WEBVTT
00:00:01.439 --> 00:00:14.320
Welcome to the On Record Podcast, a podcast where we talk with professionals and subject matter experts from diverse markets that are using video and audio technology in innovative ways to support their organization's objectives.
00:00:14.560 --> 00:00:18.320
My name is Carmen Hawk, and my co-host today is Kyle Schalstedt.
00:00:18.559 --> 00:00:32.880
Today, our guest is Annie Holland Temple, a faculty member in the Communication Sciences and Disorders Department at the MGH Institute of Health Professions, one of the country's top-rated SLP programs and part of the prestigious Mass General Brigham Health System.
00:00:33.200 --> 00:00:38.159
Annie's also the inaugural recipient of the Capsid IVS Clinical Education Seed Grant.
00:00:38.320 --> 00:00:39.119
Welcome, Annie.
00:00:39.280 --> 00:00:40.320
Thank you for joining us.
00:00:40.640 --> 00:00:41.439
Thank you.
00:00:41.759 --> 00:00:48.479
Annie, would you share a bit about yourself with our listeners, about your journey and how you came to be in your current role at MGH?
00:00:49.280 --> 00:00:49.679
Sure.
00:00:50.079 --> 00:00:58.960
I'm a speech language pathologist and I had always been really interested in communication and connection from a very early age.
00:00:59.119 --> 00:01:06.959
And that really led me to the profession when I was, you know, selecting what I was going to do for my career.
00:01:07.120 --> 00:01:12.560
And then as I was practicing as a speech language pathologist, I specialized in pediatrics.
00:01:12.719 --> 00:01:17.200
I loved working with my clients who were very young and their families.
00:01:17.439 --> 00:01:28.719
And as I worked more with them, I really found myself wanting to teach more and, you know, engage with people who were interested in the field, students who were pursuing a degree in speech pathology.
00:01:28.799 --> 00:01:45.680
And so as I started supervising more students and engaging more in teaching, then I transitioned to my current role where I'm teaching full-time and I get to teach graduate students, mostly speech language pathology master students, but we also do a lot of interprofessional teaching as well.
00:01:45.920 --> 00:01:48.079
So that's how I ended up where I am.
00:01:48.239 --> 00:01:51.439
And it's it's been such a such an awesome journey.
00:01:51.519 --> 00:01:53.359
And I I love every minute of it.
00:01:53.680 --> 00:01:57.120
I understand that you've worked across some really diverse clinical settings.
00:01:57.280 --> 00:01:59.280
Would you share more about that with us?
00:01:59.840 --> 00:02:00.239
Sure.
00:02:00.400 --> 00:02:17.439
I think that's one of the big benefits of being in a health profession like Speech Language Pathology, is that you really can work in different kinds of settings with different groups of people as your interests change or as your personal needs or family needs may change.
00:02:17.759 --> 00:02:22.319
And so I've been really fortunate to be able to work within school systems.
00:02:22.560 --> 00:02:27.680
I've worked in pediatric hospital systems, which has been really excellent.
00:02:27.840 --> 00:02:34.400
In my time at the pediatric hospital, I was doing some work in the neonatal intensive care unit, which was awesome.
00:02:34.560 --> 00:02:36.960
Our neonatal follow-up clinic there.
00:02:37.680 --> 00:02:54.560
So lots of lots of different ways have I've been able to work with children and their families in the speech pathology realm and working on communication and then also on feeding, which is another aspect of speech pathology that we get to work with children and adults on is feeding and swallowing.
00:02:54.639 --> 00:02:57.039
So doing that in different settings too.
00:02:57.439 --> 00:03:00.719
Does that follow or does that uh fall into your published research?
00:03:00.879 --> 00:03:04.080
Will we tell us more about your published research that you focused on?
00:03:04.639 --> 00:03:05.039
Sure.
00:03:05.199 --> 00:03:15.840
Um, one of my more recent publications is within my teaching of speech pathology graduate students, one of the things that I teach a lot about is clinical writing.
00:03:16.000 --> 00:03:24.400
And so I have really seen the power that good clinical writing can have and also other clinical writing can have.
00:03:24.560 --> 00:03:32.800
You know, the way that we write about the sessions we conduct can really do a lot to educate a family on our impressions.
00:03:32.960 --> 00:03:38.159
It can let other people on the care team know what we're thinking or what the recommendations are.
00:03:38.479 --> 00:03:48.560
And clinical writing that's done in such a way that is inaccessible or laden with biases and things can do a lot to harm that message.
00:03:48.639 --> 00:04:03.840
And so I really feel passionate about teaching my graduate students how to be really strong clinical writers so that all the important work that we do is communicated in really accessible and accurate ways for the family, for the patient, and for the whole care team.
00:04:03.919 --> 00:04:09.439
So everybody knows what's going on, what we're thinking, what we bring to the care plan.
00:04:10.240 --> 00:04:17.519
And so one of our more recent projects we did was have the students look at a piece of clinical writing.
00:04:17.759 --> 00:04:36.879
And in that piece, we put in lots of instances of problematic or biased language with the intention that our students could see some of those pop out and flag them as problematic and have a chance to rewrite those instances.
00:04:37.040 --> 00:04:47.439
And so we gave them that piece of writing and had them go through that exercise identifying any of those pieces of language and trying to correct it.
00:04:47.600 --> 00:05:03.920
And then we engaged them in a training module where we talked about clinical writing, we talked about anti-oppressive practice, we talked about how we can make our clinical writing a lot stronger, a lot more accurate, and move away from using biased language.
00:05:04.079 --> 00:05:10.480
And then we had them do that same piece of writing again and complete the activity after the training.
00:05:10.639 --> 00:05:22.079
And in our research project, we were able to compare pre-top-training and see the change that our students were able to demonstrate after they received the information, which was really fun.
00:05:22.639 --> 00:05:32.639
The ability to keep bias out of this practice seems like it would be incredibly challenging, um, especially uh what you're talking about, clinical writing.
00:05:32.720 --> 00:05:43.920
So, do you have any specific examples of somebody who was able to remove their bias um from their clinical writing or specific tools they might use?
00:05:44.399 --> 00:05:44.800
Sure.
00:05:44.959 --> 00:05:45.279
Yeah.
00:05:45.439 --> 00:06:03.120
I mean, it's important in speech pathology and in other fields, but especially in speech pathology where we work on communication, and communication is so deeply tied to a person's identity, to their culture, to things that are really important for us to know about and to be able to talk about.
00:06:03.279 --> 00:06:26.160
It's important for us to know different aspects of a person of a person's race or ethnicity or cultural factors that are important because when we're thinking about their communication, there are really real um, you know, pieces that we we want to make sure we don't misidentify as a disorder if it's something that actually is very acceptable given parts of their background.
00:06:26.240 --> 00:06:36.720
And so um, you know, in thinking about different times that we sometimes see biases, we have worked with our students a lot about the gender binary.
00:06:36.800 --> 00:07:13.519
So we have students who sometimes come into our program and will come from the lens of having this idea of a more rigid gender binary, and especially in working with children, you know, boy using the He series pronouns only, and girl use it and really working with our students to say, you know, these are pediatric patients and these are they're coming from families of all different kinds and all different, you know, feelings around this, and teaching students how to talk with families about how they want us to address their children and speak with their children.
