AI Special Feature: “The Cyborg Will See You Now”

 

How Artificial Intelligence Could Infuse Genuine Humanity Back into Healthcare

 

 

By Fabio Gratton
Chief Listening Officer
inVibe Labs

 

Michael Spitz
Chief Strategic Storyteller
Blitz Strategy

 

 

 

 

The Good Old Days

Once upon a time the “family doctor” was a trusted and beloved professional who arguably dispensed as much emotional support as medical advice. The physician-patient relationship was characterized by a spirit of “benign paternalism” where, in exchange for giving physicians complete control at the point of care, patients received sufficient office time and personal attention to feel that their healthcare needs were being satisfactorily met.

Since the 1950s the physician-patient relationship has been shaken by seismic changes in medical science, marketing/regulation, and technology. The economic prosperity of the post-war era saw the birth of reproductive, cancer, and cardiac treatments culminating in ulcer, SSRI, and statin blockbusters. “Detail men” transformed into an army of pharmaceutical sales reps for HCPs, while DTC advertising came into its own after the first broadcast TV ad in 1982.

Access to all this information empowered physicians and patients, transforming their relationship into a more egalitarian one as the flood of data became a tsunami with the advent of digital. The “ePatient Revolution” over the past two decades has encouraged patients to take control of their healthcare, giving marketers unprecedented opportunities to find and engage both professionals and consumers with increasingly personalized and contextually relevant content.

But the benefits of digital health are still offset by systemic problems of the healthcare system, buckling under the pressure of too many patients, too few physicians, and a steadily deteriorating point of care experience. Ironically enough, the culprit is not only dwindling resources, socio-economic disparities, and fragmented infrastructure, but the electronic medical records that were put into place to ostensibly heighten efficiencies and encourage growth.

Designed and developed to facilitate billing, EHRs have had the unintended yet foreseeable side effect of destroying human interaction between physician and patient at the point of care. All of us are now accustomed to appointments that are not only rushed, but characterized by impatient and exhausted physicians staring into computer screens, barking out questions, and frantically typing on keyboards. Medical school graduates have become data entry clerks.

Worst of both worlds, the physician-patient relationship not only remains stubbornly paternalistic, but is now mostly transactional. As digital marketers, we have opportunities to target and educate doctors through EHRs, but that begs the question of why the proven power of personalized medicine has been subsumed by clunky digital platforms that turn doctors into robots. If the miracle of digital got us into this mess, then its infinite potential can get us back out.

 

From Hype to Hope

As savvy healthcare digital marketers, most of us are familiar with the Gartner Hype Cycle. The oscillating curve plots the trajectory of emerging technologies from their birth as an “innovation trigger” and up into their hyped “peak of inflated expectations.” Most plummet forever into a “trough of disillusionment,” but some rise into a “slope of enlightenment” and rarer still, into a “plateau of productivity.”

Innovation excites, but it often befuddles and blinds. Calling BS is hard, especially since many of the greatest accomplishments over the past couple decades have been so unexpected and surprising. While pundits predicted world domination for Yahoo! and bankruptcy for Jeff Bezos, Facebook now has over 2 billion daily users and Amazon pushes a market cap of over a trillion dollars. Healthcare predictions are even trickier, because the stakes couldn’t be higher.

Healthcare is also data-intensive, making the industry particularly ripe for digital transformation. To give you an idea as to how much medical data is out there, an “exabyte” is literally one billion gigabytes: in 2013 the medical community generated 153 exabytes; in 2020 it’ll generate 2,314 exabytes – that’s 2.314 trillion gigabytes – out-producing the current storage capacity of the world’s servers. Given all that data, it’s no wonder MDs have become clerks!

Every challenge encourages a solution. In this instance, the world needs one to figure out how to best assimilate, analyze, and activate a mind-blowing amount of health data. At the heart of personalized medicine is the quantified self, a way of meaningfully measuring the 37 trillion cells in the average human body; and at the core of personalized marketing is digital engagement, a way of effectively targeting millions of consumers eager for the most accurate and actionable content.

The convergence of healthcare, communications, and tech is upon us, in what Eric Topol has christened “the creative destruction of medicine.” With that special focus on big data, which innovative goodies along the Gartner Hype Cycle catch our eye as potentially viable for fixing what’s broken and ultimately improving the healthcare experience that counts the most – that foundational and essentially emotional relationship between physician and patient at the point of care?

We’ve all heard the buzz words: cryptosecurity and blockchain; quantum computing and neural networks. Some innovations – including virtual and augmented reality, surgical robots, and smart pills – have already made their way to the plateau of healthcare productivity. Most recently and arguably most importantly, “Deep Learning” has tipped over the peak of inflated expectations. What is that, exactly, and how might Artificial Intelligence transform healthcare?

