Artificial intelligence: what we know so far about how it works for health

From the research lab to the bedside to the back office, AI and machine learning are transforming the ways we think about care and how we develop treatments.
By Mike Miliard
08:00 AM
Share
Artificial Intelligence

All November long, Healthcare IT News and other HIMSS Media brands have been focused on artificial intelligence. AI, of course, is a complex proposition for many. Some are wildly excited about what it might augur for humanity. Some are much more skeptical – if not downright wary – of what the future of learning machines might portend.

Just look at the discordant war of words that erupted this past year between two of the tech world's towering titans, Mark Zuckerberg and Elon Musk.

"With AI, especially, I'm really optimistic," said Zuckerberg. "In the next five to 10 years, AI is going to deliver so many improvements in the quality of our lives."

Musk, for his part, looking a bit further into the future perhaps, took a darker view. AI is the "biggest existential threat" we face, he said, and a "fundamental risk to the existence of civilization."

For now, at least, let's just agree to disagree?

But some of these concerns are surely playing out, in miniature, within healthcare. Is AI technology coming to supplant clinical and imaging jobs? Or merely to supplement them, a benevolent augmentation to existing processes that, properly harnessed, can enable huge advances in how care is delivered? What about ethical concerns?

Some of these questions will surely continue to iron themselves out in the years ahead. But so far this month we've highlighted some signs of exciting real-world progress in clinical and operational settings across healthcare that point to big progress and a bright future for AI.

But that's not to say that its progress comes without hurdles, of course. (Or, at times, absurdities.) Tune in next week for advice on the big challenge of the ensuring your data is optimally governed and groomed to make the most of machine learning. And before the month is over, keep a look out for HITN's first attempt at art criticism, as we team with Harvard Professor and Cyft CEO Len D'Avolio to examine the eye-bending pleasures of bad AI stock art!

In the meantime, here are some things we've learned so far this month about artificial intelligence in healthcare.

Most hospitals and health systems have big plans for AI. And big expectations, too. But many of them are still not well-positioned to capitalize on it, we showed. "Healthcare executives expect artificial intelligence to be among the most impactful technologies fueling innovation, but few are crafting strategies to advance emerging AI applications. That means now is the time to not only invest but also take on a proactive role in developing new tools."

It matters what you call it. All machine learning is AI, but not all AI is machine learning. But what about cognitive computing and neural networks? And what do supervised, unsupervised, semi-supervised mean in this context? Augmented intelligence? Adaptive intelligence? We try to explain here.

Persuading executives invest may require some hand-holding. Convincing the C-suite to try a test deployment of a new or perhaps overhyped technology means explaining to execs that there may be substantial ROI in a project whose value may seem intangible at first. But it can be done with the right data points.

It's already hard at work. From routine colon screenings to cardiac care to advanced precision medicine, AI is closer than many realize to changing the outlook for how treatments are developed and care is delivered.

FDA is catching up. After years of lagging technological progress, the U.S. Food and Drug Administration has signaled a new era for its approach to healthcare AI and has already given the nod to many clinical algorithms.

Payers use it too. The biggest breakthroughs for data crunching are in more sophisticated predictive machine learning algorithms, of course. "The applicability and opportunity on the insurers side is fantastic," said one tech exec.

The momentum will continue in 2019 and beyond. And CIOs should be prepared. A new report from IDC shows that in the years ahead, some 70 percent of CIOs will "aggressively apply data and AI to IT operations, tools, and processes" as they work to curtail spending, improve enterprise IT agility and accelerate innovation.

Focus on Artificial Intelligence

In November, we take a deep dive into AI and machine learning.

Twitter: @MikeMiliardHITN
Email the writer: mike.miliard@himssmedia.com

Healthcare IT News is a publication of HIMSS Media.