AI Medical Interpretation in an Ophthalmology Practice

Rivka Allouche (00:08)
Welcome back to Care Culture Talks. I'm so excited about today's guest, Corinne from Pacific Eye, honestly a masterclass in Change Management. She runs daily operations for a leading multi-location surgical team orchestrating clinicians and billing, scheduling and patient service into a seamless engine. But here's what makes her story so good

is that she identified a challenge and she implemented a change. So this would be my first question. Tell us more about your journey at Pacific Eye.

Corinne Knowles (00:43)
Thank you, Rivka. I'm Corinne, Executive Director for Pacific Eye Associates. As you said, we're a multi-specialty practice. We're actually the most comprehensive multi-specialty ophthalmic private practice in the San Francisco Bay Area. And so naturally we have a very diverse patient base.

And so I represent surgeons across different cultures and languages and patients across different cultures and languages. And we found that we were very much in need of a solution that could help us better communicate with our patients.

Rivka Allouche (01:18)
And at the very beginning, how did you identify

that there was a challenge or there was an issue?

Corinne Knowles (01:24)
So the history of Pacific Eyes, we actually did have at one point two different full-time interpreters, one for Chinese, Cantonese and Mandarin, and one for Russian and Polish. And while these were incredibly helpful staff,

When they moved on, we struggled with a solution that could scale to the size of our practice and reach all of our patients as they needed. So what we ended up transitioning to before No Barrier was a live call-in agent service.

The nice part of that was that we could lean on and know that we could utilize whatever language we were needing that we would be able to connect with a live agent eventually. But we found that the wait times on these calls and the repetitions of what we had to go through in these calls was quite a barrier, if you will.

So we knew that we needed something that would help us be able to streamline, be more efficient, and have more correct language. We found even these live interpreters did not always understand and did not always have all of the terminology that we needed to properly communicate medical procedure and business structure to these patients.

Rivka Allouche (02:49)
And just to give us a bit of data, what's the patient population breakdown? So I heard you said more about Asian languages, but what would be your top five languages right now?

Corinne Knowles (02:59)
Yeah, it's very diverse. Our top languages would be Cantonese and Mandarin. So Chinese first, very close behind is Russian. And then Spanish, Vietnamese, Korean, Japanese. We honestly see on average more than 20 different languages being interpreted for our patients throughout the year.

Rivka Allouche (03:19)
So before No Barrier you had to invest in several equipments like iPads for like how many iPads for how many doctors and nurses?

Corinne Knowles (03:27)
It was like maybe even a dozen going around the practice. We found obviously a patient's journey isn't just lying in that exam room. They had to be able to communicate with front desk, with our schedulers, with surgical coordinators. They had to be able to communicate with their technician to understand their workup and instruction for different types of testing. And then finally also,

in exam room in that chair facing their clinician and their physician.

Rivka Allouche (03:55)
So I assume that your staff on the front line just came to you and say, okay, it's just, it doesn't work. And then you made the research or how was it?

Corinne Knowles (04:03)
Honestly, the first pain point came directly from our physicians. They found it was really tough to one, track down the equipment and the tablet to make sure that they had the equipment necessary to call in to an agent. Two, if an agent was previously called with a technician or with front desk, that didn't mean that that call was active and ready for the physician once they got into exam room.

So, queue in wait times. We would on average be waiting like five or 10 minutes for these patients to even get on the line with a live agent. And then the repetition. There was a lot of difficulty for the agent to be able to hear or understand the physician and difficulty in hearing or understanding the patient.

And so it became very arduous to go through even basic conversations where previously, like live agents and multilingual, like bilingual staff were able to run that much faster. So we knew that our exam times were going up, our patient appointment times were going up

while our volume was also increasing. And so we needed a new way that we could better serve these patients without all of these roadblocks and these bottlenecks of communication.

Rivka Allouche (05:19)
So you were saying that even the technology, like the audio was not so good, not even for someone who has an ear issue or who is hearing impaired, like just for a standard patient, let's say.

