications range from lowering stress as part of surgery preparation, to entheogen simulation (or “digital psychedelics”) that promotes empathy or resets the brain by inducing forms of cross-talk in its circuitry.

VR is experienced not just with clumsy goggles, but can takes on many forms. For instance, it could be a ring that vibrates or glows; or new fabrics that enable recognition of specific gestures; or a device that attaches to our keychain and naturally forms part of our everyday life. The integration of VR with physiological devices is a field that is largely unexplored; computer mediation of our body senses is an opportunity we have yet to seize.

The stars of the show are the gadgets, and there were many on display: too many to list here. The critical factor in their growth is not so much their dropping price and growing functionality. What is really changing is the level of convenience they offer for the emotional benefit they bring.

Wearables have to be fashionable or invisible, with the constraint on both typically being the size and longevity of the battery. The ideal is for a device that never has to be taken off, be it for aesthetic, comfort or recharging needs.

Some are gentle evolutions of existing categories, like hybrid smartwatches that deliver additional sensing functions. Others are far more radical and active in their effect, such as a “mood control” electrostimulation system. The use profile for this kind of device is far from the “geek” early adopter crowd, with a notably high level of female and millennial participation.

Bringing it together: the Web of Life

What we are seeing is the arrival of a new platform, somewhat similar to the emergence of the PC and desktop metaphor, or the Internet and the Web. This platform, centred on the human, integrates a “full symphony” of activities, such as eating, exercise, socialising, and meeting life challenges. When it acts as a life companion, it no longer suffers the engagement decline typical of products that only address a small portion of the puzzle.

This platform has a critical threshold of enablers and their integration. This is similar to how the mouse, bitmapped display and window paradigm transformed computing in the 1980s; or how the touchscreen, kinetic design motif and mobile broadband made the smartphone ubiquitous a decade ago.

This “Internet of Health” or “Web of Life” is the next iteration of our technology platform, following on from the basic Web (and search), communications (Skype, WhatsApp, Facebook), and control systems (Nest, Tesla). What is missing are the protocols and platforms that would allow different wellness products to communicate with one another and share information in a way that makes the whole far more than the sum of its parts.

This integration needs to occur in both a “horizontal” form, between wellness products, as well as in “vertical” form, such as between clinical use and wellness use. For instance, the Fitbit is being used in over 100 clinical trials, and there is a growing requirement to integrate and merge self-test data, activity data, and clinical scans and monitoring.

Superhuman powers for normal people

So, who gets to use these clever technologies first? The early adopters are often found among more extreme user groups: astronauts, the military, first responders, or athletes. There are also pioneering uses among the more vulnerable populations of pre-term births, elderly care, and those with degenerative diseases of the nervous system.

The elderly highlight a key barrier, which is privacy. They don’t want to be spied upon, or to have their vulnerability and need become obvious or burdensome. This requires working with indirect measures of social life and activity level to create a continuous wellness metric.

There are also segments of society who are socially disadvantaged or oppressed. In this case these tools can be a substitute for alcohol and pills. When people are treated as people, not patients, it may be better to prescribe augmented meditation and electronic mood improvement rather than off-label Xanax.

What all these groups have in common are cycles of stress, recovery and activity, such as when an athlete gets an injury. There is then a need to enter into a different fitness and nutrition regime in order to regenerate his or her performance and competitiveness. The lessons from higher-stress environments can be mass-reproduced for all of us.

Cui bono? Et cui debitum?

“The answer is ‘money’. Now, what is your question?” quips an industry colleague. Who will stump up the cash for this new industry?

Consumers themselves will drive a significant level of direct spending. US millennials already spend ¼ of their income on wellbeing products and services. These will increasingly become subscription-based.

The new “pay as you live” model will only be able to charge based on impact and outcomes. If you want consumers to pay (directly or indirectly), you have to be able to prove that you are helping people. This is challenging: you live or die based on your reputation, as you cannot patent a lifestyle or diet.

The healthcare industry will also undergo its own, slower, transformation from ‘fee for service’ to ‘fee for value’. This will, eventually, shift the money from quantity of procedures to quality of prevention. We have seen some early signs of this with the removal of payment for readmissions in the US following surgery.

