David: [00:00:00] Hey everyone. Welcome back. This is our second crossover podcast with Aleks and Giovanni. Hello.
Aleksandra: Hello.
David: Aleks, how are you?
Aleksandra: Good. Great to be back with you. I’m super excited. I’m always super excited to talk to you guys, so I’m super happy to be here again.
David: Yeah.
Giovanni: Same here, same to you very excited. Happy to see you again, happy to share with you.
David: What we’re doing today is we want to get together a couple times a year and just talk about some of the hot topics that us well-read individuals follow. We also post our podcasts, so we talk to some of the experts and maybe we like to think that more.
Aleksandra: Better read individuals, cool.
David: Yeah, exactly. So, maybe we like to think that we’re at the forefront of what’s going on and here’s a chance for us to talk as a group about what we’re seeing and hope you guys enjoy it. So we’re gonna talk about the year of AI, we’re gonna talk about CPT adoption and how that’s going the new digital pathology CPT codes.
And we’re gonna be talking about a [00:01:00] subject that is very near and dear to my heart, software hardware integration. I might end up having a cry session at the end, depending on how that goes so.
Aleksandra: Let’s see if we can like, have some cool exchange, David, about this.
David: We’ll see. I don’t know, that’s something that, that’s obviously one of the main things that, that I develop and yeah, it causes a lot of headaches, so we’ll see. Who wants to go first? Who wants to do the topic first?
Aleksandra: Let’s start with the year of AI, because AI I think is gonna be definitely important for my topic as well.
David: All right cool. So here’s the premise. Earlier this year, I think it was the Atlantic Magazine came out with an issue stating that this is the year of AI, and I think one of the main.
Aleksandra: Is that the first time they stated this, or did they do it last year?
David: It’s probably not the first time they’ve stated it, but I think it’s extremely relevant and appropriate to talk about it because what is like, what is the hottest topic on the planet? On the internet right now? It’s.
Aleksandra: AI chat GPT.
David: Exactly, it’s chat GPT, so exactly. So I guess, I eventually want to get to is what is the [00:02:00] chat GPT going to be for digital pathology?
And maybe we can have some predictions. But first, this chat GPT is the real thing. Everybody that I talk to, they use it in some way, shape, or form.
Aleksandra: I use it.
David: I’m personally scared to use it because I’m like afraid. I’ll give you my like concern as a member of society is that I think chatbot, GPT can write pretty well for people and are we gonna lose our writing skills?
Aleksandra: I dunno.
David: I’m scared of that I place incredible value on writing skills and now, like you have, the younger generation that’s going to class and they have to, they’re, they wait till everyone else. They wait till the last minute to, go to write their essay the night before it’s due and, oh look, you got Chap GPT right there that can just write it for you. That’s a, it’s a scary concept to me.
Giovanni: I can tell you I lost completely, and it’s not that I had a big sense of orientation. It was always very broken to start with, but now with the all this GPS system location maps, I don’t even pay attention anymore of where I’m going. Everywhere I’m going I [00:03:00] have to put the map to take me back home, because I’m completely oriented and I just relinquish any, again, I wasn’t good to start.
But the few orientation capabilities that I have, they got rusty and probably disappear. I don’t regret it, I’m okay with that.
Aleksandra: So let me tell you about my use of chart GPT and to comment on if we’re gonna be losing our writing skills.
Probably yes and then it’s gonna be taken to a new baseline where, because I use it a lot for content creation for social media. And then I prompt this chat GPT. So you give this spot a question and it answers you the question. For example, let’s say we have this podcast and I’m gonna ask chat GPT, Hey, give me a one pager summary based on the transcript of this podcast and use a few bullet points.
And it gives me this one page summary. And then I go through the summary because let’s say 30% of it is not so coherent, and then a little bit of that doesn’t really sound like in the [00:04:00] voice that I’m usually talking to my people. So I go through it, I adjust it, and then it’s mine and it’s faster, so that’s how I use it.