00:07:13.680 --> 00:07:42.160
I think that's um really important for our students to know that we don't just see something and immediately make an assumption and let that sort of direct our work and show up in the writing, but just really open things up for conversations with our families and with our patients so that we're using the language that um matters to them and makes sense for how they want us to talk with them, talk about them, and that carries over into how we write about them.
00:07:42.959 --> 00:07:44.079
I that's excellent.
00:07:44.240 --> 00:07:50.560
I was gonna ask you to elaborate on some examples of why it matters so much in speech language, but you you tackled that.
00:07:50.720 --> 00:07:56.720
That's incredibly interesting from the perspective of healthcare in general, but speech language pathology specifically.
00:07:56.959 --> 00:08:01.519
Um, is there a room for AI to assist in this space?
00:08:01.759 --> 00:08:05.759
Or is this um human thought only that should be documented?
00:08:06.399 --> 00:08:10.160
I think we've seen so much good that can come from AI.
00:08:10.319 --> 00:08:21.279
I think, especially for me, I love learning about new technologies that can help me spend more time with the people that I work with and spend more time in my teaching with students.
00:08:21.360 --> 00:08:31.519
And so I'm really excited to learn about how artificial intelligence and different programs can give me back some of that patient care or student teaching time.
00:08:31.680 --> 00:08:38.639
And I think that we're seeing a lot of strengths in some of these programs for assisting us in writing our clinical reports.
00:08:38.799 --> 00:08:45.759
It's an important piece of any speech language pathologist and many other healthcare professionals' day is the documentation.
00:08:45.919 --> 00:08:51.360
It's important for reimbursement, it's important for communication among the care team.
00:08:51.519 --> 00:08:55.120
So it's something that we all need to do in our daily practice.
00:08:55.200 --> 00:09:13.519
And we're seeing that there are many AI programs that are able to take in a session or take in our notes and create something that could be very workable and could save us a lot of time as far as um, you know, a session note or an evaluation report, even.
00:09:13.759 --> 00:09:17.840
So we are seeing a lot of strengths, but then I think it's also important.
00:09:17.919 --> 00:09:48.320
And the reason why I engaged in this particular research study I'm doing now is we also have to be really aware and open about the um the risks as well, because there are um things that we're seeing in the output of generative AI platforms that we need to be aware of and we need to correct, and we need to make sure don't perpetuate um, you know, long-term biases and healthcare inequalities that we know um exist and can easily be perpetuated.
00:09:49.200 --> 00:10:02.399
Do you see these tools as potentially replacing speech language pathologists, or are they more of just a way to assist you in your workload?
00:10:03.279 --> 00:10:37.600
I think one of the things that really makes me happy about speech language pathology right now, and I think for the future, is that it is pretty resilient to um being fully replaced by AI systems because the human connection piece of the communication and the feeding that we do is something that incorporates a human-to-human um response and interaction that cannot be duplicated by these various um algorithms or machines.
00:10:37.759 --> 00:10:49.440
But I do see a lot of artificial intelligence really assisting us and being able to do our jobs better, not only with the clinical documentation, but in our augmentative communication systems.
00:10:49.759 --> 00:11:02.320
The AI that's being put into systems that are helping give voice to the people that we work with are showing such incredible strides in what they're able to help our patients communicate.
00:11:02.480 --> 00:11:17.919
And so, in those ways, knowing how AI works and being able to be the human that uses the AI in responsible and smart ways is helping our patients make so many more improvements so much faster.
00:11:18.000 --> 00:11:22.960
And so I do see it as something that really, really supplements and enhances what we do.
00:11:23.279 --> 00:11:36.000
But I'm happy to say that I don't think it can replace the human component that is so um so important to speech language pathology, communication, the feeding, um, but it can really help.
00:11:36.399 --> 00:11:38.000
That was a really brave question, Kyle.
00:11:38.080 --> 00:11:41.039
I wasn't sure how that response was getting on.
00:11:41.120 --> 00:11:56.639
I'm just the reason I ask is because like you always see people marketing their AI tools as as a a piece that can, you know, one, make your job easier, like this is the solution for everything.
00:11:56.799 --> 00:12:12.960
Uh whereas like I just I wonder if clinicians, uh pathologists like yourself um are necessarily interested in that or just interested in ways to kind of lighten your workload, make it give you more time to focus on the patient.
00:12:13.039 --> 00:12:17.440
Um I never even thought about the augmented the augmentation part for the communication of the patient.
00:12:17.600 --> 00:12:19.039
I never thought about it that way, too.
00:12:19.360 --> 00:12:27.360
It's not just the benefit of uh the communicator, it's also like the learner or the the recipient of the therapy.
00:12:27.440 --> 00:12:28.240
That's fascinating.
00:12:28.559 --> 00:12:34.559
Sorry, there was I don't there was a question there, probably Kyle, but I thought that was cool about the augment um augmented communication.
00:12:35.120 --> 00:12:35.600
For sure.
00:12:35.840 --> 00:12:47.679
Yeah, and we know we've had certain aspects of our jobs that have had artificial intelligence sort of woven in for a long time that we have really um benefited from our patients in using.
00:12:47.759 --> 00:13:09.200
And so there are aspects of the care that we provide and some of the assessment and treatment that we do that have been successfully transferred over to certain AI systems, especially if there's a patient who's working on voice and they're able to see their um their various voice outputs on a computer and be able to self-monitor.
00:13:09.279 --> 00:13:11.759
It's especially powerful for adult clients.
00:13:11.919 --> 00:13:24.879
So there are particular things that we had always been doing face-to-face provided by humans, that we transition certain things over to using AI systems with the help of a human facilitator.
00:13:25.039 --> 00:13:33.360
Um, and we're really open to that because again, it makes it makes us be able to do our jobs better and it makes us um even more effective.
00:13:33.519 --> 00:13:40.240
So I think um there's always always room for the human connection piece in speech pathology.
00:13:40.480 --> 00:13:55.519
We, you know, are still going to be the humans driving behind all of these things, but I think it's important for us to learn about the advances and be able to bring these things in, especially where it makes sense given the diagnosis.
00:13:56.799 --> 00:14:02.480
And you had introduced a little bit about what that current research looks like and how it relates to the conversation we're having.
00:14:02.639 --> 00:14:04.080
Well, the cat's out of the bag.
00:14:04.159 --> 00:14:08.799
You are our inaugural recipient of the CAPSID IVS Clinical Education Seed Grant.
00:14:08.879 --> 00:14:10.399
So, congratulations to you.
00:14:10.639 --> 00:14:19.440
Um for our listeners, the IVS Clinical Education Seed Grant is a financial award that was offered through CAPSID and funded by Intelligent Video Solutions.
00:14:19.600 --> 00:14:26.399
It's designed to support innovations in clinical education, training, and service delivery and communication sciences and disorders.
00:14:26.639 --> 00:14:32.320
Annie, what does this funding now make possible that you may have not been able to do otherwise?
00:14:32.799 --> 00:14:33.279
Yes.
00:14:33.600 --> 00:14:46.639
So we've been so happy to be able to use this funding to create a research project where we have generated AI platforms generate speech language evaluation reports.
