I was taken aback by the question I got from an executive with an artificial intelligence (AI) company the other day. "What do you have against AI?" he asked. "Nothing. Why? I love the technology," I responded.
"It just seems like everything you write about questions the use of the technology," he said.
I tried to remain objective and understand his concern. Yes, I do look at things from slightly different perspectives than others may. That's my role as a consultant -- to look at everything from every angle, to understand the pros and cons of the situations (in this case the technology), and to never sugar coat it.
I never was a Pollyanna and never will be. Unfortunately, when doing my research, I often find contradictory information as well -- one study says this, while another says something different about the same thing -- and not just differing opinions either. That doesn't help matters at all.
AI seems to get the most questions about market growth but has the fewest answers. It's not for lack of trying, though. Every day I read a newly published study about how AI can make the radiologist's job better, faster, and easier.
Sadly, there is often little substantiation behind these claims. Substantiation is ultimately the key to gaining support for AI, beyond just the radiologist's desire to use the technology.
It's not just AI that has interest, either. I am often asked where the PACS, enterprise imaging systems (EIS), and teleradiology markets will be in the next three to five years. Projecting where any market will be in the next three to five years, months, weeks, days, hours, or even minutes is sometimes asking a lot. No one could possibly have predicted the impact of COVID-19 on imaging. There were and are just way too many unknowns.
Using AI should be a no-brainer, but in the end, there are usually just two questions that need to be answered:
- Will it make me money?
- Will it save me money?
These questions can be answered in a plethora of ways, as long as they are answered. These aren't the only questions that need to be addressed by any means, but they often constitute the lion's share of the decision-making process.
While hospitals make up the vast majority of the entities that are making decisions to acquire technology like AI, it's not just healthcare providers that are using or looking at using the technology.
Healthcare payers are starting to use AI technology to evaluate prospective beneficiaries to determine their acceptability prior to underwriting a policy. AI would be used with an imaging procedure (typically a CT scan) and combined with other clinical data to create a numerical score that "rates" the potential beneficiary. Too high or too low a score and a policy would not be issued.
Even though the upfront cost of such a use case would be fairly significant when factoring in the cost of the technical component and AI usage, most seem to feel the proactive value of AI in this situation is well worth it. At the very least, the use of AI here would be a fraction of what a single hospitalization would cost. Like everything else though, the value proposition needs to be shown. Sadly, most AI companies doing this simply don’t know how to present the value proposition of their products to the stakeholders and others.
When you look at how AI is promoted, often you hear statements about time savings, improved diagnostic capability, improved patient outcomes, etc. All of these are very real, but showing how these help the facility paying for AI is another story.
Time savings is very real but requires trust in the software. I can be skeptical at times -- most times the truth be told -- but there are people who put me to shame in terms of questioning things. The industry brought skepticism to the table by saying years ago that AI was better than radiologists. Back then, study after study was quoted, yet it simply wasn't true.
Even now, where AI has improved to a much higher level, many people still don't believe AI's claims, believing the statement "fool me once, shame on you, fool me twice shame on me." Even something as basic as using AI to calculate breast density has been met with skepticism.
Yes, using AI can save time, can automatically populate the radiology report, and even be more accurate as well. I have seen this in clinical practice and believe it. Yet when you chat with others about the benefit AI brings to the table they make it out to be no big deal.
It's like listening to Ethel Merman singing "Anything You Can Do I Can Do Better" in "Annie Get Your Gun," with radiologists who have never even seen AI let alone used it leading the chorus. "No you can't. Yes I can."
So when will AI be "real? Many imaging facilities are using it now, but how many have actually paid "hard dollars" for it remains to be seen. Many academic facilities obtained AI on an evaluation basis or through grant monies without any out-of-pocket or near-zero cost.
In 2020 the new technology add-on payment (NTAP) program was launched by the U.S. Centers for Medicare and Medicaid Services (CMS) to assist facilities in justifying AI financially as well. Sadly, less than a handful of the 200+ companies offering AI products cleared by the U.S. Food and Drug Administration have been able to take advantage of the NTAP program. NTAP for AI was also supposed to end this year, but it was extended for yet another year in the hopes to get more people to embrace the solution. Still, that is a short-term solution at best.
It's not until there are current procedural terminology (CPT) codes associated with AI that the technology will take off and finally be considered "real." Some vendors have already gotten CPT III codes issued; however, payment against a CPT III code still remains voluntary.
Once a CPT code goes from a category III (with no assigned relative value unit, or RVU; the payer determines payment or not) to a category I (with a fixed associated payment amount) the value proposition of AI can be shown in hard dollars.
At this stage, skeptics will then have one less piece of ammunition to fight against AI, and the technology can move forward another small step at a time.
Michael J. Cannavo is known industry-wide as the PACSman. After several decades as an independent PACS consultant, he worked as both a strategic accounts manager and solutions architect with two major PACS vendors. He has now made it back safely from the dark side and is sharing his observations.
His healthcare consulting services for end users include PACS optimization services, system upgrade and proposal reviews, contract reviews, and other areas. The PACSman is also working with imaging and IT vendors developing market-focused messaging as well as sales training programs. He can be reached at [email protected] or by phone at 407-359-0191.
The comments and observations expressed are those of the author and do not necessarily reflect the opinions of AuntMinnie.com.