But I’m with Giovanni, I’m like, terrible with maps without gps, I will not get anywhere.
David: Yeah, Aleks, that’s really interesting that you use it in that way because it’s somewhat of a correlation to how we’re trying to use it in pathology. We are feeding in imaging data and it’s spitting back to us some kind of results about a diagnosis, but it’s not telling us, this is the diagnosis.
It’s suggesting that this may be the diagnosis, or this may be where you wanna look. So we’re still in that phase where we need the human to put their own personal touch on it and I use this term all the time, but essentially police the algorithm. So in your writing with chat GPT you are still policing what’s being written.
You’re still taking the, essentially the first draft, and you are making it in your own voice. And I think it’s the same with where we are in artificial intelligence, like real clinical applications and pathology [00:05:00] is, yeah, we’re getting suggestions, but we’re still, like the human pathologists are still making the final calls.
Aleksandra: Yeah. I’m very strong on QCing, the algorithm on policing, the algorithm on, when you have the algorithm metrics, performance metrics. You have the numbers and even the numbers like they’re already in the nineties, they’re al already in, when you compare to pathologists, they’re either non-inferior or even superior or a combination like for the page algorithm, that is a computerated diagnosis.
It shows where the pathologist just supposed to look for prostate cancer, right? It has high numbers, but you still have to visually check if it’s not producing those numbers from some funky places that don’t have anything to do with what you’re looking for. So in this way, I think, I just think AI isn’t our tool in terms of society.
David, my husband, is with you. He thinks it’s very disruptive for society, and a lot of people are gonna lose their jobs because data entry or this type of transcript is their job, which yes, probably is gonna be the case. But it was [00:06:00] the case when cars were invented and when trains were invented and when other technological revolutions happened.
It has a double-edged sword, it’s a double-edged sword.
David: Giovanni, are you using any AI tools in your day-to-day?
Giovanni: As a test only, we are not really, we have been able to deploy any algorithms but we are testing them using parallel, like I do my diagnosis and I use the AI, I compare AI to what I’m doing and I don’t report any of the AI exclusively.
I just my separate report of what the correlation is sure and hopefully all these preliminary observations will be pulling up in a paper that we plan to publish and different in different areas. We have general algorithms, we have the more sophisticated AI that makes actually diagnosis and classified too much and we are just, I’m delighted testing it.
David: What’s it gonna take for you to actually deploy one of these algorithms for real clinical use?
Giovanni: I think when we finish with [00:07:00] this process, with this kind of been like an ongoing validation because we haven’t set the number when we’re gonna finished. But when we feel comfortable, that’s one side and the other is when we have a full integration, we wanna, right now we have all these algorithms on a separate parallel flow. So meaning you have to export the image to whatever other software you are using and then do the analysis and then you transfer to whatever decision or whatever conclusion you have from the AI into your report.
And in order for us to be able to do those, to do this in a big scale with all the pathologists, remember that we are, this includes people who are not technology savvy, people who are not that happy about digital or AI to start with. So we need to make this a very easy process with just a few clicks and having to exit one system to get into another one and come back.
It’s not really sustainable for a big department, I think it’s only doable if you [00:08:00] have a highly motivated pathologist to do all these extra steps.
Aleksandra: These are some words of wisdom. Giovanni.
David: Yeah.
Aleksandra: I think you should always start with those highly motivated people, not to discourage those who are not that motivated, and let those motivated figure out the all the kinks and then all the others have smooth right? And can basically benefit from this without having to troubleshoot because every time somebody is doing something first, there’s troubleshooting is part of the deal. If you’re not willing to troubleshoot, then maybe that’s not the best thing for you to do.
David: So the premise for this topic was in popular science or popular culture, this has been coined the year of AI.
