AI can help hospitals read scans faster, shorten MRI appointments and create additional imaging capacity. But radiology leaders say many patients still struggle to get imaging for reasons no algorithm can solve: transportation, childcare, insurance coverage and the ability to take time off work.
As health systems invest heavily in AI-powered imaging tools, a question is emerging alongside the excitement: What does it actually mean to say AI is expanding access?
For some organizations, access means reducing wait times and increasing scanner capacity. For others, the biggest barriers lie outside the radiology department entirely.
At Chicago-based Northwestern Medicine, AI has become a key tool for increasing imaging capacity without adding staff or extending operating hours.
One of the most significant gains has come in MRI. Historically, a total spine MRI could take 60 to 90 minutes. With AI-enabled software, Northwestern has reduced some studies to 30 to 45 minutes. The result has been a substantial increase in throughput.
“In many of our hospital areas we’ve seen upwards of almost 30% increase of volume with same staffing resources, same hours of operation,” Aaron Stonecipher, system director of nuclear medicine & molecular imaging, told Becker’s.
The health system measures success through metrics such as scanner utilization, daily imaging volumes and study duration. Mr. Stonecipher said the added capacity is an access improvement because it creates more appointment availability while reducing the amount of time patients spend inside scanners.
Northwestern is also using AI beyond image acquisition. The health system has deployed tools to help accelerate prior authorization workflows and support follow-up care through a results management program that identifies incidental findings in radiology reports and helps ensure patients receive appropriate notifications and referrals.
But Mr. Stonecipher acknowledged that throughput is only one piece of the access equation.
When asked why patients delay or forgo imaging, he cited appointment availability, patient understanding of why a study is needed and cost as some of the most common factors. The organization has not seen evidence that patients are avoiding imaging because AI is involved in the process.
For rural health systems, the distinction between operational efficiency and real-world access can be even more pronounced.
At Allen Parish Community Healthcare in Louisiana, AI has primarily been deployed through stroke imaging software that helps identify potential large vessel occlusions and prioritize critical CT scans. The technology can accelerate stroke evaluations, support telestroke programs and improve after-hours workflows when radiologist coverage is more limited.
Yet Lucinda Daigle, director of radiology at Allen Parish Community Healthcare, said the most significant barriers to imaging remain financial and logistical. Patients most commonly delay imaging because of “financial concerns, transportation issues, scheduling challenges and limited local specialty access rather than distrust of AI technology,” she said.
Even when AI improves turnaround times or speeds the identification of critical findings, many patients still face obstacles such as high deductibles, lack of insurance, difficulty taking time off work, transportation challenges and limited childcare options, according to Ms. Daigle.
“AI can improve operational efficiency and support faster clinical decision-making, but it does not fully address the socioeconomic barriers that affect access to healthcare in rural communities,” she said.
Valley Children’s Healthcare in Madera, Calif. has reached a similar conclusion.
The pediatric health system uses AI-powered critical finding triage tools across its 12-county service area and tracks metrics such as image acquisition time reduction and scanner utilization. But Mark Toatley, director of imaging services for radiology administration, said radiologist capacity is rarely the primary reason families delay care.
“The primary reasons families delay or skip imaging appointments stem from systemic social determinants of health, including geographic and transportation barriers,” he said.
Finding childcare, transportation challenges and the realities of rural life often create greater obstacles than scanner availability, he added.
Mr. Toatley said AI can also sometimes create road blocks elsewhere in the care continuum.
“When clinical AI successfully speeds up image acquisition and radiologist reads, it can inadvertently create a downstream bottleneck on the patient-facing side when our administrative workflows are not automated to match,” he said.
As a result, Valley Children’s measures access more broadly than imaging speed alone.
“We define ‘improved access’ by the total elapsed time from the initial physician order to the finalized treatment plan, rather than just the speed of the machine,” Mr. Toatley said.
The differing perspectives highlight a challenge facing health systems as AI adoption accelerates: Operational access and care access are not always the same thing.
AI is helping organizations create more imaging capacity, shorten scan times, improve prioritization of critical cases and streamline administrative tasks. But technology alone cannot overcome many of the barriers that determine whether patients ultimately receive care. For health systems evaluating the return on radiology AI investments, the question may be less about whether AI improves access and more about which kind of access it improves.
At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.