Corinne Knowles (05:31)
You make a good point. In ophthalmology, our average patient age is between like 40 to 70. And so it's not always just that language barrier, but also audio issues, obviously site issues. So trying to communicate with these call-in agents, there was no way to be able to like read or transcribe what was being said. And there was definitely

a roadblock in just that repetition. If a patient's already having difficulty hearing, sometimes it would almost turn into a shouting match between the call agent and the physician, the call agent and the patient, to finally trial and error into what is attempting to be said.

Rivka Allouche (06:14)
So how was the switch from the traditional interpretation method to the new technology? So like from the iPad and like the visual interpretation right now, so you have like both audio and visual.

Corinne Knowles (06:26)
Yes, so now with No Barrier, we get that audio and visual, and I have to say it's one of my physician's favorite things. For one, there's no more repetition. The AI understands what's being said. It feeds it back and it transcribes in both languages for the physician and for the patient. We found even significant others, family members, have been able to

like review the transcriptions and be able to ask follow-up questions. The technology is intelligent enough to recognize like not just two people in the room, but if there's that significant other asking questions or even a tech that's coming in to help clarify on like prescription instructions, it can capture and understand that and feed that back.

My physicians have very much appreciated being able to look through those transcriptions and better understand what the patient is communicating to them, and also to be able to show those transcriptions and for the patients to see those transcriptions in real time to better understand their clinician.

Rivka Allouche (07:27)
So it's very interesting what you're saying. It means like also your patient from 40 to 70 plus, they are also willing to use the technology and they seem to be happy with the technology as well.

Corinne Knowles (07:38)
It's incredibly user friendly. It's very simple and the text is very easy to read off of the transcription side. The audio comes in very clear and honestly competes with our native speaking employees. We have some staff or like we're lucky enough to have staff that speak Cantonese or Mandarin. We just can't clone them across 20 exam rooms and have them be in 40 places at once across different clinics.

But we found that the technology is intelligent enough to compete with those bilingual staff and patients have been really grateful and quite happy with it.

Rivka Allouche (08:17)
I'm so happy to hear. And can we say that you completely switched from phone-based interpretation to AI interpretation right now, or you are still using phone-based interpretation?

Corinne Knowles (08:29)
I'd say that the only time that we've found potential need for phone interpretation is once you get up into a large group. If there's a lot of sound going on in a room or there's more than like three, four people that are trying to communicate at once, it can be more difficult for the application to recognize who is speaking. But honestly, once we swapped to No Barrier, we were able to drop off almost 100%

of our other live calls. It was that fast and it was that simple. And really like it just, it was like a night and day difference across the clinic, honestly.

Rivka Allouche (09:06)
Wow and this leads to my next question. So do you have a practical case of human escalation for instance from the time you've been using No Barrier? Like maybe a patient that required to switch to a human interpreter or provider that just felt that it was better to use a human interpreter?

Corinne Knowles (09:24)
We have seen some cases of that, but honestly I have had patient cases, especially with Cantonese, Chinese, in which maybe they begin a conversation with one of my live staff, for example, like our technician trainer. They might start a conversation in, say, Mandarin.

And then maybe this trainer needs to be pulled into a more emergent role. We've actually like seen a pretty seamless handoff from a live conversation in native tongue to be able to move to the AI for the remainder of their visit or for the time that they're sitting in with physician.

And our patients have actually asked and been surprised that it's even an AI that is speaking to them. They're surprised to hear. We're careful to go through consents. So that way all patients understand the technology that we're using with them and that they consent to that. But to have it feel like essentially there's a native speaker in the room with them

even after they just finished an in-person conversation, it really speaks to the advance of the technology and the success in the technology.

Rivka Allouche (10:33)
I'm so happy to hear and I heard from Eyal, our CEO, that your team was so amazing. It didn't even go through a proper, usual training. So can you tell us, how did you manage? What's your recipe?

Corinne Knowles (10:49)
It was thorough integration across the practice, all parties, and then we ended up going down into smaller subset training with our super users. I will say that we were careful to have our super users, our higher admin staff, be able to take a little bit of sit down time with the physicians to ensure that they understood the application and the technology.

But honestly, we're talking about maybe a five, 10 minute training for each of these physicians. And then for lower level staff, testing it on each other once or twice, solidified the process and made it really simple to just bring in and at the click of a button, be able to start utilizing with live patients.