So the question is this the year of AI for pathology? And even further, what is the chat GPT of pathology going to be? I think we’ve already answered the question, and Giovanni, I’m probably gonna have to agree with you on this one, that the answer is no. This is probably not the year of AI for pathology.
Aleksandra: I heard for the same. Not quite yet.
David: Just [00:09:00] because we have. I think we need to continue to take baby steps to get there because the tools we have available are great, but they are complex and they require a lot of buy-in from individuals that I don’t think they’ve, they’re not quite there yet in terms of the motivation.
Giovanni: No, definitely. No. And I think when we have the technology that’s gonna be able to be self-sufficient, that you don’t need to babysit it, there is when everybody will be able to buy in knowing that of course the final decision will be the physician’s decision who’s making the diagnosis, right? But at least he will get the AI opinion or input in a direct way, in a direct straight way that they don’t have to be hesitant about if there is an error that they made when they were trying to process the information.
David: So then the kind of, the second part of this is what would the chat GPT for pathology look like? And the ideal, pie in the sky idea, is you [00:10:00] drag an image into this platform and it tells you literally everything, every single thing about it.
If you look at what’s out there in the market, there are a lot of platforms that can tell us a lot already. We have platforms that are, some are FDA approved to tell us a diagnosis. We have these really cool research platforms that can now tell us incredible amounts of detail from an image about a tumor microenvironment.
They can tell us incredible amounts of information about a single cell, but I think we’re falling short on, and that needs to be part of this new platform is algorithms need to start, or developers need to start looking at how can we use AI past what human pathologists are looking at? Can we look at, can we use, can we develop algorithms to look at images, to find things that we, like pathologists haven’t been trained to see?
So the first couple things that go that come to mind are prognostic biomarkers. Can we find things that help predict treatment outcomes? Yeah. Aleks, go ahead and chime in. You got something? You got something I’m sorry.
Aleksandra: Am I gonna be interrupt, [00:11:00] otherwise I’m gonna keep like talking.
David: No, just go. Just go. It’s fine.
Aleksandra: No, I’m gonna raise my hand whenever I have a comment. You can decide who speaks but.
David: Go for it.
Aleksandra: Spoiler alert for this very topic, you know who agreed to join my podcast, hopefully in, we’re gonna be recording in April. Anant Madabhushi, who is very much working in this space, developing or predicting biomarkers from H&E.
My very question is, okay, we have this in the research space and it’s, every week you have a new paper, which marker was predicted. What will it take to get it into the clinic? What kind of validation? Because visual validation by a pathologist, what Giovanni is doing, they’re doing it. You guys are doing it. Once this is done, you have confidence. The final say is the pathologist, and we figure out how not to babysit AI.
What about. Like this, what you’re mentioning, David, the predictions, the prognosis from image without knowing like where it comes from, but [00:12:00] basically trusting that it is correct.
David: Yeah.
Aleksandra: So I hope to ask him, like we mentioned at the beginning, we often invite very smart and very well read people. So I consider him to be somebody like that.
Let’s see what he says. Another comment, and then I close my mouth for a little bit. Is so chat GPT, if we like when I translate it into images, that would be what we just said. But we can also use the language models for pathology. And I have seen companies working in that space that are supporting report.
So Giovanni, do you guys dictate things for reports?
Giovanni: Yeah, we have a Dictaphone, which is, it’s actually very good and, but we, I personally prefer to do the input myself with codes and build a report myself, which is find that is easier because the way I can check it at the same time and release it and don’t have to go back.
But we do have a Dictaphone machine software that is very good.
Aleksandra: So imagine if you could just like, instead of typing, you could say, Hey, in my pathology chat [00:13:00] GPT, could you put this code here and that code here and that I don’t know if you’ll probably have like specific building blocks of the reports.
That’s how I remember it from veterinary diagnostic it’s you, now your diagnosis you say, oh 0.01 paste there. If you could just do it with voice without going into the computer, I would imagine this to be the like direct translation of chat GPT into pathology. I could use this for my reports as well.