00:14:46.799 --> 00:14:51.519
So part of this funding has made it so that we can access various platforms.
00:14:51.600 --> 00:14:57.759
We don't want to be biased and only look at one program and how that is generating speech and language reports.
00:14:57.840 --> 00:14:58.879
We wanted a variety.
00:14:59.120 --> 00:15:03.600
So in that way, we've been able to use this grant for access to those.
00:15:03.759 --> 00:15:18.159
And then we also have used this grant to be able to put together a team of experts, which is really important in this kind of research because we know that when we're talking about biased language, there's so much that comes with that.
00:15:18.399 --> 00:15:22.480
Language is an ever-evolving piece of human life.
00:15:22.720 --> 00:15:33.279
And we, when we're thinking about what we would consider to be biased or problematic language, we need many different minds to think about that.
00:15:33.360 --> 00:15:46.639
And we need people who do research in that particular space who have interests in AI as well, to really be able to comment on what they see to be potentially biased language in those reports.
00:15:46.799 --> 00:16:01.440
And so this C grant has allowed us to put together a team of a variety of individuals, all of whom have extensive experience in biased language, all of whom also have a strong interest in artificial intelligence.
00:16:01.600 --> 00:16:03.679
And they come from all over.
00:16:03.840 --> 00:16:12.879
We connected with colleagues from many different universities, healthcare systems, and we were able to present our ideas for the project.
00:16:13.120 --> 00:16:25.120
And every single one, for the most part, was fully on board to review the reports generated by the various AI platforms and then comment on the bias language that they found.
00:16:25.200 --> 00:16:38.559
And so it's really made it possible so that we could assemble that team and be able to generate these reports in a way that really is representative of how people are working right now.
00:16:38.960 --> 00:16:54.080
Because we also wanted it to be using systems and using those systems in a way that we know our students are using in their work right now with their in their clinical placements and that practicing speech pathologists are using right now to help them in their work.
00:16:54.879 --> 00:17:08.960
If you were referenced by a colleague to help them implement AI documentations tools in their student body or to help them use and implement documentation tools with AI, is there a warning you'd give?
00:17:09.119 --> 00:17:10.319
Is there a caveat?
00:17:10.400 --> 00:17:15.839
What would you advise if someone was looking to explore these tools, maybe in an adjacent industry?
00:17:16.880 --> 00:17:17.279
Sure.
00:17:17.599 --> 00:17:36.799
I would definitely point them to the prompt research, the research that we've seen in terms of how we prompt these systems and letting that organization know that there are risks, there are biases, and we need to be sure we're doing this in really smart and ethical ways.
00:17:37.039 --> 00:17:56.559
And directing them to some of the literature in prompting these systems so that the output they receive is hopefully the best representative of the output because we've been learning a lot about the way that these various platforms are prompted can result in a lot of different kinds of output.
00:17:56.720 --> 00:18:13.759
So that's that's what I would probably mostly caution the person against is not to just fly in with no plan because I there are these risks of problematic language showing up, perpetuation of bias, um, you know, some issues with ethics here.
00:18:13.920 --> 00:18:30.720
I would say tailoring your prompts and making sure that the people in your organization are using really well-written prompts so that the output is more of what they're looking for, and so that they aren't having to prompt and reprompt and edit.
00:18:30.799 --> 00:18:41.039
Because again, we know that with that comes more and more use of these systems and that has environmental impacts, you know, the amount of times that we're doing all of these things.
00:18:41.119 --> 00:18:48.960
So if we can teach people the importance of prompting in very smart and efficient ways, that's another way that we can support better outputs.
00:18:49.359 --> 00:18:50.960
I think that's very interesting.
00:18:51.039 --> 00:19:00.480
Um, the fact that what you put in will ultimately get you more quality on the other end and how you're wording things.
00:19:00.720 --> 00:19:13.759
Um, obviously, adoption has to be there, and there has to be a belief that this is going to be a tool that helps save time or or money or has um a good investment uh on the output.
00:19:14.079 --> 00:19:16.640
Is there adoption concerns?
00:19:16.799 --> 00:19:18.640
And do you see peers or colleagues?
00:19:18.799 --> 00:19:21.279
You are working with a lot of people excited about AI.
00:19:21.519 --> 00:19:24.000
What does the other side of the coin look like?
00:19:24.240 --> 00:19:25.119
The resistors.
00:19:25.839 --> 00:19:26.880
Yes, for sure.
00:19:27.119 --> 00:19:43.039
So when we started as a department talking about how we were going to bring AI into our teaching and think about it for teaching our students, we know that practicing clinicians are using this, so we know our students are going to be asked to use this in their clinical placements.
00:19:43.359 --> 00:19:51.599
We expected there to be pushback, and and there were some good discussions among different members and talking about it.
00:19:51.759 --> 00:19:57.039
But what we didn't expect was how much pushback we would receive from our students.
00:19:57.279 --> 00:20:18.160
And so that's been a really interesting, um interesting piece that we didn't fully prepare for is that a lot of our students there in their early 20s, and they have a lot of very legitimate reasons for why they do not want to engage with AI in their practice as speech language pathologists.
00:20:18.240 --> 00:20:33.440
And so we've actually had to do a lot of thinking about that and talking with our students about the reality that this is something that they will very likely be tasked to use in their clinical placements or in jobs.
00:20:33.599 --> 00:20:43.519
And so, how can we teach them to use it in ways that feel comfortable for them, given that there's some very strong resistance there for very legitimate reasons?
00:20:44.000 --> 00:20:51.039
Honestly, that's that's something we've seen in past conversations that the demographic is most um resistance.
00:20:51.279 --> 00:20:54.319
You think it'd be maybe the alternative, but no, sorry, Kyle.
00:20:54.720 --> 00:21:00.079
No, I was just yeah, that was I'm surprised that it came from students, not like tenured faculty.
00:21:00.319 --> 00:21:08.400
Um in a this is completely irrelevant, but I saw some reel the other Instagram reel the other day where it was some commencement speaker.
00:21:08.799 --> 00:21:11.920
You're not even getting your news from TikTok against it.
00:21:12.960 --> 00:21:15.039
But but basically the same thing, whatever.
00:21:15.119 --> 00:21:22.319
Uh, but the commencement speakers started like talking about AI and the future of it, and the whole all the student graduates boo, yeah, you saw that.
00:21:22.559 --> 00:21:25.839
Yeah, it was a couple months ago I saw it probably just recently for you on Reels.
00:21:26.640 --> 00:21:31.279
But that was like my first exposure to like, oh, maybe students like aren't fully on board with this.
00:21:31.759 --> 00:21:32.079
Yeah.
00:21:32.240 --> 00:21:46.400
And you know, we make these, we make these assumptions and we think, oh, we know exactly how how to advance our teaching, and we need to bring in the best and the newest and the and the most sort of next next step sorts of things.
00:21:46.480 --> 00:22:04.319
Um, but I think it's really important for anybody in any line of work that we do to to really kind of gauge with our audience, engage with our audience and figure out okay, actually, let's let's take a needs survey here or Let's actually figure out what the students are feeling about it.
00:22:04.480 --> 00:22:10.799
So that again, we need, we have the responsibility to teach them about artificial intelligence.
00:22:10.960 --> 00:22:16.240
We must teach them about how it's being used in speech language pathology clinical practice.
