WEBVTT
00:00:01.760 --> 00:00:08.720
Welcome to the Advancing Surgical Care Podcast, brought to you by ASCA, the Ambulatory Surgery Setter Association.
00:00:08.960 --> 00:00:19.039
ASCA represents the interests of outpatient surgery setters of every specialty and provides advocacy and resources to assist them in providing safe, high-quality, cost-effective patient care.
00:00:19.199 --> 00:00:26.399
As with all of ASCA's communications, please check to make sure you are listening to or viewing our most up-to-date podcasts and announcements.
00:00:26.800 --> 00:00:28.800
Hello, and welcome to the podcast.
00:00:28.879 --> 00:00:35.840
I'm Mandy Hawken, CEO of Three Hawks Advisors, past president of ASCA, and the host of this recording.
00:00:36.079 --> 00:00:50.399
Earlier this year, the ASCA Board of Directors created a working group to study and explore how advances in artificial intelligence, or AI, could improve the efficiencies of both the administrative and clinical operations of surgery centers today.
00:00:50.640 --> 00:00:56.399
This is a subject that fascinates me, and I immediately said yes when I was asked to chair the new working group.
00:00:56.640 --> 00:01:17.599
I was particularly excited when I learned the group would include several very distinguished ASCA colleagues, including our new president Todrier, past president Mark Weiner, and ASCA board members Dare Meeks, Jessica Rodriguez, Alfonso Delgranado, and Lisa Cooper, along with ASCA CEO Bill Prentiss and his outstanding team.
00:01:17.920 --> 00:01:28.319
As I mentioned a minute ago, we intend to explore both administrative and clinical parts of AI that I imagine will require a number of educational tools for our members.
00:01:28.560 --> 00:01:35.760
For this introductory podcast, I want to focus on some of the ways that AI can provide administrator support at their surgery centers.
00:01:36.000 --> 00:01:48.159
To start this conversation, I've asked my really good friend and colleague, Becky Ziegler Otis, the assistant executive director of the ASC Quality Collaboration, to join me on the podcast.
00:01:48.480 --> 00:01:58.640
Back in May, Becky was part of an outstanding panel on AI at ASCA's annual meeting, and she too is enthusiastic and an early adopter of several AI tools.
00:01:58.799 --> 00:02:05.519
But before I bring Becky into the conversation, I want to say a few more words about the startup ASCA's AI Working Group.
00:02:05.680 --> 00:02:12.080
In our inaugural meeting, the consensus among the group was that AI indeed holds great promise for ASCs.
00:02:12.479 --> 00:02:16.560
But many of us are also discovering that there is a need to proceed with caution.
00:02:16.719 --> 00:02:22.879
We started by aligning on an understanding of the foundational elements of AI and healthcare for our working group.
00:02:22.960 --> 00:02:33.919
And to confirm we are on the right track, we asked ASCA members to complete a short survey about their AI usage and where they thought ASCA could be most helpful in terms of education and support.
00:02:34.080 --> 00:02:41.599
And convincingly, the respondents in the survey said AI, compliance, and privacy guidance were a top priority for them.
00:02:41.840 --> 00:02:53.599
From there, we invited a couple of leading healthcare attorneys to join us for a briefing, and work is already underway to develop some useful tools for ASCA members to remain safe and compliant when adopting AI.
00:02:53.840 --> 00:02:56.159
There will be more about this in the near future.
00:02:56.400 --> 00:02:58.719
So let's bring Becky into the conversation.
00:02:58.960 --> 00:03:01.039
Becky, welcome to the podcast.
00:03:01.360 --> 00:03:04.240
Thank you, Mandy, and thank you so much for having me today.
00:03:04.319 --> 00:03:07.680
I really appreciate being able to be a part of the podcast with you today.
00:03:07.919 --> 00:03:09.039
We're so glad you're here.
00:03:09.120 --> 00:03:18.400
And I know our members are thankful that you are here and for getting us kicked off at ASCA this year at the AI panel that you were a part of.
00:03:18.639 --> 00:03:27.520
I notice AI can be a little intimidating for some, and for others, it almost gives them a sense that they're going to have to learn something complicated and new.
00:03:27.599 --> 00:03:29.039
And who has time for that?
00:03:29.360 --> 00:03:37.759
Would you talk about a couple of practical administrative applications that are available to surgery centers today that don't have steep learning curves?
