Everyone knows Moore's law - that the number of transistors on a semiconductor and hence their performance will double every 18 months. It is the reason why the iPhone in your pocket has more computing power than the world’s fastest supercomputer from twenty-five years ago. What people may not realize is that the doubling of computation has been going on since at least the 1850's through more than five different technologies from Charles Babbidge's mechanical difference engine, through the electromechanical calculators used in the 1890 U.S. census, to the Colossus, the Allies code-breaking, vacuum tube computer, past the transistor-based machines used in the first space launches, all the way to the silicon semiconductors that Moore's Intel was manufacturing, and that power of current generation of computers. This phenomenon was described by Ray Kurzweil, who went on to realize that the doubling of performance was not just isolated to computation, but was a function of the evolutionary nature of technologies. He called this the Law of Accelerating Returns that results in what others call exponential technologies.
"Evolution applies positive feedback in that the more capable methods resulting from one stage of evolutionary progress are used to create the next stage. Each epoch of evolution has progressed more rapidly by building on the products of the previous stage." – Ray Kurzweil
A business or industries capacity to harness this law of accelerating returns depends on the degree to which it is completely digital and information enabled. As Salim Ismail writes in “Exponential Organizations”:
"Once any domain, discipline, technology or industry becomes information-enabled and powered by information flows, its price/performance begins doubling approximately annually. - "Exponential Organizations" - Salim Ismail
This is the point that we have arrived at in healthcare. For over 30 years, we have been digitizing health. While progress has been slow and frustrating as we have taken existing processes, many of which were never optimized, and digitized them; we are almost through the digitization phase. Parallel to this, consumers are now carrying supercomputers in their pockets and we are all connected to the internet all the time. Understanding that most information is now digital, that all the end-points of healthcare interactions are using digital devices, and connectivity is ubiquitous, then you can see that healthcare is on the precipice of becoming information enabled and subject to exponential forces.
If you lead a healthcare organization, the realization should be even clearer. The efforts to put EHR’s in place and the struggles to digitize physician workflows were not an end unto itself but are the beginning of a digital strategy in health.
What I mean by digital strategy, is one where you are able to leverage digital infrastructure to transcend the limitations of:
Is the interaction best delivered
When each interaction is viewed through these lenses, and every interaction is measured and that information is fed back to improve the next, then healthcare has truly become information enabled. Once healthcare organizations start this process, they will begin to deliver exponential improvements in health outcomes. The changes will be small at first, but they will continue to double every couple of years, transforming healthcare.
Embracing a digital strategy for health, allows for audacious goals like Lucien Engelen’s vision for Radboud UMC in the Netherlands to remove 70% of services from the hospital in the next 10 years. These services will either be done elsewhere, at home or in the community, or they will not be done at all because of prevention strategies that address the health need upstream.
Leaders in healthcare organizations that embrace and develop real digital health strategies will have massive transformations in the next decade that will improve health outcomes in unimaginable, exponential ways; sort of like those who imagined twenty-five years ago that the world’s fastest computer would one day be in a device they carry in their pocket.
Samuel Hoskins is a 52 year old man, recently diagnosed with Type II Diabetes. His story is not atypical for many in Western cultures. Over the years, his weight increased while his activity decreased. At first it was not noticeable, relaxed fit instead of slim cut, but now on this October night leaving his family doctor, Samuel is now a diabetic. His doctor, who had known him since his early 30’s, was knowledgeable and reassuring. Diabetes was manageable, with effort it could be controlled and complications avoided. The first step for Samuel was to start a new medication Metformin. His doctor was also concerned about his blood pressure and cholesterol. “If they don’t improve, we have meds that will help” – for Samuel this part was more ominous than reassuring.
In medical terms we would say that Samuel has diabetes, hypertension and borderline cholesterol and he is being treated by his primary care physician, but that does little to describe the actual work being done. To better understand we need to introduce a concept: “health jobs”. I’m not talking about employment, doctors, nurses, pharmacists, etc. When I use the term health jobs, I am referring to the specific tasks that are being done as part of the provision of care for a person. Health jobs are the tasks that patient’s need done. The health jobs themselves may or may not be specific to a disease or diagnosis; they may be done by health professionals for patients; or they may be done by patient’s themselves. They may be related to diagnosis and treatment, but also to prevention.