00:22:16.400 --> 00:22:20.000
We would be doing our students a disservice to just say, oh, forget it.
00:22:20.079 --> 00:22:20.799
Then we won't teach it.
00:22:20.880 --> 00:22:23.039
If you, if you all don't want to use it, we won't teach it.
00:22:23.119 --> 00:22:25.920
That would be a real disservice to them.
00:22:26.079 --> 00:22:32.960
So we do need to be honest with them about that, be really transparent with the fact that this is something.
00:22:33.039 --> 00:22:49.359
It's going to show up, and your ability to use it in ethical and responsible ways is going to make it so that it provides the best with the least amount of negative impact or resource use as possible.
00:22:50.079 --> 00:23:02.400
I'm curious, all of the focus that's going on this research and your work in this space, how does that impact your video observation and recording in these clinical education environments?
00:23:02.559 --> 00:23:03.440
Do they go hand in hand?
00:23:03.599 --> 00:23:04.960
Are they separate?
00:23:05.039 --> 00:23:12.319
How are you using them together or do you keep some maintain uh different spaces for these two topics?
00:23:12.880 --> 00:23:16.079
We keep it all folded in together.
00:23:16.240 --> 00:23:26.720
So all of our in our on-campus clinic, all of our sessions are video recorded and we are watching our students via that live stream video in the moment.
00:23:26.880 --> 00:23:43.599
And for that, we are able to give the students the space in that treatment room with the patient and with the family to take some risks and to feel like they can try out things without their faculty member being physically in that space.
00:23:43.839 --> 00:23:57.839
And so in that way, we're able to let our students have some of that autonomy and learn how to be a practicing clinician without us necessarily jumping in at every at every moment.
00:23:58.000 --> 00:24:14.400
And in terms of our use of AI, we are able to fully fold in the fact that they're going to be writing a session note about what's happening in that recorded session for every single time they see a patient in our clinic.
00:24:14.640 --> 00:24:30.480
And so we're able to work with them and replay that video recording of the session and pull things out and, you know, work with them to use different ways of writing up that session note in a way that is accurate to what happened.
00:24:30.559 --> 00:24:38.400
And we're always talking to our students about using objective language and removing any bias language that we find.
00:24:38.559 --> 00:24:52.319
So when we have AI supports in helping us with that session note output, we want our students to know there is a strong potential that there will be some bias or some subjective language here.
00:24:52.480 --> 00:25:11.039
So I need you to be able to identify it and I need you to be able to rewrite those moments and come to me with any questions that you might have about that because it's really difficult, especially as a new clinician who is just learning how to do the job to then learn how to write about it is so new.
00:25:11.119 --> 00:25:18.400
And it's really tricky when you haven't put those various things into words many, many times already.
00:25:18.480 --> 00:25:21.599
It does require some help, and that's why we're there for them.
00:25:22.240 --> 00:25:25.519
So Noli's technology is reflection, is a huge theme.
00:25:25.680 --> 00:25:35.519
Reflecting on the auto prompts, reflecting on what you're getting as the quality output from AI, reflection on the physicality and what's happening in the actual spaces with AV capture.
00:25:35.680 --> 00:25:44.559
What is a positive outcome that has come from using AV, whether it's in tandem with your clinical documentation with AI or if it's just AV?
00:25:45.599 --> 00:26:01.200
I think one of our most favorite things that we do is at the end of every semester, our students pick a video clip from their sessions that they've had in the on-campus clinic, and they share it with everybody.
00:26:01.359 --> 00:26:08.400
So we have a very large cohort of students, and each student has two patients in our on-campus clinic their first year.
00:26:08.559 --> 00:26:11.839
And so at the end of every semester, we all get together.
00:26:11.920 --> 00:26:20.960
That large cohort is able to put up a selected clip by them, and they're able to explain to their classmates here's the patient I'm going to show you.
00:26:21.119 --> 00:26:30.960
Here's what we're going to do in the session, and what I was really proud of, or what I was really challenged with, or what I really would love your, would love your input on.
00:26:31.039 --> 00:26:40.640
And then they show a clip from those recorded sessions, and then every student in the cohort is able to see a patient that they're not familiar with, which is really important.
00:26:40.799 --> 00:26:43.519
They get more exposure and variety.
00:26:43.839 --> 00:26:47.519
And they're able to think about, okay, what would I have done in this situation?
00:26:47.759 --> 00:26:54.319
Or I really appreciated how my peer did something in this way.
00:26:54.480 --> 00:26:59.440
I'm going to do that with my client that I have or my patient that I've got.
00:26:59.680 --> 00:27:16.319
So that's probably the most fun things that we can use our video recordings for is really celebrating each other's successes and exposing the whole cohort to all the various different patients that come through our on-campus pro vono clinic.
00:27:16.559 --> 00:27:19.920
It's really, really fun to do that at the end of every semester.
00:27:20.240 --> 00:27:28.160
Annie, we have a lot of listeners that might be interested in contacting you to learn more about your work or just to follow you and your research.
00:27:28.640 --> 00:27:32.640
How is the best, what is the best way for those listeners to get a hold of you?
00:27:33.200 --> 00:27:33.519
Sure.
00:27:33.680 --> 00:27:36.480
I'm more than happy to connect via email.
00:27:36.640 --> 00:27:42.559
My email is akholland at mghihp.edu.
00:27:43.279 --> 00:27:53.200
And you can find our more recently published report on the biased language in the pre and post-training that I talked about earlier.
00:27:53.359 --> 00:27:59.279
That's in the Teaching and Learning and Communication Sciences and Disorders journal, which is open access.
00:27:59.680 --> 00:28:01.200
So you can find that there.
00:28:01.359 --> 00:28:03.359
I'm also on LinkedIn, Annie K.
00:28:03.519 --> 00:28:04.000
Holland.
00:28:04.319 --> 00:28:07.279
So I'm happy to connect with anyone anytime.
00:28:07.680 --> 00:28:08.640
Thank you, Annie.
00:28:08.799 --> 00:28:09.759
It was very enlightening.
00:28:09.839 --> 00:28:15.119
We learned a lot today, and it's it's good practice and information for every individual human being to hear.
00:28:15.680 --> 00:28:21.680
It's really nice to know that people are uh doing some research on these AI tools.
00:28:21.759 --> 00:28:33.279
And um, and so maybe we have some like uh facts grounded in data um as to how they should be used, and we aren't just kind of like going with it and hoping that it all works out.
00:28:33.440 --> 00:28:36.000
So we really appreciate you doing this work.
00:28:36.480 --> 00:28:38.240
Yes, thank you so much.
00:28:38.480 --> 00:28:47.039
And if you or if someone you know is interested in learning more about the IVS Clinical Education Seed Grant, visit capsa.org forward slash scholarships.
00:28:47.200 --> 00:28:48.240
We'll see you next time.
00:28:48.400 --> 00:28:49.279
Thank you.
00:28:50.559 --> 00:28:58.079
OnRecord is a podcast presented by Intelligent Video Solutions, hosted by Mike Anzolone and Carmen Hawk, and produced by Kyle Shellstead.
00:28:58.319 --> 00:29:05.440
Find us online at ipivs.com slash on record and on social media at Intelligent Video Solutions.