00:03:38.080 --> 00:03:39.280
Mandy, absolutely.
00:03:39.360 --> 00:03:40.159
I would love to.
00:03:40.319 --> 00:03:46.240
And I want to start by saying that the practical starting points aren't really all that flashy.
00:03:46.479 --> 00:03:49.919
They're really things that you're probably already halfway using.
00:03:50.000 --> 00:03:52.000
So I'm just going to point out a few of those.
00:03:52.240 --> 00:03:59.120
I suspect many centers are already using Microsoft Office Suite, which has Copilot that's built right into it.
00:03:59.280 --> 00:04:03.919
And you might already be leaning on Copilot to help you draft some emails.
00:04:04.080 --> 00:04:08.319
Because as you write your email at the end of it, you'll see a little pencil that shows up.
00:04:08.400 --> 00:04:16.319
And if you click on that pencil, it can give you some options to help you rewrite the content of the email, maybe even allowing you to change the tone.
00:04:16.480 --> 00:04:21.600
You can say you want it to be more casual or more professional, and it helps rewrite it for you.
00:04:21.759 --> 00:04:27.439
And I think what's important to know is that you have the ability to replace it or ignore it.
00:04:27.680 --> 00:04:30.959
And that's the piece that I think many people really get nervous about.
00:04:31.120 --> 00:04:38.000
They think that AI is just going to do the answer for them and they won't be able to help out with the response, but that's not how it works.
00:04:38.240 --> 00:04:48.879
Once Copilot does the draft, you absolutely have the ability to look it over and make the decision, have the human in the loop to make sure that you decide what you wanted to say.
00:04:49.120 --> 00:04:52.079
And speaking of Copilot, I have another practical way.
00:04:52.240 --> 00:04:55.360
It can also help you outside of drafting emails.
00:04:55.519 --> 00:04:57.199
Maybe it can help organize your day.
00:04:57.360 --> 00:05:06.959
For example, let's say you've been off for a bit, maybe it's over the weekend, and you come in on Monday morning and you want to find out what's been happening over the weekend in all of your emails.
00:05:07.040 --> 00:05:07.839
What's your priorities?
00:05:07.920 --> 00:05:09.120
How do you get your day started?
00:05:09.279 --> 00:05:16.560
So a simple thing you can do is just go into Copilot and say, let me know what I've missed over the past 48, 72 hours.
00:05:16.720 --> 00:05:23.120
And it will scan those emails, come up with what's the high priorities, what are the things that you really need to focus on.
00:05:23.279 --> 00:05:26.160
So it helps you with your efficiency.
00:05:26.399 --> 00:05:32.480
And if you haven't tried this, I have another pearl, and that has to do with agenda or minute preparation.
00:05:32.720 --> 00:05:35.600
And if you haven't tried it, I would say give it a try.
00:05:35.920 --> 00:05:42.879
As we all know, we have to be responsible for committee meeting minutes, whether it be the governing body, quality meetings, staff meetings.
00:05:43.040 --> 00:05:46.000
We're always putting together agendas and then we have to do the minutes.
00:05:46.160 --> 00:05:56.000
So the next time we're preparing for an agenda, for example, for your quality meeting, for your governing body, just say to Copilot, put my agenda together for me.
00:05:56.240 --> 00:06:03.600
I think you're going to be pleasantly surprised because you're going to come up with an agenda that is probably about 80% complete.
00:06:03.759 --> 00:06:16.639
And it's really going to help you with your efficiency because you're going to be able to go through it, tweak what it is that you want to change, have that human in the loop, making sure that you make the decision on what's in that agenda.
00:06:16.879 --> 00:06:20.079
But it's going to get you there so much faster.
00:06:20.319 --> 00:06:22.560
And I've really focused a lot on Copilot.
00:06:22.639 --> 00:06:29.600
And I just want to bring up that you can also use some other AI tools such as chat, GPT, or Claude that are available.
00:06:29.839 --> 00:06:32.079
And let's say you have to draft a document.
00:06:32.160 --> 00:06:33.680
And this is very helpful.
00:06:33.839 --> 00:06:45.120
Whether that document be a policy, whether it be a job description, a competency checklist, an orientation checklist, for example, think about using that AI tool to assist you.
