Artificial intelligence (AI) has the potential to revolutionize dentistry, though many clinicians may be unfamiliar with the technology. The Dental AI Council (DAIC) recently spoke with Teresa Duncan, a consultant, public speaker, and author who focuses on revenue and management issues facing dental healthcare providers, about AI and how it may impact dental practices.
DAIC: Can you please provide a little background on how you came to recognize the potential of AI in dentistry?
Teresa: For the last decade or so, we’ve been seeing a lot of changes in the medical side of AI. For example, we’ve seen radiographs evaluated for cancer analysis. It’s really exciting that this is now entering dentistry. I wasn’t sure how it was going to translate, but it looks like the quality control side is going to be a real boon to the industry in terms of being able to double-check our diagnoses.
DAIC: What have you seen in the dental market in terms of AI?
Teresa: I think two main areas are really interesting.
The first area is in achieving consistency in treatment planning and in diagnosis. By having a database of images, clinicians can learn whether their diagnosis is as accurate as it could be. The wonderful benefit is that a provider can learn from lots and lots of data. When you’re in dental school, you have only a couple of patients to go through, whereas with AI you have the possibility of looking at hundreds of different images.
And then the second area, which really interests me because of my insurance background, is the ability to use it on the carrier side for detection of fraud, waste, and abuse. I know from the provider side that we need to really keep an eye on outlier providers, because those are the ones that give dentistry a bad name. With AI, it’s easy to catch those outlier providers.
DAIC: In terms of that teaching, is there a role for AI?
Teresa: I think it will be helpful. I think AI will help dentists calibrate their diagnosis skills. A dentist coming out of dental school sees a decent number of radiographs while they’re in dental school. First couple years, they see a pretty decent number of images too.
There’s a learning curve, though. Not many dentists who are two years out know exactly what they’re looking at a hundred percent of the time. By the time they’re in practice for 10 years, they’ve seen so many radiographs that it really is not that hard for them to pinpoint the issue.
Radiographs are pretty subjective. It’s really your interpretation based on the teaching and experience that you’ve had. I think just having a large database of images to work from is going to be very helpful for students to reach that level of excellence faster.
And also, the carriers could do more with, “Here’s what typically we see submitted and what would be acceptable documentation in order to get paid.” That doesn’t mean acceptable for having to have a treatment done, because the dentist is the one who ultimately decides that. The carrier’s job is to take that image and decide if it’s going to be paid, not whether or not the work should have been done.
DAIC: One of the potential benefits people talk about AI bringing is real-time claim approvals. Do you see a potential for real-time adjudication approval and treatment, where you get the diagnosis, get approval on the treatment plan, and then go immediately treat?
Teresa: As far as far as quick turnarounds on pretreatment estimates, no, I don’t believe that you’re going to get that any time soon. It’s very different than medical. Medical will typically pre-authorize everything. Offices will usually schedule the treatment for after the pre-authorization comes in. No surgery happens without that. The hospital itself actually insists on it, because they have so much to lose in bad debt.
In dentistry, the opposite is true. We will do the work and then send in the claim. Then we end up having to justify why we did the work. So, we do all the justification on the back end, which is a terrible way of making sure you get paid for your services.
There are lots of dentists who are fee for service. They don’t participate with insurance and, while they still file for their patients, they’re not as concerned with getting approvals and estimates. On the other hand, many of the participating [PPO] offices will pre-estimate everything, and that’s really sad because it puts off the work and the patient loses that sense of urgency.
The reason offices do it this way is that they’re unsure of how much is out of pocket. They want to make sure that they’re giving the right information to the patient and so they utilize the pre-treatment estimate option. And that’s a function of dental insurance. Dental insurance is typically not the clearest enterprise. Our process is not as smooth as it could be.
DAIC: Do you see AI bringing more clarity?
Teresa: Yes. I think what we’re missing is real uniformity and standards. I don’t want to take away the art of the clinician, because they’re scientists. They’re taught their skill. Still, you put four dentists together with the same set of radiographs, and it’s been proven in studies they’ll come up with different treatment plans. It will depend on whether their diagnosis style is conservative or aggressive.
The dental consultants who are looking at the claims are usually former dentists. This means that you may have a dental consultant who is very aggressive and would crown a tooth. Or you may have one who’s extremely conservative and wouldn’t replace that restoration with a crown. They would put another filling in.
And so it just would be nice if we had standards on what this radiograph means, and should we use this radiograph to make a sweeping decision on all crown buildups? I don’t know if we’re there yet, but it would be nice if, before dentists send in anything to insurers or do any other crown buildups, they see what has been sent in [to the insurer] previously. And what an ideal radiograph looks like, which is what AI could do.
I think that would get people to think a little bit differently on what [evidence] needs to be there to get paid. Of course, I’m talking now about the documentation, not necessarily the treatment needs of the patient.
DAIC: What about AI’s potential in other areas?
Teresa: When I’m teaching and talking about waste and abuse, there’s always a disbelief that it’s really bad in the industry. In my course I present a lot of examples where dentists actually tried to justify the treatment using this [bogus or invalid] documentation. And people can see that there are a lot of bad apples out there, but I still don’t think they realize just how bad it is, especially in the Medicaid arena, because of the potential for large scale fraud with pediatric billing.