00:00:01.439 --> 00:00:14.320
Welcome to the On Record Podcast, a podcast where we talk with professionals and subject matter experts from diverse markets that are using video and audio technology in innovative ways to support their organization's objectives.
00:00:14.560 --> 00:00:18.320
My name is Carmen Hawk, and my co-host today is Kyle Schalstedt.
00:00:18.559 --> 00:00:32.880
Today, our guest is Annie Holland Temple, a faculty member in the Communication Sciences and Disorders Department at the MGH Institute of Health Professions, one of the country's top-rated SLP programs and part of the prestigious Mass General Brigham Health System.
00:00:33.200 --> 00:00:38.159
Annie's also the inaugural recipient of the Capsid IVS Clinical Education Seed Grant.
00:00:38.320 --> 00:00:39.119
Welcome, Annie.
00:00:39.280 --> 00:00:40.320
Thank you for joining us.
00:00:40.640 --> 00:00:41.439
Thank you.
00:00:41.759 --> 00:00:48.479
Annie, would you share a bit about yourself with our listeners, about your journey and how you came to be in your current role at MGH?
00:00:49.280 --> 00:00:49.679
Sure.
00:00:50.079 --> 00:00:58.960
I'm a speech language pathologist and I had always been really interested in communication and connection from a very early age.
00:00:59.119 --> 00:01:06.959
And that really led me to the profession when I was, you know, selecting what I was going to do for my career.
00:01:07.120 --> 00:01:12.560
And then as I was practicing as a speech language pathologist, I specialized in pediatrics.
00:01:12.719 --> 00:01:17.200
I loved working with my clients who were very young and their families.
00:01:17.439 --> 00:01:28.719
And as I worked more with them, I really found myself wanting to teach more and, you know, engage with people who were interested in the field, students who were pursuing a degree in speech pathology.
00:01:28.799 --> 00:01:45.680
And so as I started supervising more students and engaging more in teaching, then I transitioned to my current role where I'm teaching full-time and I get to teach graduate students, mostly speech language pathology master students, but we also do a lot of interprofessional teaching as well.
00:01:45.920 --> 00:01:48.079
So that's how I ended up where I am.
00:01:48.239 --> 00:01:51.439
And it's it's been such a such an awesome journey.
00:01:51.519 --> 00:01:53.359
And I I love every minute of it.
00:01:53.680 --> 00:01:57.120
I understand that you've worked across some really diverse clinical settings.
00:01:57.280 --> 00:01:59.280
Would you share more about that with us?
00:01:59.840 --> 00:02:00.239
Sure.
00:02:00.400 --> 00:02:17.439
I think that's one of the big benefits of being in a health profession like Speech Language Pathology, is that you really can work in different kinds of settings with different groups of people as your interests change or as your personal needs or family needs may change.
00:02:17.759 --> 00:02:22.319
And so I've been really fortunate to be able to work within school systems.
00:02:22.560 --> 00:02:27.680
I've worked in pediatric hospital systems, which has been really excellent.
00:02:27.840 --> 00:02:34.400
In my time at the pediatric hospital, I was doing some work in the neonatal intensive care unit, which was awesome.
00:02:34.560 --> 00:02:36.960
Our neonatal follow-up clinic there.
00:02:37.680 --> 00:02:54.560
So lots of lots of different ways have I've been able to work with children and their families in the speech pathology realm and working on communication and then also on feeding, which is another aspect of speech pathology that we get to work with children and adults on is feeding and swallowing.
00:02:54.639 --> 00:02:57.039
So doing that in different settings too.
00:02:57.439 --> 00:03:00.719
Does that follow or does that uh fall into your published research?
00:03:00.879 --> 00:03:04.080
Will we tell us more about your published research that you focused on?
00:03:04.639 --> 00:03:05.039
Sure.
00:03:05.199 --> 00:03:15.840
Um, one of my more recent publications is within my teaching of speech pathology graduate students, one of the things that I teach a lot about is clinical writing.
00:03:16.000 --> 00:03:24.400
And so I have really seen the power that good clinical writing can have and also other clinical writing can have.
00:03:24.560 --> 00:03:32.800
You know, the way that we write about the sessions we conduct can really do a lot to educate a family on our impressions.
00:03:32.960 --> 00:03:38.159
It can let other people on the care team know what we're thinking or what the recommendations are.
00:03:38.479 --> 00:03:48.560
And clinical writing that's done in such a way that is inaccessible or laden with biases and things can do a lot to harm that message.
00:03:48.639 --> 00:04:03.840
And so I really feel passionate about teaching my graduate students how to be really strong clinical writers so that all the important work that we do is communicated in really accessible and accurate ways for the family, for the patient, and for the whole care team.
00:04:03.919 --> 00:04:09.439
So everybody knows what's going on, what we're thinking, what we bring to the care plan.
00:04:10.240 --> 00:04:17.519
And so one of our more recent projects we did was have the students look at a piece of clinical writing.
00:04:17.759 --> 00:04:36.879
And in that piece, we put in lots of instances of problematic or biased language with the intention that our students could see some of those pop out and flag them as problematic and have a chance to rewrite those instances.
00:04:37.040 --> 00:04:47.439
And so we gave them that piece of writing and had them go through that exercise identifying any of those pieces of language and trying to correct it.
00:04:47.600 --> 00:05:03.920
And then we engaged them in a training module where we talked about clinical writing, we talked about anti-oppressive practice, we talked about how we can make our clinical writing a lot stronger, a lot more accurate, and move away from using biased language.
00:05:04.079 --> 00:05:10.480
And then we had them do that same piece of writing again and complete the activity after the training.
00:05:10.639 --> 00:05:22.079
And in our research project, we were able to compare pre-top-training and see the change that our students were able to demonstrate after they received the information, which was really fun.
00:05:22.639 --> 00:05:32.639
The ability to keep bias out of this practice seems like it would be incredibly challenging, um, especially uh what you're talking about, clinical writing.
00:05:32.720 --> 00:05:43.920
So, do you have any specific examples of somebody who was able to remove their bias um from their clinical writing or specific tools they might use?
00:05:44.399 --> 00:05:44.800
Sure.
00:05:44.959 --> 00:05:45.279
Yeah.
00:05:45.439 --> 00:06:03.120
I mean, it's important in speech pathology and in other fields, but especially in speech pathology where we work on communication, and communication is so deeply tied to a person's identity, to their culture, to things that are really important for us to know about and to be able to talk about.
00:06:03.279 --> 00:06:26.160
It's important for us to know different aspects of a person of a person's race or ethnicity or cultural factors that are important because when we're thinking about their communication, there are really real um, you know, pieces that we we want to make sure we don't misidentify as a disorder if it's something that actually is very acceptable given parts of their background.
00:06:26.240 --> 00:06:36.720
And so um, you know, in thinking about different times that we sometimes see biases, we have worked with our students a lot about the gender binary.
00:06:36.800 --> 00:07:13.519
So we have students who sometimes come into our program and will come from the lens of having this idea of a more rigid gender binary, and especially in working with children, you know, boy using the He series pronouns only, and girl use it and really working with our students to say, you know, these are pediatric patients and these are they're coming from families of all different kinds and all different, you know, feelings around this, and teaching students how to talk with families about how they want us to address their children and speak with their children.