00:06:45.279 --> 00:06:50.720
So, Mandy, all you really have to do to start the chat is ask the tool to draft a document.
00:06:50.879 --> 00:06:59.360
For example, I want to create a job description for a surgical tech for an ASC environment, and it's an ortho-specific center.
00:06:59.519 --> 00:07:05.279
The more detail you can give in the prompts, the better on how it's going to come up with the first draft.
00:07:05.519 --> 00:07:10.240
But I even learned recently that sometimes I'm not sure what I want to make for my prompts.
00:07:10.399 --> 00:07:16.079
So I will just say, draft me a job description and provide me with some clarifying questions.
00:07:16.319 --> 00:07:20.000
And then the tool will give me the clarifying questions to answer.
00:07:20.240 --> 00:07:25.360
I can come up with a draft, once again, is about 80 to 90% there.
00:07:25.600 --> 00:07:31.439
I can tweak that document, use my human in the loop, my decision making on how I want it to look.
00:07:31.519 --> 00:07:36.160
And it works much faster than if I have to just sit down with a blank piece of paper.
00:07:36.399 --> 00:07:40.720
So I hope those are some really practical tips for an administrator, Mandy.
00:07:40.959 --> 00:07:42.800
I think those are all great tips.
00:07:42.879 --> 00:07:45.279
And I love that you called out the human in the loop.
00:07:45.439 --> 00:07:53.759
The human in the loop gives you that intelligence and judgment and oversight while working with AI, and it also improves accuracy.
00:07:54.000 --> 00:07:56.800
It makes me think of self-driving car like Waymo.
00:07:57.040 --> 00:07:59.519
It just didn't just start driving on its own.
00:07:59.680 --> 00:08:09.040
A human was in the loop providing data that was used to train and guide the models, teaching it what it was supposed to do so it would have that safe driving behavior.
00:08:09.279 --> 00:08:16.480
I hope some of our members are using or beginning to try these AI tools you talked about in their everyday work or even home life.
00:08:16.800 --> 00:08:18.959
They absolutely can save you time, Andy.
00:08:19.040 --> 00:08:23.680
And as a leader, one of the things we can certainly benefit from is saving time.
00:08:24.000 --> 00:08:24.800
Saving time.
00:08:24.879 --> 00:08:26.000
That's what we all need.
00:08:26.240 --> 00:08:30.720
Becky, we have to take a short break to hear a few words from our podcast sponsor.
00:08:30.879 --> 00:08:32.960
Stand by and we'll be right back.
00:08:33.200 --> 00:08:42.080
This recording is being brought to you by Surgery Partners, a leading owner and operator of ambulatory surgery centers with more than 200 locations nationwide.
00:08:42.320 --> 00:08:54.960
Surgery Partners collaborates with ASC leaders and physicians to acquire, develop, and optimize surgery center operations and helps ASCs scale efficiently, strengthen performance, and drive long-term growth?
00:08:55.200 --> 00:08:58.159
To learn more, visit surgerypartners.com.
00:08:59.440 --> 00:09:15.039
Becky, let's talk about business associate agreements or BAAs, which should be well known to ASC administrators, but take on an added level of sensitivity when using AI tools that are coming in contact with PHI.
00:09:15.279 --> 00:09:23.440
Do administrators need to know about BAAs, even with some of these free AI tools that don't require an enterprise agreement?
00:09:23.759 --> 00:09:25.039
That's a great question.
00:09:25.200 --> 00:09:37.200
And here's the one thing I'd really want the listeners to walk away with, and that is paying for an AI tool does not mean automatically that you have a business associate agreement in place.
00:09:37.360 --> 00:09:39.360
They're really two different things.
00:09:39.519 --> 00:09:42.159
And you should never assume that you have one.
00:09:42.320 --> 00:09:52.320
You should always verify that it's in place before putting in any proprietary information or patient personal health information in any AI tool or chat.
00:09:52.559 --> 00:09:58.320
Absolutely, someone has to confirm that there's a BAA in place for that specific product.
00:09:58.559 --> 00:10:00.720
And it can't be just that we have an account.
00:10:00.799 --> 00:10:03.440
It's for what account and what plan.
00:10:03.600 --> 00:10:05.440
It has to be very specific.
