Artificial intelligence: Savvy hospitals are deploying AI and its technological brethren cognitive computing and machine learning in specific use cases at this point – while industry luminaries are predicting that their advancement will soon start happening more quickly than previously anticipated.

"I've never in my career seen the acceleration of technology as fast as what we've witnessed in machine learning during the last two years," said Dale Sanders, executive vice president at Health Catalyst.

Sanders, it's worth noting, has a U.S. Air Force background working on stacked neural networks and fuzzy logic, which used to be called deep learning, as well as serving as the CIO of both Northwestern University and national health system of the Cayman Islands.

"The rate of improvement happening in machine learning," Sanders added, "is way beyond what Moore's Law is to chips."

Hospitals already deploying AI
As the next generation of both patients and caregivers – including clinicians, doctors, nurses, specialists, even executives and administrators – starts taking a foothold in the healthcare workforce, hospitals looking for a first-mover advantage already know that AI is on the verge of becoming a critical component across the entire organization, and not just IT.

"AI and machine learning are exciting opportunities for us to accelerate," Carolinas HealthCare Chief Information and Analytics Officer Craig Richardville said. "To be successful you have to understand how that will fit within your market and your patient population, and you have to be knowledgeable about how to use it."

[Also: Hospital datacenters: Extinct in 5 years?]

Today, that means picking opportunities akin to low hanging fruit for modern AI capabilities. Carolinas, for its part, is working to develop self-service applications that provide tools to patients for self-diagnosis and self-treatment in very targeted scenarios where the science enables clinicians to understand what the right methods are, Richardville said. 

The hospital is also eyeing AI to capture patient information and bring it into data lakes or warehouses, which is paramount because Richardville said that only 20 percent of relevant patient information for Carolinas clinicians resides in its EHRs.

"Applying more intelligence to that data continues our transition from the art of medicine to the science of medicine," Richardville added.

The revenue cycle is another area ripe for machine learning, according to Stuart Hanson, senior vice president of Change Healthcare.

"Healthcare organizations have started to become more information-centric, and the next level of that is taking a personalized view," Hanson said.

Hanson cited two examples: the ability to predict what is relevant for a particular patient and deliver smart messaging, such as wellness and prevention tips and price transparency, as well as the opportunity to drive down costs associated with useless billing by better understanding how patients interact with various types of payment statements.

"There's clear ROI in the revenue cycle for physicians and hospitals," Hanson said.

While such work among payers and at Carolinas and other leading hospitals is admittedly cutting-edge, Health Catalyst's Sanders is hardly alone in believing AI, cognitive computing and machine learning will outpace the processing power advances that Moore's Law illuminated.

Read more Innovation Pulse columns from Healthcare IT News.

Beyond the futuristic hypothetical
Intel co-founder Gordon Moore predicted in 1965 that the then-current pace of computer chips doubling in power every year would continue into the future; Moore's Law was amended a decade later as processing power was doubling every two years. And depending on whom you ask, that rate pretty much held steady for the better part of 50 years.

Indeed, a lot has happened during the last five decades, genomics advances not least of all. When a company named 454 sequenced DNA co-discoverer James Watson's genome in 2007, it cost $2 million – which was down just a dram from the $3 billion the Human Genome Project spent on its first sequencing in 2003. 

"Now we're down to $1,000, and we'll get to an era of $100 per genome," said Bryce Olson, global marketing director of health and life sciences at Intel. "It's going to become the next big thing."

AI is exploding quickly. Healthcare providers will be able to diagnose disease by DNA in the near future, Olson said, and the industry is on the verge of making the technologies faster, better and cheaper.

"We're seeing it right now in the genomic space and with machine learning algorithms," Olson added. "It's a lot faster than Moore's Law." 

Twitter: SullyHIT
Email the writer: tom.sullivan@himssmedia.com