everythingpossible - Fotolia
- Scott Robinson, Lucina Health
I'm a fan of cognitive computing. My affection for analytics, artificial intelligence and big data is perpetually on display. But cognitive computing -- the mashup of the way the human brain and computers work, plus large dollops of analytics and AI -- is swiftly building a bridge between humans and machines.
My enthusiasm is largely attached to cognitive computing in healthcare in my own field -- healthcare analytics -- where the sheer pace of new knowledge has made analytics and AI-driven methodologies not only a novelty, but also a mission-critical necessity.
Specialization hurts patient relationships
I like to keep up with how cognitive computing is doing in the marketplace and came across a TedMedLive Talk by Basit Chaudhry, an MD specializing in the design of clinical service delivery systems for chronic disease care. His talk is a little dated, 2013, but what he had to say then applies very much to the state of today's healthcare industry.
"It's not possible [anymore] to try to keep up with everything that's going on, ... for one person to fit everything known in medicine inside of their head, regardless of how talented they are," Chaudhry said. The "breathtaking growth of medical knowledge," he said, has forced clinicians to specialize so they can cope with their own cognitive limits and focus on a subset of all things medical. As a result, the quality of clinical care has suffered, according to Chaudhry, who is the founder of Tuple Health and a former IBM medical scientist.
"At the heart of the doctor-patient relationship was a narrative," he said. "It was a story that a patient would initially tell a physician, and at its best, it was a story that would be created by both of them as equals. The narrative nature of medicine has been lost." In the era of specialization, he added, "it's easy to lose sight of the entire individual."
Chaudhry pointed out that as medical knowledge expands exponentially, the available facts per decision a clinician must cope with are rising as well, while the time per patient during clinical visits is rapidly shrinking, mainly due to economic factors. This situation, he said, makes it impossible for the clinician to keep up because human cognition will remain constant, despite these dynamic shifts in clinical care.
Cognitive computing in healthcare a big data prescription
On top of this trend is the era of big data. All this new, prodigious, unstructured data is arriving in various forms, freely created by humans rather than traditional medical information systems and, therefore, not amenable to querying or any other insight-generating methodology. And it's here, Chaudhry said, that cognitive computing can play a key role.
IBM's Watson, probably the world's best-known cognitive computing system, has demonstrated the potential for building bridges between the oceans of amorphous, unstructured data that contains human narrative and the needs of the healthcare clinicians and other professionals who are drowning in data. Cognitive computing in healthcare can distill unstructured input into interactive algorithms, and applying machine learning can help yield a small set of potential responses. We all saw this play out in 2011, when Watson appeared on Jeopardy! and easily defeated the two winningest contestants in the long-running game show's history.
Chaudhry imagined a hypothetical clinical setting in which Watson would parse language to extract ideas, map the ideas to medical concepts and link the concepts to deliver a small set of potential answers. Those answers would then be ranked after being validated via patient history, textbooks, best practices, diagnostic tests, journal articles and so on. Then out pops a recommendation that awaits a human decision.
But this scenario of cognitive computing in healthcare is more than hypothetical. Chaudhry pointed to Memorial Sloan Kettering Cancer Center in New York, where Watson collated thousands upon thousands of individual patient history details. The goal was to achieve a doctor-patient narrative that would enable oncologists to craft highly individualized care management plans. That was a perfect application of cognitive computing that yielded very positive outcomes.
Cognitive computing in healthcare really does have the potential to bridge humans and machines -- a sweet spot where computers and the human brain can truly merge.
Dig Deeper on Enterprise applications of AI
Potential for AI tools in healthcare lies in the long game
Digital transformation technologies power CX makeovers
Data cognition engines can save millions in big data costs
APAC healthcare providers reaping digital transformation returns