00:10:05.759 --> 00:10:16.399
Certainly when it comes to a free Chat GPT or personal Claude or even your own personal co-pilot account, none of those come with a business associate agreement.
00:10:16.559 --> 00:10:20.399
Even a paid personal subscription does not count.
00:10:20.639 --> 00:10:26.240
It has to be a business or an enterprise account, specifically one that is configured for HIPAA.
00:10:26.720 --> 00:10:32.159
And where I can really see this going wrong for an organization isn't really one of those dramatic breaches.
00:10:32.480 --> 00:10:41.360
To me, it's someone on your team is just trying to be efficient and ultimately ends up doing something that gives your organization some exposure.
00:10:41.519 --> 00:10:45.360
For example, let's say they're drafting a response to a patient complaint.
00:10:45.519 --> 00:11:03.200
Maybe a family member has complained about their patient care experience, and they just want to paste that information very quickly into Chat GPT to help reframe what they're saying, but they accidentally forget to remove the patient's name or some other portion of the demographics.
00:11:03.360 --> 00:11:09.840
That information is now part of the large language model and in that overall chat, and you have some exposure.
00:11:09.919 --> 00:11:19.679
So you can see it's really just a simple mistake, which is absolutely why you need to have some guardrails in place, such as policy depicting AI governments.
00:11:19.919 --> 00:11:28.480
So my most practical advice is to make sure you always strip out any identifying or proprietary information before entering into a chat.
00:11:28.639 --> 00:11:34.080
Because once it's in the chat, it's part of that large language model and it's out there.
00:11:34.320 --> 00:11:37.759
Becky, I really like that you said confirm it's in place.
00:11:38.000 --> 00:11:41.440
That's such an important piece for all of our members to understand.
00:11:41.600 --> 00:11:49.840
When you're getting a new contract, as well as with vendors that have now added automation, review and make sure it is there.
00:11:50.159 --> 00:11:58.559
Hopefully, as we continue to build on our AI working group, we'll be able to provide some more education and resources for our members to get these guardrails in place.
00:11:58.799 --> 00:12:08.720
Well, Mandy, I think a lot of ASCs are struggling with knowing what is the right thing to do and not even knowing where to begin from the governance perspective.
00:12:08.879 --> 00:12:12.960
So the effort on behalf of this work group will certainly be well appreciated.
00:12:13.120 --> 00:12:14.159
Thank you so much.
00:12:14.399 --> 00:12:19.600
We've talked about document prep and protecting against HIPAA violations.
00:12:19.840 --> 00:12:29.360
I know that you've also had some personal experience with AI agentics that have become lifelike, intuitive, and engaging in conversations with patients.
00:12:29.600 --> 00:12:34.080
Can you share some personal interaction that you've had with your AI agent?
00:12:34.399 --> 00:12:36.960
This one was very, very exciting for me.
00:12:37.120 --> 00:12:49.840
I had a trained AI agent that was being used to complete the pre-operative phone call that, as we know, is typically done by the pre-op nurse prior to a procedure or a phone call in the ASC environment.
00:12:50.000 --> 00:12:53.519
Or in some cases, we go out to a portal and we complete that information.
00:12:53.759 --> 00:12:57.840
Well, I decided to give it a try using the trained AI agent.
00:12:58.000 --> 00:13:03.919
And when I received the call, it sounded very lifelike and it was very engaging.
00:13:04.080 --> 00:13:06.240
And in my case, it was a woman's voice.
00:13:06.399 --> 00:13:10.399
She asked me all of the required questions of your typical pre-op phone call.
00:13:10.559 --> 00:13:15.919
And my goal really was to try to stump this woman, to stump the AI agent.
00:13:16.080 --> 00:13:19.600
I wanted to see what would happen if I asked a lot of questions.
00:13:19.840 --> 00:13:21.840
So I really had a lot of fun with this.
00:13:22.000 --> 00:13:25.679
For example, I told her, geez, I know I'm on some medications.
00:13:25.840 --> 00:13:30.000
One is for my blood pressure, and I'm not sure if it starts with a P or a T.
00:13:30.240 --> 00:13:36.320
And the agent responded back with me on some very clarifying questions, and I was not able to answer them again.
00:13:36.399 --> 00:13:38.320
And she said, You know what, Becky?
00:13:38.480 --> 00:13:39.440
And she used my name.