00:07:13.680 --> 00:07:42.160
I think that's um really important for our students to know that we don't just see something and immediately make an assumption and let that sort of direct our work and show up in the writing, but just really open things up for conversations with our families and with our patients so that we're using the language that um matters to them and makes sense for how they want us to talk with them, talk about them, and that carries over into how we write about them.
00:07:42.959 --> 00:07:44.079
I that's excellent.
00:07:44.240 --> 00:07:50.560
I was gonna ask you to elaborate on some examples of why it matters so much in speech language, but you you tackled that.
00:07:50.720 --> 00:07:56.720
That's incredibly interesting from the perspective of healthcare in general, but speech language pathology specifically.
00:07:56.959 --> 00:08:01.519
Um, is there a room for AI to assist in this space?
00:08:01.759 --> 00:08:05.759
Or is this um human thought only that should be documented?
00:08:06.399 --> 00:08:10.160
I think we've seen so much good that can come from AI.
00:08:10.319 --> 00:08:21.279
I think, especially for me, I love learning about new technologies that can help me spend more time with the people that I work with and spend more time in my teaching with students.
00:08:21.360 --> 00:08:31.519
And so I'm really excited to learn about how artificial intelligence and different programs can give me back some of that patient care or student teaching time.
00:08:31.680 --> 00:08:38.639
And I think that we're seeing a lot of strengths in some of these programs for assisting us in writing our clinical reports.
00:08:38.799 --> 00:08:45.759
It's an important piece of any speech language pathologist and many other healthcare professionals' day is the documentation.
00:08:45.919 --> 00:08:51.360
It's important for reimbursement, it's important for communication among the care team.
00:08:51.519 --> 00:08:55.120
So it's something that we all need to do in our daily practice.
00:08:55.200 --> 00:09:13.519
And we're seeing that there are many AI programs that are able to take in a session or take in our notes and create something that could be very workable and could save us a lot of time as far as um, you know, a session note or an evaluation report, even.
00:09:13.759 --> 00:09:17.840
So we are seeing a lot of strengths, but then I think it's also important.
00:09:17.919 --> 00:09:48.320
And the reason why I engaged in this particular research study I'm doing now is we also have to be really aware and open about the um the risks as well, because there are um things that we're seeing in the output of generative AI platforms that we need to be aware of and we need to correct, and we need to make sure don't perpetuate um, you know, long-term biases and healthcare inequalities that we know um exist and can easily be perpetuated.
00:09:49.200 --> 00:10:02.399
Do you see these tools as potentially replacing speech language pathologists, or are they more of just a way to assist you in your workload?
00:10:03.279 --> 00:10:37.600
I think one of the things that really makes me happy about speech language pathology right now, and I think for the future, is that it is pretty resilient to um being fully replaced by AI systems because the human connection piece of the communication and the feeding that we do is something that incorporates a human-to-human um response and interaction that cannot be duplicated by these various um algorithms or machines.
00:10:37.759 --> 00:10:49.440
But I do see a lot of artificial intelligence really assisting us and being able to do our jobs better, not only with the clinical documentation, but in our augmentative communication systems.
00:10:49.759 --> 00:11:02.320
The AI that's being put into systems that are helping give voice to the people that we work with are showing such incredible strides in what they're able to help our patients communicate.
00:11:02.480 --> 00:11:17.919
And so, in those ways, knowing how AI works and being able to be the human that uses the AI in responsible and smart ways is helping our patients make so many more improvements so much faster.
00:11:18.000 --> 00:11:22.960
And so I do see it as something that really, really supplements and enhances what we do.
00:11:23.279 --> 00:11:36.000
But I'm happy to say that I don't think it can replace the human component that is so um so important to speech language pathology, communication, the feeding, um, but it can really help.
00:11:36.399 --> 00:11:38.000
That was a really brave question, Kyle.
00:11:38.080 --> 00:11:41.039
I wasn't sure how that response was getting on.
00:11:41.120 --> 00:11:56.639
I'm just the reason I ask is because like you always see people marketing their AI tools as as a a piece that can, you know, one, make your job easier, like this is the solution for everything.
00:11:56.799 --> 00:12:12.960
Uh whereas like I just I wonder if clinicians, uh pathologists like yourself um are necessarily interested in that or just interested in ways to kind of lighten your workload, make it give you more time to focus on the patient.
00:12:13.039 --> 00:12:17.440
Um I never even thought about the augmented the augmentation part for the communication of the patient.
00:12:17.600 --> 00:12:19.039
I never thought about it that way, too.
00:12:19.360 --> 00:12:27.360
It's not just the benefit of uh the communicator, it's also like the learner or the the recipient of the therapy.
00:12:27.440 --> 00:12:28.240
That's fascinating.
00:12:28.559 --> 00:12:34.559
Sorry, there was I don't there was a question there, probably Kyle, but I thought that was cool about the augment um augmented communication.
00:12:35.120 --> 00:12:35.600
For sure.
00:12:35.840 --> 00:12:47.679
Yeah, and we know we've had certain aspects of our jobs that have had artificial intelligence sort of woven in for a long time that we have really um benefited from our patients in using.
00:12:47.759 --> 00:13:09.200
And so there are aspects of the care that we provide and some of the assessment and treatment that we do that have been successfully transferred over to certain AI systems, especially if there's a patient who's working on voice and they're able to see their um their various voice outputs on a computer and be able to self-monitor.
00:13:09.279 --> 00:13:11.759
It's especially powerful for adult clients.
00:13:11.919 --> 00:13:24.879
So there are particular things that we had always been doing face-to-face provided by humans, that we transition certain things over to using AI systems with the help of a human facilitator.
00:13:25.039 --> 00:13:33.360
Um, and we're really open to that because again, it makes it makes us be able to do our jobs better and it makes us um even more effective.
00:13:33.519 --> 00:13:40.240
So I think um there's always always room for the human connection piece in speech pathology.
00:13:40.480 --> 00:13:55.519
We, you know, are still going to be the humans driving behind all of these things, but I think it's important for us to learn about the advances and be able to bring these things in, especially where it makes sense given the diagnosis.
00:13:56.799 --> 00:14:02.480
And you had introduced a little bit about what that current research looks like and how it relates to the conversation we're having.
00:14:02.639 --> 00:14:04.080
Well, the cat's out of the bag.
00:14:04.159 --> 00:14:08.799
You are our inaugural recipient of the CAPSID IVS Clinical Education Seed Grant.
00:14:08.879 --> 00:14:10.399
So, congratulations to you.
00:14:10.639 --> 00:14:19.440
Um for our listeners, the IVS Clinical Education Seed Grant is a financial award that was offered through CAPSID and funded by Intelligent Video Solutions.
00:14:19.600 --> 00:14:26.399
It's designed to support innovations in clinical education, training, and service delivery and communication sciences and disorders.
00:14:26.639 --> 00:14:32.320
Annie, what does this funding now make possible that you may have not been able to do otherwise?
00:14:32.799 --> 00:14:33.279
Yes.
00:14:33.600 --> 00:14:46.639
So we've been so happy to be able to use this funding to create a research project where we have generated AI platforms generate speech language evaluation reports.
00:14:46.799 --> 00:14:51.519
So part of this funding has made it so that we can access various platforms.