Only I couldn’t because a lot of this is very proprietary information, so it would have to be something that is in-house built for human medicine, HIPAA compliant for labs, G L P compliant and not some like open source thing where everybody can paste, like all the transcripts of whatever they’re doing.
It would have to be very secure, so that would be my concern about that.
David: The reporting that is a really intriguing idea and I hope people are listening, innovators are listening to this and getting to work on that right now. Because when I talk to pathologists, like one of their top pain points and inefficiencies is indeed the reporting.
So anything we can do [00:14:00] to help, right?
Giovanni: Something that is very overdue. Is the fact that in pathology, for instance, we have all these different diagnostic buckets, right? That we are surgical pathology, we deal with the images, but all the other laboratory divisions are also part of pathology, hematology, chemistry, immunology, cytogenetics, molecular.
If you can combine everything into one comprehensive object sent report, that the clinician can look at it and see everything together without having to be checking for glass over here, labs over there. Everything comes predigested to you, morphology, immunology, molecular I think we need that. Go ahead.
David: No, you don’t have to. If no one’s.
Aleksandra: If no one’s talking I can just talk. I’m impossible I know. I’m notorious in interrupting people when I wanna say something. It’s not good for podcast hosts.
David: I just do it. Sometimes people talk too long and you have to interrupt them, there’s no way around it.
Aleksandra: Okay, so Giovanni, I have to tell you, so in direct development you do have something like that, but it’s not automated in any way [00:15:00] because where you work for developing a drug and data from different places are coming in, pathology is one piece of data. Then there is in live data, which would be clinical pathology or some labings, like a lot of data pieces, a lot of different people responsible for that.
And as per FDA, the one person that’s responsible for compiling this final report is Study Director. So they compile the report and then when it goes in it, when it leaves this phase, when this study was done, it goes as a final report to the next step, which either is the sponsor who wanted the study to be conducted, or regulatory agencies.
And I see if I understand correctly, you would love something like that to be at a patient level and that tragedy would help.
Giovanni: It will help for sure the doctors, treating doctors that are trying to put everything together, especially in complicated cases like cancers where you have so many sources of information and you have to start exploring them one by one, separately, and sometimes you may miss some data that is [00:16:00] important for this particular case because when you go to labs, for instance, if you’re reviewing the data for this particular cancer.
Maybe the patient had another aubin before where multitude of immunologic tests have been ordered, and then you have to sort through all of those and find which ones are important for this particular case, rather than having all the information pertaining to this disease or this encounter in one.
Easy to view all encompassing for, that’s my wish for the chat GPT. We’re asking wishes like I have three wishes. So I’ll think about the other two.
Aleksandra: Okay, sure.
David: We got two more topics to cover, so I’m gonna close the, I’m gonna, maybe those that’s where user two wishes. But I’m gonna close the discussion on this one and just summarize and say that I think we all agree that it’s probably not the year of AI for pathology, but let’s remain optimistic.
Let’s hope that for pathologists use the platforms that are available and get more comfortable with the assistance and the diagnosis first. That’s going on [00:17:00] this year, perhaps this is my prediction, is we’re gonna see a lot of really interesting publications about tapping into those prognostic biomarkers, and I think that’ll be a very interesting development that I look forward to following this year.
We have the prognosis plus the diagnosis, plus some other things, maybe some chat GPT elements. Gonna have some really cool platforms in the future, but my hope is that at the very least this year, that more pathologists get on board and get more comfortable with the capabilities of AI.
Giovanni: Definitely some steps. Some steps will happen. Yep. Some forward steps for sure.
David: Okay. All right, Aleks your topic.
Aleksandra: Okay, so my topic, which is integration of hardware and software, which is what you live and breathe in your business and basically is what would need to happen for Giovanni to stop babysitting all the systems.
