When it comes to artificial intelligence (AI), it can be hard to tell the hype from the reality. Tech reporters tend to cover it with breathless enthusiasm. Venture capital floods AI-oriented firms at an astonishing rate. Amidst all this, healthcare leaders might be wondering: just how much hope should we invest in this powerful new technology?
Steve Jackson, president of NRC Health, brings some clarity. In this Q&A, he discusses where AI stands in healthcare, where it’s going, and what it means for the industry’s future.
Where does AI technology stand in the healthcare industry today?
Most of the big healthcare AI innovations are medical. It’s not an exaggeration to say that AI has the potential to revolutionize the way we research, diagnose, and treat several conditions. Today, algorithms are already outperforming radiologists at spotting malignant lung masses, for example, and AI is already guiding researchers in how they select candidates for clinical trials.
This trend is likely to accelerate, too. Data collection is only going to get better; hardware performance is going to improve; data storage is going to get cheaper; the algorithms we work with are going to get nimbler and cleverer.
Suffice to say, exciting things are coming. The precise improvements are impossible to predict, but the pace and momentum of the AI revolution were never in doubt.
Why don’t we hear as much about AI applied to the experience side of care?
AI does best with objective data. That’s why the clinical side of things is so inviting for AI innovators. Crunchy, quantitative data like lab-values and vital signs are ripe for algorithmic processing.
Experience data are trickier. They revolve around patients’ feelings, which aren’t always easy to capture and quantify. Without that ready-made data, it’s harder to create AI products that can make use of what we observe from patients’ experiences.
That’s changing, however. We’re getting savvier about how we understand what patients go through, and we’re able to bring new models of machine learning to subjective-experience data. It’s cutting-edge technology.
What’s an example of experience-facing AI technology in use today?
The best example is probably Natural Language Processing (NLP). Basically, NLP is an algorithmic process that “reads” verbal information and uses probabilities to classify what it means.
It’s a technology we all use every day. Inbox filters, for example, flag emails that contain the word “money,” because NLP algorithms have found that such emails are highly likely to be spam.
Obviously in healthcare, the stakes are a little higher. A type of NLP process called sentiment analysis can tease out invaluable qualitative information from patient comments, including how they felt about their encounters. For instance, if a patient’s comments contain the word “forever,” chances are good that they have a complaint about their wait times.
The advantage of NLP, as opposed to employing humans to sort through patient comments, is that they can do it en masse, and instantly. This enables health systems to spot problematic trends much earlier than they would without NLP on their side.
What experience-oriented AI innovations are coming down the road?
Two different technologies really excite me.
One of them is predictive analytics. These are algorithms that use past data to make predictions about the future. Banks are using these algorithms all the time, to judge if a borrower is likely to default on a loan.
The potential for predictive analytics for care experiences is enormous. Imagine if your health system could anticipate patient needs, instead of just reacting to them. What if you could foresee, on a population level, when you’d need to expand clinical capacity, and when you could afford to trim it down? Predictive analytics could make that possible.
The second technology with undeniable potential is personalized experience engines. Think of how Netflix or Spotify makes recommendations to you—and picture how helpful that could be for patients.
With such a tool, your system could create a concierge care-experience for any patient who walked in your door. It would be mass personalization.
It almost sounds utopian. And indeed, this kind of technology is of a much higher order of complexity than what we have today. But there’s no doubt in my mind that such innovations are coming—likely in the next five years.
What does all this mean for the healthcare workforce?
Discussions of AI inevitably rouse some anxiety. People in every industry have legitimate fears of being “automated” out of their jobs. In healthcare, though, this is unlikely to come to pass.
AI will always be a supplement for human attention. It can never replace it. The relationship between a patient and their provider is intimate and personal; no machine could ever capture it.
This means that, while AI processes will be able to tell us what to pursue, it will come down to us, the very human healthcare workforce, to execute on the insights that AI uncovers. Mass personalization means that we’ll all have to do a little more to create the kinds of experiences that our patients deserve.
News from our Partners