Skip to main content

Using AI to Help Clinicians Get a Clear Picture

Analysis  |  By Eric Wicklund  
   June 20, 2024

Cleveland-based University Hospitals is partnering with Aidoc to help radiologists screen CT scans and make care delivery more efficient

Clinicians dread the missed spot on a CT scan or X-ray that leads to a serious health concern. Now health systems are using AI to make sure those mistakes don’t happen.

University Hospitals recently announced a partnership with AIdoc to deploy the company’s aiOS platform across all 13 hospitals and dozens of outpatient facilities in the Cleveland-based health system. The technology aims to assist clinicians by giving them another tool to analyze images.

‘[We’re] looking to see if we’re finding things that we would have otherwise not seen,” says Donna Plecha, MD, the health system’s Chair of Radiology. “We work with AI – it is not replacing our reads. And I think most studies that look at AI with a radiologist, that combination usually does better than either one by itself.”

The distinction—is AI artificial or augmented intelligence?—encapsulates both the promise and the peril of the technology, which has drawn comparisons for its effect on healthcare to both the printing press and the Terminator. Advocates say AI will work best as a tool that clinicians can use to improve their work and their workflows, rather than a replacement for a doctor or nurse.

Plecha notes the difference, saying clinicians will always be reviewing AI output for accuracy. She says the presence of false negatives and false positives in early AI results supported that position.

“I think they’re realizing how careful they have to be and not believing everything that AI is marking,” she says.

As for the potential, UH officials point to the opportunity for AI to pick up on infinitesimal aspects of a CT scan or X-ray that might bypass the naked eye. That tiny spot could be a sign of a pulmonary embolism, aortic dissection, vertebral compression fracture, or pneumothorax. Identifying those and other acute health concerns early means the patient is moved more quickly to the appropriate care provider and treated more quickly and efficiently.

“The technology identifies both expected and unexpected findings, helps physicians prioritize urgent cases, and ensures all flagged conditions are reviewed by the care team,” the health system said in a press release announcing the partnership.

Plecha says the health system will review all the data collected by the AI platform for accuracy and outcomes before expanding the platform to other departments and use cases. That review process will also help clinicians better understand how to use AI and what to look for.

Aside from improving accuracy and care team efficiency, Plecha says the tool will also help University Hospitals make the most out of its limited supply of radiologists, addressing workforce shortages that are plaguing health systems and hospitals across the country. It will, she says, enable radiologists 9and, eventually, other clinicians) to work with more confidence and at the top of their license.

The idea of using AI to improve workflows isn’t new. Texas-based CHRISTUS Health, in announcing a partnership this week with Abridge to implement a clinician conversation tool, noted the effect on “cognitive load,” or the amount of mental effort needed to complete a task.

According to CHRISTUS officials, the AI tool helped reduce physician burnout by some 78% during a pilot earlier this year. With the AI tool, they said, physicians were under less stress and were able to perform their task better and more efficiently.

“I feel much less distracted with patients since I can focus on the conversation and history without pausing to take extensive notes or re-ask questions I missed during notetaking,” Myriah Willborn, MD, a family medicine doctor at the CHRISTUS Trinity Clinic in Corpus Christi, said in a statement issued by the health system.

The concern, of course, is that clinicians become too reliant on the technology, expecting it to be perfect and catch anything they miss. That’s where continuous review comes into play, along with the understanding that clinicians always have the final say in care and are using AI only as a tool to improve their decision-making.

To that end, Plecha says she sees a future where AI not only reads an image, but combs through all other information databases, from the EHR to other tests and exams, even outside sources reflecting social determinants of health, to form a more complete picture of the patient and recommend diagnoses and other treatments.

“In the future it’s going to be impossible to be a radiologist and not use AI,” she says.

Eric Wicklund is the associate content manager and senior editor for Innovation at HealthLeaders.


Health systems and hospitals are using AI as a tool to read X-rays and other images, picking up on small things that a radiologist or other clinician might overlook.

Clinicians are using the AI technology as a tool to improve their decision-making, not as a replacement for their services, and it’s important that a clinician always review the AI tool’s output before making a clinical decision.

 Healthcare organizations need to evaluate AI as a complement to clinical care, rather than pitting the technology against the clinician.

Get the latest on healthcare leadership in your inbox.