 

“I’m Sorry Dave, I’m Afraid I Can’t Do That…”

AI is tough enough to understand, let alone AI for healthcare. Perhaps the best way to describe what the technology is and can do, is to reveal what it isn’t and can’t. Let’s start by doing what doctors shouldn’t and go to Wikipedia for a crowdsourced definition: “Artificial intelligence” is often used to describe machines (or computers) that mimic “cognitive” functions that humans associate with the human mind, such as “learning” and “problem solving”.

The key here is “problem solving,” which within a healthcare context is central to the Hippocratic Oath of promising to use treatment (ideally the best one) to heal the sick. Tweaking the definition somewhat and looking at it through the lens of medicine, AI is essentially using digital technology to help address medical needs such as properly recognizing symptoms, diagnosing a condition, prescribing an optimal medication, etc. – essentially mirroring the duties of a doctor.

Clearly machines haven’t replaced physicians, and as we’ll soon see, they can’t and shouldn’t. We all keep hearing about this alleged tension, though, so where does it stand and what is it all about? To succinctly answer that and better understand how AI can and will actually impact healthcare, the physician-patient relationship, and our capabilities as marketers, let’s take a quick look at two peaks of inflated AI expectations, and how the industry recalibrated for future success.

 

The Brain-in-a-Box Failure

Research into artificial intelligence formally began at Dartmouth College in 1956 and tapped into real minds including Allen Newell and Marvin Minsky. Dovetailing off the world war’s astonishing advances in computers, the AI geeks started where John von Neuman and Alan Turing left off: they would program machines that wouldn’t simply execute commands, but think, feel, and interact with the world in ways essentially indistinguishable from flesh and blood human beings.

 

The Machine Learning Renaissance (and Crash)

Although Minsky and his crew’s software played checkers, spoke English sentences, and did cool math, funding was pulled by the mid-70s because little progress was actually made toward a “sentient machine.” But all was not lost: by the early 1980s their goals became more practical and computational power exponentially increased, triggering a new era of real wins where “AI” did what machines do best: analyze mountains of data and recognize patterns.

Suddenly giant computers loaded with “AI” were doing the seemingly impossible: in 1997 they beat Gary Kasparov in chess, and by 2011 defeated the world’s best Jeopardy! players. Although they weren’t “human” in the way science fiction movies had characterized them – neither having the empathic neuroses of Kubrick’s HAL-9000 nor the strategic viciousness of Cameron’s SkyNet — “AI” nonetheless began to outperform humans in specific areas of expertise.

 

Healing Healthcare with AI

These unexpected and phenomenal successes, thanks to moving the goal post, focusing on pattern recognition, and increased digital processing power, inspired innovators to try and transform healthcare. As we’ve seen, healthcare is not only data-intensive, but most if not all problems in medicine can be solved through information processing: the input is patient biometric data; the analytics is differential diagnosis; the output is treatment recommendations and follow up.

Trend-setters and paradigm-shifters leaped at the chance. Most notably the same “Deep Blue” computer that defeated chess grandmasters and celebrity game show players fueled IBM Watson and its special focus on healthcare. Their immediate goal: help oncologists read through and process thousands of journal articles to stay current with the literature and integrate the very latest data into actionable and effective diagnoses that result in the best cancer outcomes.

But as Topol describes in his recent Deep Medicine: How Artificial Intelligence Can Make Healthcare Human Again, the mountains of data in the medical corpus and across electronic health records network is all unstructured and therefore beyond the machine learning capabilities of current AI systems. Although these systems can still outperform humans when it comes to pattern recognition, they still have a long way to go before human physicians should feel threatened.

In fact, the theme of Topol’s book is that human physicians need never feel threatened by machines. Successful use cases and startup companies abound for AI already entering the plateau of productivity in healthcare, including drug development, patient recruitment, differential diagnosis, targeted therapies, and more. Throughout all these proven examples the technology supplements and enhances the still-valued and indispensable role of the medical professional.

Rather than replace doctors, Topol sees AI as increasing productivity in a way that can ultimately enhance the physician-patient relationship. “What I’m most excited about is using the future to bring back the past: to restore the care in healthcare. By giving both the gift of time to clinicians… and empowerment to patients.” The Terminator was a perfect killer, part human and part machine. Dr. Cyborg will heal you now, the machine enabling the human to return and thrive.

As healthcare marketers, we can perhaps best understand Artificial Intelligence in terms of Animal Instinct: precision targeting algorithms, programmatic demand side and data management platforms offering increasingly sophisticated means to better understand, find, engage, and inspire our audiences. As the technology becomes more “intelligent,” we become more adept at getting to the heart of what patients and professionals need in order to extend and improve human life.