When looking at health jobs, it is useful to step back and examine the goals a particular patient may have. For Samuel and his diabetes, his goals are straightforward, and at this stage are all about avoiding the bad things that diabetes can bring:
Looked at from this perspective of goals, Samuel’s health jobs become simpler to understand. First and foremost, he needs to control his blood sugars as that has been definitively shown to reduce both microvascular and macrovascular complications as well as infections and co-morbidities. Secondarily, he needs to monitor and control his risks for these complications. And finally he will want to screen for and diagnose early any complication or co-morbidity so that it can be optimally managed. The table below outlines his specific health jobs, as well as whether these are jobs for Samuel or for his team of health professionals.
Type II Diabetes is of course a disease, and the goals and management are correspondingly healthcare oriented, yet within the array of jobs to be done to reach Samuel’s goals, many fall within his capacity and capabilities. Of the jobs performed by his health team, I see four underlying reasons why they are done by professionals and not by Samuel:
These four reasons are reasonable, broadly accepted and widely applicable across all health jobs. Given the potential adverse effects and interactions, most accept that healthcare professionals should be responsible for prescription medications as well as the administration of immunizations. Having healthcare professionals required to order screening, monitoring and diagnostic tests makes sense from a cost-control basis but also because of the expertise necessary to interpret the data from tests in the full context of a patient. Many diagnostic and therapeutic procedures require a level of technical ability only mastered after 15+ years of post-graduate training so it makes sense that consumers do not attempt these things at home. Finally, the role of trust and confidence cannot be underestimated when making important decisions about our health and wellness.
The emergence of the digital health stack promises to disrupt this logic directly by leveraging artificial intelligence to provide consumers with personalized insights allowing them to correctly interpret, and safely act upon information, at marginal costs that over time will approach zero.
For Samuel, one can imagine the application of the digital health stack for diabetes pharmacotherapy, providing a learning algorithm that uses a vast array of data relevant to him:
This algorithm would compare this information to the complete set of medical literature as well as vast datasets of outcomes from other diabetics along with Samuel’s previous data in order to make a therapeutic recommendation. This insight would be precise and personal and probably significantly safer than anything coming from a healthcare professional relying only on traditional means. With the emergence of 3D printing for medications, Samuel’s dosage could even be personalized with delivery to his home by drone. His response to the medication would be tracked immediately and fed back so that the algorithm could adjust and learn. Over time the efficacy of the algorithm will improve and the cost will come down.
The real promise for diabetes with the emerging digital health stack lies in leveraging learning algorithms to predict and prevent the initial onset of diabetes. In many ways this will be the path forward; shifting care from a healthcare, disease orientation into the wellness, betterment perspective of prevention and optimization. The same approach used for the pharmacotherapy of diabetes could be used to direct activity and lifestyle changes with the potential use of therapeutic nutrition and supplements.
Beyond diabetes one can see some common, broad categories for Health jobs:
From this list we can see that learning algorithms will play a key role in shifting health jobs from healthcare professionals to consumers because they will improve safety by having more complete information and built in feedback monitoring of any and all interventions. The technical ability required to perform health jobs will diminish as it becomes consumer grade with computation and machine intelligence providing better quality assurance and interpretation than any individual technician’s capability. The marginal cost of the underlying exponential technologies will approach zero. Finally, the measurement of health outcomes and the feedback loop necessary will allow consumers to trust and have confidence in learning algorithms over time.
For Samuel these technologies hold the promise that he will be fully in control in of his health. He will have the tools to understand, manage and change the course of his diabetes. To do this he will be willing to share his data with the applications, algorithms and services necessary to achieve his goals. When it comes to avoiding the complications of diabetes, privacy fears seem like remote risks.