I really hope AI is helpful in catching that because, as humans, we only have so many eyes. And if the computer can recognize that, “Hey, this is not kosher, these radiographs aren’t showing what we think they should show,” I think we’ll catch a lot more of the bad apples out there.
For example, we’ve had providers using the same radiographs over and over to get their claims paid. An AI system would spot that pretty quickly. Dentists who want to abuse the system and overbill are going to find ways to do it. And it’s up to not only the dental industry, but the dental insurance industry to catch that before it becomes too costly.
The other part of that is, when I talk about AI or the data analysis and data mining that insurance carriers do, there are a lot of dentists and team members that don’t think of the fact that carriers look at it as a big business.
In dentistry, we’re small businesses, for the most part. So if we looked at our business the same way the carriers did, we would also be employing ways to cut down on waste and abuse in our facilities. With doctors who are mentoring, AI is a great way to catch it and to use it as examples. I am really excited about the fact that we could lower the rate of billing abuse in the industry. I would love to see that happen.
DAIC: Do you have any sense of how much could be saved?
Teresa: I think that would be impossible to put a number on. What I can say is in my close to 30 years of working with dental insurance, we’ve gone from seeing just messages of, “Let us know if you see any suspicious billing” to, now, whole departments and hotlines built around combatting abuse.
The Affordable Care Act provided money for something called a Recovery Audit Contractor (RAC). There was a huge rise in those after the ACA got passed, because they anticipated more fraud and abuse. On the healthcare side of fraud, it’s unbelievable how much is happening. Dentistry is still a small part, but I think the carriers have realized that it’s easier to spot than on the medical side. And I think it’s a really good return on investment for them to have those processes in place.
DAIC: What about sentiment around AI in the dental community? Is it an impediment to adoption?
Teresa: When I discuss in classes that AI is coming to dentistry, there’s always suspicion among the providers that it’s a Big Brother type operation––that the only reason they want to use AI is to catch us and to deny more claims.
And it’s not bad apples saying that. I mean, they’re taking my class. Bad apples don’t take my classes. The ones who are taking my class are the ones who are trying to do it better. They already feel frustrated by the pressure from carriers.
And when you add AI into the mix, they’re thinking, “This is another way for them to keep me under their thumb.” So, there’s a little bit of frustration on the provider side. They really just don’t know what to expect from it, because they haven’t seen it in action.
That’ll change as it becomes more widespread with the more veteran dentists. But you talk to them about AI in mentoring and that, I think, is very exciting for them, because they can use it to teach, which is what they want to do. In my experience, though, it’s not been a hundred percent open arms yet, because there’s still a lot of suspicion about how it’s going to be.
DAIC: How do we address that suspicion?
Teresa: We should focus on how we can better treat patients. If more dentists who are using it can come out and discuss that, and how it’s a benefit for the patient, that’s ultimately what we’re in the business for. When you focus on how the carriers are going to use it, that’s more of a numbers thing. It doesn’t really appeal to the healers that we’re talking to.
I think if you hear from more dentists who are using it on the mentoring side, and also from the teaching centers, talking about how they can help train the new breed of dentists coming into the field, I think that’s how you get the awareness out. You’ll be letting people know about the waste and abuse, which is almost a dirty little secret in dentistry.
People know about it, but they never know anybody who does that. It’s always somebody we read about. But, in reality, most dentists have had some sort of run-in with somebody, either embezzling, wasting, or committing fraud in their office.
I think there could be more use of cases. AI vendors will need to be a lot more visual when teaching AI usage. Hearing about it is one method, reading about it is still another, but actually seeing the images and how they’re used is absolutely the most impactful method to show its use.
DAIC: Do you have any questions about AI that you’d like the DAIC address?
Teresa: I’m curious to see whether the AI is going to integrate into practice management systems, or if it’s going to be a standalone system. Because my experience with standalone systems is that they’re used initially and then they tend to fall by the wayside. So I’m really curious to see how it gets integrated into the practice management systems.
And then I’m curious to hear how the Council is going to reach every member of the dental industry or actual dental practice. Because I think for any new technology, the team needs to be brought in so that we all understand why we’re going in the same direction.
DAIC: Great. One last question. Can you suggest another great mind that you’d like to hear from on the topic of AI in dentistry?
Teresa: I’d love to hear what Dr. Paul Feuerstein has to say. His feedback would be priceless in this situation. He’s a tech guy with lots of experience and very well respected. He’s seen everything come to pass, and he’s just got such a unique way of looking at trends.
Teresa Duncan is a consultant, public speaker, and author who focuses on revenue and management facing dental healthcare providers. For more information about Teresa and her work, visit odysseymgmt.com/about.
The Dental AI Council (DAIC) is a nonprofit devoted to helping define the future of artificial intelligence (AI) in dentistry through research, education, and thought leadership. To join to the DAIC’s effort, visit dentalaicouncil.org/membership.