00:14:51.600 --> 00:14:57.759
We don't want to be biased and only look at one program and how that is generating speech and language reports.
00:14:57.840 --> 00:14:58.879
We wanted a variety.
00:14:59.120 --> 00:15:03.600
So in that way, we've been able to use this grant for access to those.
00:15:03.759 --> 00:15:18.159
And then we also have used this grant to be able to put together a team of experts, which is really important in this kind of research because we know that when we're talking about biased language, there's so much that comes with that.
00:15:18.399 --> 00:15:22.480
Language is an ever-evolving piece of human life.
00:15:22.720 --> 00:15:33.279
And we, when we're thinking about what we would consider to be biased or problematic language, we need many different minds to think about that.
00:15:33.360 --> 00:15:46.639
And we need people who do research in that particular space who have interests in AI as well, to really be able to comment on what they see to be potentially biased language in those reports.
00:15:46.799 --> 00:16:01.440
And so this C grant has allowed us to put together a team of a variety of individuals, all of whom have extensive experience in biased language, all of whom also have a strong interest in artificial intelligence.
00:16:01.600 --> 00:16:03.679
And they come from all over.
00:16:03.840 --> 00:16:12.879
We connected with colleagues from many different universities, healthcare systems, and we were able to present our ideas for the project.
00:16:13.120 --> 00:16:25.120
And every single one, for the most part, was fully on board to review the reports generated by the various AI platforms and then comment on the bias language that they found.
00:16:25.200 --> 00:16:38.559
And so it's really made it possible so that we could assemble that team and be able to generate these reports in a way that really is representative of how people are working right now.
00:16:38.960 --> 00:16:54.080
Because we also wanted it to be using systems and using those systems in a way that we know our students are using in their work right now with their in their clinical placements and that practicing speech pathologists are using right now to help them in their work.
00:16:54.879 --> 00:17:08.960
If you were referenced by a colleague to help them implement AI documentations tools in their student body or to help them use and implement documentation tools with AI, is there a warning you'd give?
00:17:09.119 --> 00:17:10.319
Is there a caveat?
00:17:10.400 --> 00:17:15.839
What would you advise if someone was looking to explore these tools, maybe in an adjacent industry?
00:17:16.880 --> 00:17:17.279
Sure.
00:17:17.599 --> 00:17:36.799
I would definitely point them to the prompt research, the research that we've seen in terms of how we prompt these systems and letting that organization know that there are risks, there are biases, and we need to be sure we're doing this in really smart and ethical ways.
00:17:37.039 --> 00:17:56.559
And directing them to some of the literature in prompting these systems so that the output they receive is hopefully the best representative of the output because we've been learning a lot about the way that these various platforms are prompted can result in a lot of different kinds of output.
00:17:56.720 --> 00:18:13.759
So that's that's what I would probably mostly caution the person against is not to just fly in with no plan because I there are these risks of problematic language showing up, perpetuation of bias, um, you know, some issues with ethics here.
00:18:13.920 --> 00:18:30.720
I would say tailoring your prompts and making sure that the people in your organization are using really well-written prompts so that the output is more of what they're looking for, and so that they aren't having to prompt and reprompt and edit.
00:18:30.799 --> 00:18:41.039
Because again, we know that with that comes more and more use of these systems and that has environmental impacts, you know, the amount of times that we're doing all of these things.
00:18:41.119 --> 00:18:48.960
So if we can teach people the importance of prompting in very smart and efficient ways, that's another way that we can support better outputs.
00:18:49.359 --> 00:18:50.960
I think that's very interesting.
00:18:51.039 --> 00:19:00.480
Um, the fact that what you put in will ultimately get you more quality on the other end and how you're wording things.
00:19:00.720 --> 00:19:13.759
Um, obviously, adoption has to be there, and there has to be a belief that this is going to be a tool that helps save time or or money or has um a good investment uh on the output.
00:19:14.079 --> 00:19:16.640
Is there adoption concerns?
00:19:16.799 --> 00:19:18.640
And do you see peers or colleagues?
00:19:18.799 --> 00:19:21.279
You are working with a lot of people excited about AI.
00:19:21.519 --> 00:19:24.000
What does the other side of the coin look like?
00:19:24.240 --> 00:19:25.119
The resistors.
00:19:25.839 --> 00:19:26.880
Yes, for sure.
00:19:27.119 --> 00:19:43.039
So when we started as a department talking about how we were going to bring AI into our teaching and think about it for teaching our students, we know that practicing clinicians are using this, so we know our students are going to be asked to use this in their clinical placements.
00:19:43.359 --> 00:19:51.599
We expected there to be pushback, and and there were some good discussions among different members and talking about it.
00:19:51.759 --> 00:19:57.039
But what we didn't expect was how much pushback we would receive from our students.
00:19:57.279 --> 00:20:18.160
And so that's been a really interesting, um interesting piece that we didn't fully prepare for is that a lot of our students there in their early 20s, and they have a lot of very legitimate reasons for why they do not want to engage with AI in their practice as speech language pathologists.
00:20:18.240 --> 00:20:33.440
And so we've actually had to do a lot of thinking about that and talking with our students about the reality that this is something that they will very likely be tasked to use in their clinical placements or in jobs.
00:20:33.599 --> 00:20:43.519
And so, how can we teach them to use it in ways that feel comfortable for them, given that there's some very strong resistance there for very legitimate reasons?
00:20:44.000 --> 00:20:51.039
Honestly, that's that's something we've seen in past conversations that the demographic is most um resistance.
00:20:51.279 --> 00:20:54.319
You think it'd be maybe the alternative, but no, sorry, Kyle.
00:20:54.720 --> 00:21:00.079
No, I was just yeah, that was I'm surprised that it came from students, not like tenured faculty.
00:21:00.319 --> 00:21:08.400
Um in a this is completely irrelevant, but I saw some reel the other Instagram reel the other day where it was some commencement speaker.
00:21:08.799 --> 00:21:11.920
You're not even getting your news from TikTok against it.
00:21:12.960 --> 00:21:15.039
But but basically the same thing, whatever.
00:21:15.119 --> 00:21:22.319
Uh, but the commencement speakers started like talking about AI and the future of it, and the whole all the student graduates boo, yeah, you saw that.
00:21:22.559 --> 00:21:25.839
Yeah, it was a couple months ago I saw it probably just recently for you on Reels.
00:21:26.640 --> 00:21:31.279
But that was like my first exposure to like, oh, maybe students like aren't fully on board with this.
00:21:31.759 --> 00:21:32.079
Yeah.
00:21:32.240 --> 00:21:46.400
And you know, we make these, we make these assumptions and we think, oh, we know exactly how how to advance our teaching, and we need to bring in the best and the newest and the and the most sort of next next step sorts of things.
00:21:46.480 --> 00:22:04.319
Um, but I think it's really important for anybody in any line of work that we do to to really kind of gauge with our audience, engage with our audience and figure out okay, actually, let's let's take a needs survey here or Let's actually figure out what the students are feeling about it.
00:22:04.480 --> 00:22:10.799
So that again, we need, we have the responsibility to teach them about artificial intelligence.
00:22:10.960 --> 00:22:16.240
We must teach them about how it's being used in speech language pathology clinical practice.