And I’m seeing, let’s say, there is enough confidence now in that, okay, you can diagnose, you can evaluate on a digital image. That’s okay. Everybody more or less is fine with the quality of [00:18:00] image. It’s fine, it’s good, but everything that is before, unlike in radiology, the analog didn’t disappear, you add on top of the analog.
So you scan a glass slide, you have to do, you have to put those slides in the scanner. And there is so many manual steps and what I’m seeing happen now in several companies that I interviewed on my podcast is that it’s been taken to the next step where image analysis is happening at the time of scanning, doing automated qc, and there is automation involved and out.
By automation, I mean something beyond just barcoding the slide actually like using machines, robots to pick the slide, put it in the scanner or they put it in wherever, like basically automation in the lab, automation of these things that are now an additional overhead, especially the qc there are currently hours of man hours, manpower spent on QCing scan sites.
And when everybody looks, [00:19:00] when the payers or those who are responsible for that investment, look at it from the outsider was it not supposed to be cheaper or faster? And you just added another full-time employee to do this job that was supposed to not be necessary anymore, so this is where this automation comes in.
And David, I don’t know how difficult it is to integrate, I know that I love my smartphone it’s a very nice integration of hardware and software but if we could take the smartphone concept into the pathology lab, at least for some tasks, that would automatically increase the, not only the perceived the actual value and the actual savings and digital pathology.
Let me know what you think about it and let me know how you see, how you interact with it or how does it work in real world? In what you.
David: I have a lot of thoughts on that. First of all, this idea is something that will greatly help the overall adoption of digital pathology tools, mainly because, the more [00:20:00] manual let me back up for a second.
We already know that buying digital pathology hardware is expensive. You know that buying digital pathology software is expensive and not on top of that, you have to hire more people, which adds to the expense because like you said, there are still all these manual steps. So three major expenses for to solve one problem right?
So that’s the challenge. The question is how can you automate some of those steps? I know a company that tried to essentially automate, like with a robot, like the entire H&E process. It’s a really interesting concept, but spending a lot of time in pathology, gross labs, that work is very chaotic.
And it’s very labor intensive and it’s just a tough thing for me to imagine it in the near future, being completely automated. Think about all of the manual steps that go into getting the tissue, cutting it with the microtome, the staining procedures it’s incredibly complex.
So Giovanni, you, you’re a pathologist, you spend your lot of time in the labs. How do you [00:21:00] see this? Is it even possible to automate all those things?
Giovanni: Sure, I think so. I most clinical laboratories that have been doing this for decades now, where they started first doing like small steps and now they’re fully automated. The only piece missing in those is the human part that probably will also be caught in a fraction of the time by AI eventually, meaning the final report will be analyzed by some by ai.
First to give you some sort of summary that maybe you just have to approve. And in our case, I think all these steps will be automated one by one, making it easier and less convoluted. And even the fact that we are facilitating now making, reducing the time and delivery of glass because there are no glass to deliver, then we’ll have to use or retrain this people to be your scan technicians or QA technicians that make sure that everything goes well, the entire process.
And I think the, yeah, I don’t, I think everything [00:22:00] automated is not far away I think that everybody is gonna be surprised on how fast this process is gonna be done just because it has been done in another areas of pathology previously.
Now what I would like to see is the glass disappear and going from fresh tissue to images, virtual stains that I think that you may have some insight.
David: I do. It’s once, it’s similar to the AI conundrum, I’m learning that it’s taking baby steps to get there and as we’re doing it we can automate some of the things but there are some things that we create more manual steps, so it’s like, you mentioned in your thoughts on this that you are converting technicians to be the scan technicians. It doesn’t actually solve the problem, it’s just moving one manual labor intensive process to another manual labor intensive process because I think having.
Aleksandra: At least you’re not adding another one.
David: True. You’re, but you’re, but we’re just shifting one pro. Yeah, so I think the question, I [00:23:00] think the question is how can you eliminate some of the manual labor when it comes to, cause you, you still have to make the glass slides and you still have to scan them. So you still, you talk, you mentioned how you can automate some of it, but you still need humans to do a lot of that, and that’s very costly.
