AI moves from hype to reality.
HEALTHCARE IS NO STRANGER TO THE infamous Gartner hype cycle. As a journalist covering the industry, it’s sometimes intriguing to figure out where an emerging technology is on this graph, which charts the typical cycle from moment of innovation for an new technology until it starts producing results.
Artificial intelligence is making its way along the ups and downs of the hype cycle. Right about now, AI appears to be near the peak of inflated expectations, perhaps slipping over the precipice onto the downward slope toward the trough of disillusionment.
Most HIT executives know the pattern— they’ve seen overhyped expectations for new technologies, where hopes and dreams obscure the hard work that lies ahead. For example, high expectations surrounded the Internet and electronic prescribing in the late 1990s, but it took years before those technologies matrured. Only after early dreams are dashed does the real work of innovation, and achieving benefits, set in.
I’d like to suggest that there’s a concurrent cycle of exaggerated hysteria with new technology. I saw this during the past two RSNA conferences regarding AI. In 2016, AI was the monster to be dreaded, because radiologists feared it would obviate the need for them to interpret images. Last year, however, those fears were nearly gone— and nearly every vendor insisted that it had incorporated AI into its products.
Expectations for AI in healthcare are high, particularly when it comes to expectations for achieving measurable improvements in patient care and financial performance. However, reality is hitting home—there are many stumbling blocks to overcome before advanced decision-aiding technology is widely adopted, reports Linda Wilson in our cover story for this issue.
In her reporting, Wilson notes that Paul Chang, MD, professor and vice chairman of radiology informatics at the University of Chicago School of Medicine, says “We are not at the solution level yet; it is a long way between initial promising results in the lab and true clinically validated solutions.” But early work on using and benefiting from AI is under way, she finds. Her story begins on Page 16.
AI is competing for IT resources with other IT initiatives, and one of those involves information security. Attacks on providers are commonplace, and instances of ransomware attacks have been increasing over the past two years. In this issue, starting on Page 20, Managing Editor Greg Slabodkin takes a microscope to examine the attack last year on Erie County Medical Center. Recovery from the attack was extensive, time-consuming and expensive, and ECMC used many workarounds to ensure care could be delivered safely to patients.
Slabodkin’s account of the story is insightful, and we applaud ECMC for allowing us to tell its story. Such openness to share information with others in the industry is crucial in allowing provider organizations to inform others about hacking gambits and best practices at defense and recovery.