A large part of the economic burden will be taken up by corporations and employers, both in their role as health insurance buyers, but possibly more importantly as creators of high-performance teams and workplaces. For instance, it was found that bank employees make 50% fewer errors when not stressed. The techniques the military use to manage work pace and energy left can be transposed to the workplace.

The insurance industry is both best motivated and placed to exploit this Wellness as a Service opportunity. The evolution of life insurance will extend it to include wellness during life, and will more resemble the existing wealth and asset management businesses.

Finally, these platforms will monetise data and insight directly. This is similar to platforms like the Sony PlayStation, where 10% of its revenue is usage data.

An economy dependent on sickness

The development of Wellness as a Service faces many barriers. The core one is its relationship to the existing ‘sick care’ industry. Should it exist as an extension of healthcare, like how banks offer payment protection plans? Or conversely more like an “over the top” player, more like how Skype treated telcos? Or as some middle space in the fringes, rather akin to PayPal’s relationship to the finance industry? Or all the above?

Who will “own” the relationship with the consumer customer? Will it be physicians as “trusted advisors”? A personal health coach? Or a more direct kind of relationship with a consumer lifestyle brand?

How will the ‘sick care’ industry respond to its essential failure being highlighted due to competition? Only 2-3% of the US population has all four wellness indicators (smoking, eating, activity, sleep). That’s a pretty catastrophic outcome, and a political hot potato.

Looking wider, what happens when the Wellness industry provides the incontrovertible data to prove the quasi-genocidal impact of the food industry? Its promotion of sugars and high carb diets likely dwarfs the impact of smoking on health, and the corn lobby is a strong one. What happens if and when a Coke is the new Marlboro?

How will the Wellness industry overcome the effects of marketing designed to undermine us in our self-control challenges? How will it deal with a government system that is often apathetic and indifferent to the health and wealth crisis facing much of the population?

Can physicians with no training in wellness make the transition to ‘well care’? If not, will they become ceremonial and decorative items? How will robohealth change attitudes to costly professionals in white gowns dispensing accredited but faulty advice? Will a new class of “lay priests” emerge to dispense medical help to the needy?

What will happen to cultural attitudes over who is legitimate to deliver health and wellness guidance? Will we endure a lasting “carbon chauvinism” in our need for human contact? Or will people be able to relate to machine-generated wisdom offered via a synthetic voice or touch?

How will we skill and staff up this new industry? What if the core skill for ‘well care’ is computational linguistics? Why are we teaching kids to code when what matters are data science skills? And if biology has anointed Python as its programming language, why are we teaching C++ or Java?

And finally, how can we inspire a new generation to take on the massive task of redesigning our technology and economic system to meet human needs? How can we evolve biology from being an abstract subject learnt from textbooks to an experienced one from hacking one’s own body?

The future of wealth is wellth

The size of the prize is enormous if we can develop a wellness culture. Our lives today are primitive in comparison to what is possible; the emotionally violent dark age we unwittingly inhabit can come to and end.

The future “life GPS” will help us to navigate around the reefs that shipwreck too many lives today. A “Waze for health” will help us to cope with a fast-changing world. We are what we pay attention to, and the symbiosis of computers and humans will increasingly design for human potential and direct that attention wisely.

As the healthcare industry evolves from its present impairment model, we will see more primary lifestyle interventions, and far fewer secondary ones of surgery and drugs. This will alleviate a lot of human suffering and iatrogenic damage.

Each of us will be offered new freedom to define our life goals. You want to be more like a weightlifter or cyclist? A top intellectual or a fine actor? Based on our goals we will be guided towards maximum performance.

As individuals, we will come to regard personal nutrition as the norm, irrespective of genetics. Every restaurant menu will be digital and interactive, even if printed superficially on paper. Indeed, our whole environment will become quantified.

This will cause a cascade of secondary change, deeper than that of personal wellbeing in isolation. The placing of stores, businesses, footpaths etc. will all become optimised. We will personalise our homes and communities to our culture and values. Hyperwellbeing is inherently hyperlocal.

The end game of Wellness as a Service will be a transformation of society, as the wider structures that make us sick, like insecurity due to a lack of assured basic income, come to be questioned and replaced. The possible impact of hyperwellbeing transcends hyperbole.

For the latest fresh thinking on telecommunications, please sign up for the free Geddes newsletter.