It’s very time consuming. Yes avoiding the slide entirely would help with that but was it gonna create new manual laborers elsewhere? I hope not, but I think in the time being from what I’m seeing cause I’m doing it the answer is yes. It’s not perfect and it’s, it’s got a long ways to go.
Giovanni: I think it’s a different type of task, I think it’s less labor intensive, less physical, let’s say, and more intellectual and cause previously you have people walking miles a day delivering glass everywhere, especially in a institution like ours. Now, if you got that and you repurpose these people who want to be trained and we’ve found that’s the best way to do it because you offer, you provide the training, they learn new skills, they [00:24:00] make themselves more valuable.
And rather than walking miles now just they’re supervising that the process is adequate. I think we always need humans involved in all these steps, even if you have a good machine that does the qa, you still need somebody who’s gonna sign on that QA and letting it go, because that’s something that the machines cannot do.
They cannot touch their own errors and for humans usually it’s very easy to see when the machine is making errors because for us are obvious. Those errors that the machine makes are very, they just, if it’s not something that comes in the straight lines that they are used to dealing with, they just shut down and give something that is, is obviously not right for a human.
They don’t work with the system, right? If the system per shows them some variation, they just can’t handle. So that’s where the human comes in place and I think we’re gonna see more of this task raising us the labor intensive board goes down and yeah, that’s gonna be the same story it’ll make us less active and just be most of our [00:25:00] jobs will be probably.
Aleksandra: So are you going to pathology vision this year?
Giovanni: Yeah.
Aleksandra: Yeah?
David: Always yeah.
Aleksandra: So I listened to the last time, Giovanni, you were interviewing people, right? And I listened to the CPT code interview and why do you guys have 13 codes for one thing, scanning slides like I don’t understand.
Giovanni: No it for different purposes. Like one is for the H&E one is for the.
Aleksandra: What is scanning? It’s not on the same thing.
David: That is true.
Aleksandra: Please explain it to us. I’m like, I was like, oh my goodness. So many things are gonna automate 13 codes. That’s amazing. And then I see scanning H&E, scanning IHC, scanning special stain.
I’m like, but the scanner is doing the same to find.
David: At that point, the slide is already made and you’re scanning the same amount of pixels. So it’s not like you’re, it’s not like it’s the manual process of making an H&E versus microwaving slides to get immunohistochemical stains to [00:26:00] penetrate.
So it’s a good question.
Giovanni: I don’t think the final CPTs will be this. Oh, they may be. Because what happened is that they did one CPT or H&E one CPT for immunos, but then they vary by the type of code that you’re charging like, they have a different call for CPT and a biopsy, and then for resection specimen.
Aleksandra: Are they attached to the previous procedure? Like they’re like at the end of each procedure instead of.
Giovanni: Yeah. At end.
Aleksandra: One thing that is happening but they’re already like many streams of CPT codes flowing into it. So we’ll just attach another one to the, I dunno, H&E biopsy processing.
Giovanni: No, you just add one, one of the CPTs to one of your original CPTs.
Aleksandra: Okay. Yeah, so it’s.
Giovanni: You don’t make, if you have a biopsy CPT, then this will have a new code, CPT or the digitization of that, and then you have immunochemistry that will have one that goes with the biopsy. Now, if this was a resection, it’ll have a different CPT here, and the CPT that you’re adding is also different and that’s why you have all these [00:27:00] numbers because they talk about the, all the permit.
David: Yeah, that makes sense because it’s, what you’re saying is that includes, there’s. You’re charging one CPT code for digital evaluation of immunohistochemistry slide. That includes going back to the actual physical processing of that slide.
So, which is a more, yeah, which is a more complex processing than H&E so therefore should have a higher value.