Rivka Allouche (11:36)
Wow, I'm so happy to hear you know because there is like on one hand we are working you know to be as seamless as possible but then when we get such a feedback it's such a recognition.

And what about the adoption? Because for people like you and us, we know that the technologies belong to a flow and to operations. And sometimes we just integrate some new tools, but it's not adopted for some reason. So how was the adoption?

Corinne Knowles (12:02)
Our staff saw essentially overnight how much this affected their workflows for the better. And so it was like a no question for our staff to start utilizing No Barrier over previous solutions. We saw even just overnight how much faster like patient testing was able to run through.

How much faster we could go through workup, how much more seamless, and honestly, what a better and more positive patient experience they were having behind those closed doors of exam room with a physician. And so even those that have been with us for many years and had seen us work with live interpreters, previously seen us work with live call agents,

were very impressed with the technology and started to utilize right away.

Rivka Allouche (12:54)
Wow. And you talked just before about consent, just to make sure that patients are aware that you are using AI. So the question is, did you build some protocols just to explain when do you need to use AI, when do you need to use human? Did you have to go through that in your operations?

Corinne Knowles (13:14)
No, it's more just making sure that all patients are aware of what their experience may be. So we do include like in our paper consents, the inclusion of like all technologies that may be used across our staff. But then when our live staff are approaching patients, they introduce themselves and they'll also like verbally acknowledge the technologies that are at hand and confirm verbal consent with the patient

before moving forward. Honestly, we're very based in patient education and we see it's essential to the patient experience. And while I know that there is some skepticism across different demographics for things like AI, we found that being very candid and straightforward about our processes has made it a really simple thing for our patients.

I have not seen one patient yet decline and I have not seen one patient yet complain about the product. So it's going very well.

Rivka Allouche (14:15)
So I wanted to ask, so was the next question, we talk about skepticism. So is there at least one member of your team that maybe is not like a hundred percent and so that we can hear about it because it can exist?

Corinne Knowles (14:29)
I do have, I have one physician that was with us last year that is multilingual, like trilingual. And he said he is a huge fan of the AI and started adopting the No Barrier as soon as we brought on. But he did make the comment that it's very fascinating technology and it would be interesting how multilingual users like himself might be able to better train it.

Rivka Allouche (14:54)
Wow, yeah, it's good to know. So which languages is he speaking? So also Cantonese and Mandarin or...?

Corinne Knowles (15:01)
actually a Japanese Korean.

Rivka Allouche (15:03)
Okay, I see. So we can say that it's a very positive outcome. Can you tell us more about efficiency and capacity and throughputs, the KPIs that the technology improved in your organization?

Corinne Knowles (15:16)
It was incredibly scalable. We knew that was one of the major hurdles, like major KPIs that we had to meet was that scalability. We're very large multi-specialty practice, obviously. There's many different facets and variables that are ongoing, even just for a single patient appointment. They might be seeing multiple clinicians in an appointment. They might be obtaining multiple different kinds of testing or different types of procedure.

And so being able to run through specific medical terminology and language to many different patients across different rooms, across different areas of the practice, this was a must. No Barrier was incredibly simple because we could have as many users as we need to sign in, as many devices as we need. It's as simple as an app on mobile

or tablet as you see need. And we found that dozens of our staff at once can have this in use across different areas of the practice. So no more bottlenecks, no more waiting around for the device to be at use, no more waiting on a call, like prompting into the application and starting a conversation with the patient is immediate.

Honestly, it was time more than anything. We were able to drop down time, all of that wait time.

Rivka Allouche (16:35)
When I talked to Eyal yesterday, he told me that when you're on boarded, there was like one KPI that was very striking. It was the duration of an encounter and it was an average of 22 minutes. And like last time he checked, it was like 12 minutes. So like, do you confirm this number? Is it still going as smooth and is it still going as down?

Corinne Knowles (16:49)
Yes.

Yes.

Yes, what's incredible is that an encounter doesn't just drop essentially in half. That we were able to like seamlessly integrate and like pass off from one staff to another. And so all of a sudden, like one encounter wasn't just one call to one agent for one test.

