Banner

News

Article

Medical Economics Journal

Medical Economics April 2025
Volume102
Issue 3
Pages: 31

How physicians actually use AI: Ronald Rodriguez, MD, PhD

Author(s):

Key Takeaways

  • AI implementation in healthcare raises concerns about incremental costs and who will bear these expenses, potentially impacting patients and third-party payers.
  • There is a need for better economic analysis and modeling to understand AI's impact on healthcare costs and disparities, as efficiency does not necessarily equate to reduced costs.
SHOW MORE

How Ronald Rodriguez, MD, PhD, professor of medical education and director of UT-San Antonio's dual MD/MS in AI program, uses artificial intelligence.

Ronald Rodriguez, MD, PhD

Ronald Rodriguez, MD, PhD

Dr. Ronald Rodriguez is at the forefront of educating doctors on AI tools and what they mean for medicine. However, when it comes to the future of AI in health care, one problem keeps coming to mind: cost.

“I’m actually quite worried about how this might unfold because the current economic models of AI are all directed at incremental increase costs for implementation,” Rodriguez says. “Who’s going to pay for it?”

Patients might be the easy answer, but Rodriguez points out that reimbursement comes from third-party payers that don’t have a way to identify which AI is being used in care because there are not specific AI codes.

“What ends up happening is, to make up for the expenses, we have to see more patients,” Rodriguez says. “To see more patients, third-party payers have to pay more money. If they have to pay more money, they pass on those expenses to their insured population.

“I think that for many, we’re going to see a dramatic improvement in the quality of our care, but at an increased cost.”

Rodriguez says there needs to be better economic analysis and modeling of the potential impact of AI on the costs and the disparities that might be created. The consensus is that AI will decrease the expense of administering health care, but he says that’s a naive view.

“They are making the assumption that improving efficiency means less cost, but they haven’t really considered the incremental cost of utilizing those technologies,” Rodriguez says. “They don’t come free, and the more you use them, the more they cost. All of the tech entrepreneurs are relying on that — it’s their business model. And so, they’re expending large amounts of resources with the assumption that it’s going to have this huge revenue payout, but we already expend a huge proportion of our GDP [gross domestic product] on health care services. How much higher can we go?”

Rodriguez says there are no obvious solutions, but physicians need to take a much more active role in technology development to mitigate disparities and costs.

Physicians also need a better understanding of the technology they use to protect themselves against crimes, like violating the Health Insurance Portability and Accountability Act (HIPAA). For example, the FDA considers AI a device, which requires approval for use in a medical setting. Rodriguez says using off-the-shelf generative AI like ChatGPT or Gemini to assist with note writing and summarization instead of dedicated medical tools with FDA approval could put doctors at risk. Doctors may use readily available public AI because it is cheaper than approved models.

“But in the act of providing that information in a prompt, you may be inadvertently providing patient health information and violating their privacy if their personal health information is contained in what you uploaded,” Rodriguez says. “As a consequence, you may be in the middle of violating the HIPAA Tier 1 regulation, and those violations cost per occurrence. And so, if you did that 10 or 20 times a day, they all add up.”

Related Videos