Giovanni: Yeah. And this is the production I and I wanna clarify that because all of these CPTs are for technical. Technical work, there is nothing involving analysis it’s just for the technical work of making the slide, the glasses slide H&E or making the immunohistochemistry, where now you go through the additional step of digitizing that.
So that’s what the H&E has one code, the immunohistochemistry has another code and that’s is just a technical one.
Aleksandra: Okay. But question couldn’t one code for scanning be added to each of those [00:28:00] procedures or.
Giovanni: I wonder, I asked myself the same question and I answer myself. I know if I’m right, but that this is, you know how we like to make everything complicated, right?
So this is the way of making it complicated but actually not. It’s not the reason. It’s that. No, what I actually told myself is that probably they want to kind of see where the CPT are coming, is everybody.
Aleksandra: So much of which is being done.
Giovanni: Yeah. Is everybody doing biopsies scan? It’s only big resections, it’s only breast, that’s, there is no way to know that. Just to give you an example that big wanna probably a little degree of granularity to the data.
David: Giovanni, how are you using the new, trial CPT codes right now?
Giovanni: That’s a good question. I’m embarrassed to say that we’re not. We’re not using it.
David: But that’s a problem that’s a problem. If we’re gonna get this adopted, that’s a problem.
Aleksandra: But you don’t get any reimbursement for them, they’re like just.
Giovanni: No, the reason is that we don’t want, the reason is us.
Aleksandra: Ah, okay.
Giovanni: When we made our happy plans to start, we’re gonna start on [00:29:00] first, but unbeknownst to us was all these different steps that you have to go through to make those CPT. Think about. Then we have to change all the workflow. Make sure that because we produce so many cases today, there is no way that you’re going be implementing this to every case. So we wanted to do it in the most automatic way possible.
And for that we have to iron out a lot of little things with IT, with the coders and also with the revenue people. Because there are concerns that when you start billing with CPT codes that the person receiving them that say no where they are, that they might just stop the process of paying altogether.
And this is not with, this is not with cms, but remember these are targeted to go to CMS into TMA, and they are gonna be, the ones making the decision when they go through and the value assigned to it. But this really goes to all payers and all payers needs to be alerted of what’s going on and [00:30:00] most of them, you will say, oh, they’re payers, they should know, but they don’t.
Or at least the person doing the receiving that call might not be aware, oh, these are CPT 3 calls I’m supposed to ignore them, they just don’t know. So they put them in a pile or I don’t know and then they get into this process of they need to be dealt with later.
So they in some cases the hospitals, if they have dozens of payers, hundreds like us here at osu, they have to make sure that’s not gonna happen otherwise you have these claims that are not address because they just went with information that the receiver didn’t know what to do. So all these little things, that’s the reason why we are not doing it yet, because we we thought it was gonna be just adding the CPT code at the end and people were gonna ignore, then the payers was, we’re gonna ignore because doesn’t. So we had to kinda figure it all out.
David: Yeah.
Giovanni: So we hoping that, Soon we’re gonna start using it.
David: That’s an interesting dilemma. The [00:31:00] CPT codes, for listeners that haven’t been following this, they don’t have value yet, they have to be used to show that they do have value. But what you’re telling us is.
Aleksandra: Nobody wants to use it because they comply well.
David: They can’t use right now because incredibly complex new CPT codes. So we’re gonna have to re revisit this again in six months.
Aleksandra: Yeah.
Giovanni: I’m gonna tell you it’s not for us being a big institution in state institution, it’s been harder. All the ones that I know that are using them now are more compact last.
David: I think we are wrapping up cuz we covered three awesome topics and we’ll have to revisit this.
We’re doing this every six months now. So that’s the plan is to.
Aleksandra: Yes, let’s put on our calendars.
David: Do the hot topics and we’ll release that to you guys and we can’t wait. So thanks everybody.
Giovanni: Okay guys.
Aleksandra: Thank you so much. High five. Bye bye. Bye bye.