We could seamlessly be transitioning a patient from their workup and their testing into exam room and continue to utilize that technology across that encounter. So just the smoothness of the transition and the lack of wait time involved, it was incredible what it did to our workflows. And our physicians have been singing its praises ever since.

I'd say if anything, it's only gotten more smooth as our team has regularly adopted and our patients recognize the technology and trust it. And so each follow-up appointment just becomes that much easier.

Rivka Allouche (17:53)
And do you have like data showing that maybe like your physicians could like meet more patients? Do you see that in your figures?

Corinne Knowles (18:01)
Yeah, some of our clinicians actually increased their volume by 30 % over two seasons after adopting No Barrier, which was huge. We had clinicians that were purposefully having to add blocks to their schedule or extend the time blocks for certain appointments, extend time blocks for certain patients, just with the knowledge that there would be some type of...

barrier of language and that there would be issues involved with language translation. Now all of that has gone away. I don't get a single request from my clinicians with concern for language. We can treat our patients of whatever language they speak as first language and we can have that follow them through their appointment in a primarily English speaking practice. It's amazing.

Rivka Allouche (18:51)
Because right now, let's say in one day, how many patients are considered as LEPs?

Corinne Knowles (18:57)
As English second language patients are up to 40 % of our patient base. It's very diverse here. So for us to be able to treat them all as the same time slot and to get that same kind of patient care,

Rivka Allouche (19:02)
Okay. Okay.

No scheduling.

If other leaders like you, they want to learn from you and like maybe you can share best practice. What would be the like the best practice you would give to someone who wants to drive change in their clinical operations?

Corinne Knowles (19:26)
I would say don't just take word for it. We interviewed many, many different solutions and investigated many different options off of this. But really, like where we found the value in No Barrier was putting it to the test. We actually brought the technology in with multiple different like bilingual, multilingual staff

and tested the technology to see if it can recognize different accents and different dialects of language. If it could understand both users speaking language as their second language versus their primary language. And it came up with glowing reviews across all tests. And so I would say like, put it to test, give it a test

and see how your users actually respond to it. And that's really where the value is, is putting it to work.

Rivka Allouche (20:19)
And

how many vendors do you recommend checking in the process?

Corinne Knowles (20:25)
We interviewed, I think, at least four others, including live and AI agents. And I know that this market is really booming. I don't think it hurts to hear from what options are out there, but I have to say that No Barrier really blew us out of the water with all of the nuances of language and the considerations of language that your team has been able to bring into the technology.

It was truly unrivaled.

Rivka Allouche (20:53)
Wow. And look from like, let's say if we put the challenge on the timeline, from the time you considered there was like a true gap to fill to the time you adopted No Barrier, like how much time spent.

Corinne Knowles (21:06)
I'd say that we took our time with the process and maybe we spent at least a good two months investigating. But once we acknowledged that No Barrier was where we wanted to start, it was a very quick turnaround to our go live. it was, like I said, it was very smooth and very easy training, unbelievably simple. And that made it just a no brainer for our staff.

Rivka Allouche (21:27)
I think we covered a lot, in your expertise of operations and flows. Are there any insights you want to share among everything that we said already?

Corinne Knowles (21:38)
I would just say don't be afraid of AI. I've gone to recent conferences that are still really skeptical of the capabilities of this technology. But I'd say it by and large, it's not replacing the live staff that we have here. It's not replacing our physicians, but it is helping elevate our staff's capabilities and helping provide a higher quality of care with our existing live agents

with this assistance of AI and technology. So I can't recommend it more.

Rivka Allouche (22:07)
Thank you very much. And Corinne, I thank you so much for sharing your journey with us today. We are truly happy to be your partner at Pacific Eye. And right now, I understand the enthusiasm of our CEO yesterday when we talked about your engagement and your adoption. Thanks again. And I can't wait to host you again, maybe soon, to have a little bit of feedback of adoption and how it is going. Until the next time.

Corinne Knowles (22:23)
Ha

Well,

we'd love that. Thank you. We've appreciated your time. And I know that we have an ongoing relationship of feedback with your company. So we really appreciate it, Rivka.

Rivka Allouche (22:45)
Thank you very much.

‍