00:22:16.400 --> 00:22:20.000
We would be doing our students a disservice to just say, oh, forget it.
00:22:20.079 --> 00:22:20.799
Then we won't teach it.
00:22:20.880 --> 00:22:23.039
If you, if you all don't want to use it, we won't teach it.
00:22:23.119 --> 00:22:25.920
That would be a real disservice to them.
00:22:26.079 --> 00:22:32.960
So we do need to be honest with them about that, be really transparent with the fact that this is something.
00:22:33.039 --> 00:22:49.359
It's going to show up, and your ability to use it in ethical and responsible ways is going to make it so that it provides the best with the least amount of negative impact or resource use as possible.
00:22:50.079 --> 00:23:02.400
I'm curious, all of the focus that's going on this research and your work in this space, how does that impact your video observation and recording in these clinical education environments?
00:23:02.559 --> 00:23:03.440
Do they go hand in hand?
00:23:03.599 --> 00:23:04.960
Are they separate?
00:23:05.039 --> 00:23:12.319
How are you using them together or do you keep some maintain uh different spaces for these two topics?
00:23:12.880 --> 00:23:16.079
We keep it all folded in together.
00:23:16.240 --> 00:23:26.720
So all of our in our on-campus clinic, all of our sessions are video recorded and we are watching our students via that live stream video in the moment.
00:23:26.880 --> 00:23:43.599
And for that, we are able to give the students the space in that treatment room with the patient and with the family to take some risks and to feel like they can try out things without their faculty member being physically in that space.
00:23:43.839 --> 00:23:57.839
And so in that way, we're able to let our students have some of that autonomy and learn how to be a practicing clinician without us necessarily jumping in at every at every moment.
00:23:58.000 --> 00:24:14.400
And in terms of our use of AI, we are able to fully fold in the fact that they're going to be writing a session note about what's happening in that recorded session for every single time they see a patient in our clinic.
00:24:14.640 --> 00:24:30.480
And so we're able to work with them and replay that video recording of the session and pull things out and, you know, work with them to use different ways of writing up that session note in a way that is accurate to what happened.
00:24:30.559 --> 00:24:38.400
And we're always talking to our students about using objective language and removing any bias language that we find.
00:24:38.559 --> 00:24:52.319
So when we have AI supports in helping us with that session note output, we want our students to know there is a strong potential that there will be some bias or some subjective language here.
00:24:52.480 --> 00:25:11.039
So I need you to be able to identify it and I need you to be able to rewrite those moments and come to me with any questions that you might have about that because it's really difficult, especially as a new clinician who is just learning how to do the job to then learn how to write about it is so new.
00:25:11.119 --> 00:25:18.400
And it's really tricky when you haven't put those various things into words many, many times already.
00:25:18.480 --> 00:25:21.599
It does require some help, and that's why we're there for them.
00:25:22.240 --> 00:25:25.519
So Noli's technology is reflection, is a huge theme.
00:25:25.680 --> 00:25:35.519
Reflecting on the auto prompts, reflecting on what you're getting as the quality output from AI, reflection on the physicality and what's happening in the actual spaces with AV capture.
00:25:35.680 --> 00:25:44.559
What is a positive outcome that has come from using AV, whether it's in tandem with your clinical documentation with AI or if it's just AV?
00:25:45.599 --> 00:26:01.200
I think one of our most favorite things that we do is at the end of every semester, our students pick a video clip from their sessions that they've had in the on-campus clinic, and they share it with everybody.
00:26:01.359 --> 00:26:08.400
So we have a very large cohort of students, and each student has two patients in our on-campus clinic their first year.
00:26:08.559 --> 00:26:11.839
And so at the end of every semester, we all get together.
00:26:11.920 --> 00:26:20.960
That large cohort is able to put up a selected clip by them, and they're able to explain to their classmates here's the patient I'm going to show you.
00:26:21.119 --> 00:26:30.960
Here's what we're going to do in the session, and what I was really proud of, or what I was really challenged with, or what I really would love your, would love your input on.
00:26:31.039 --> 00:26:40.640
And then they show a clip from those recorded sessions, and then every student in the cohort is able to see a patient that they're not familiar with, which is really important.
00:26:40.799 --> 00:26:43.519
They get more exposure and variety.
00:26:43.839 --> 00:26:47.519
And they're able to think about, okay, what would I have done in this situation?
00:26:47.759 --> 00:26:54.319
Or I really appreciated how my peer did something in this way.
00:26:54.480 --> 00:26:59.440
I'm going to do that with my client that I have or my patient that I've got.
00:26:59.680 --> 00:27:16.319
So that's probably the most fun things that we can use our video recordings for is really celebrating each other's successes and exposing the whole cohort to all the various different patients that come through our on-campus pro vono clinic.
00:27:16.559 --> 00:27:19.920
It's really, really fun to do that at the end of every semester.
00:27:20.240 --> 00:27:28.160
Annie, we have a lot of listeners that might be interested in contacting you to learn more about your work or just to follow you and your research.
00:27:28.640 --> 00:27:32.640
How is the best, what is the best way for those listeners to get a hold of you?
00:27:33.200 --> 00:27:33.519
Sure.
00:27:33.680 --> 00:27:36.480
I'm more than happy to connect via email.
00:27:36.640 --> 00:27:42.559
My email is akholland at mghihp.edu.
00:27:43.279 --> 00:27:53.200
And you can find our more recently published report on the biased language in the pre and post-training that I talked about earlier.
00:27:53.359 --> 00:27:59.279
That's in the Teaching and Learning and Communication Sciences and Disorders journal, which is open access.
00:27:59.680 --> 00:28:01.200
So you can find that there.
00:28:01.359 --> 00:28:03.359
I'm also on LinkedIn, Annie K.
00:28:03.519 --> 00:28:04.000
Holland.
00:28:04.319 --> 00:28:07.279
So I'm happy to connect with anyone anytime.
00:28:07.680 --> 00:28:08.640
Thank you, Annie.
00:28:08.799 --> 00:28:09.759
It was very enlightening.
00:28:09.839 --> 00:28:15.119
We learned a lot today, and it's it's good practice and information for every individual human being to hear.
00:28:15.680 --> 00:28:21.680
It's really nice to know that people are uh doing some research on these AI tools.
00:28:21.759 --> 00:28:33.279
And um, and so maybe we have some like uh facts grounded in data um as to how they should be used, and we aren't just kind of like going with it and hoping that it all works out.
00:28:33.440 --> 00:28:36.000
So we really appreciate you doing this work.
00:28:36.480 --> 00:28:38.240
Yes, thank you so much.
00:28:38.480 --> 00:28:47.039
And if you or if someone you know is interested in learning more about the IVS Clinical Education Seed Grant, visit capsa.org forward slash scholarships.
00:28:47.200 --> 00:28:48.240
We'll see you next time.
00:28:48.400 --> 00:28:49.279
Thank you.
00:28:50.559 --> 00:28:58.079
OnRecord is a podcast presented by Intelligent Video Solutions, hosted by Mike Anzolone and Carmen Hawk, and produced by Kyle Shellstead.
00:28:58.319 --> 00:29:05.440
Find us online at ipivs.com slash on record and on social media at Intelligent Video Solutions.