00:13:39.600 --> 00:13:41.279
She said, You know what, Becky?
00:13:41.440 --> 00:13:45.039
I'm going to make a note of this and I'll have the nurse call you back and discuss it.
00:13:45.200 --> 00:13:47.919
And I thought, oh, that's a great answer.
00:13:48.159 --> 00:13:51.039
And she went out again and I kept trying to stump her in other ways.
00:13:51.120 --> 00:13:56.879
And I also remember I asked her about MH when she asked about having a history of that.
00:13:57.120 --> 00:13:58.720
I said, I don't understand what you mean.
00:13:58.879 --> 00:14:00.000
Can you help me understand?
00:14:00.159 --> 00:14:03.840
And she responded very appropriately with a definition.
00:14:04.080 --> 00:14:11.360
And when I said, I still don't understand what you're talking about, at that point, she said, You know, Becky, I think we're going to have the nurse call you back and help answer that.
00:14:11.600 --> 00:14:12.960
And we finished the call.
00:14:13.039 --> 00:14:22.000
And right towards the end of the call, where I thought the agent really went above and beyond, is started to frame for me all of the questions that I had asked.
00:14:22.159 --> 00:14:26.240
And she said, These are the questions that the nurse is going to give you a call back on.
00:14:26.399 --> 00:14:31.200
And she hit every one of them, even the ones that I had forgotten.
00:14:31.440 --> 00:14:39.919
So I felt very comfortable that the person that I was talking to, which was a trained agent, was a caring person, felt like it was a real person.
00:14:40.320 --> 00:14:46.799
And this isn't an actual ASC practice management software that we have available to all of us if we want to use it.
00:14:47.039 --> 00:14:51.039
Wow, this can really help with efficiency and effective use of staff time.
00:14:51.200 --> 00:14:54.399
I thoroughly enjoyed learning about this and being a part of the call.
00:14:54.480 --> 00:14:57.360
And to frame it all up, it was exciting.
00:14:57.679 --> 00:14:59.840
I also have had the same experience.
00:15:00.000 --> 00:15:12.320
I find it fascinating, the quality of their responses, how they change their rhythm and tone, almost how graceful they are when they're addressing you and not interrupting.
00:15:12.480 --> 00:15:19.440
So it's nice to see that you can really train that agent to do exactly what you want it to do, keeping that human in the loop.
00:15:19.679 --> 00:15:34.000
When we kicked off the working group, we had an executive speak to us from Google Health, and they had a lot of experience in this with pre-calling patients and how they've trained their agent to have those conversations, and it's been really successful for them.
00:15:34.159 --> 00:15:40.960
So it's nice to hear that people outside of the ASC space are using this and that's extra support for us to do the same.
00:15:41.200 --> 00:15:43.120
Those are some great insights, Mandy.
00:15:43.200 --> 00:15:53.679
And actually, when you were talking about it, it's amazing how when you listen to those trained agents, you have a preconceived notion that it's going to sound like a robot, that there's going to be uncomfortable gaps.
00:15:53.919 --> 00:15:56.080
That's not how it really works out.
00:15:56.399 --> 00:15:56.960
It isn't.
00:15:57.120 --> 00:15:58.559
It is not how it works out.
00:15:58.879 --> 00:16:07.360
Well, Becky, I would be remiss if I had you on the podcast without giving you a chance to tell us to what's the latest at the ASC Quality Collaboration.
00:16:07.600 --> 00:16:09.840
I know you have a new assessment tool.
00:16:10.159 --> 00:16:12.320
What are your responses looking like?
00:16:12.399 --> 00:16:15.519
And how are you thinking about AI and ASC Quality?
00:16:15.759 --> 00:16:16.720
Well, thank you, Mandy.
00:16:16.879 --> 00:16:24.639
The ASC Quality Collaboration has recently just finished their annual ASC Safety and Quality Assessment, which is complimentary.
00:16:24.879 --> 00:16:26.639
It's free, complimentary.
00:16:26.799 --> 00:16:30.559
So any ASC across the country can help participate in this.
00:16:30.720 --> 00:16:38.240
And what you're doing is you're benchmarking your center against national performance on quality and safety best practices.
00:16:38.480 --> 00:16:49.919
And this year we had it open for a six-week period of time and we had over 700 submissions, which compares very positively to a little under the 400 submissions we had the year prior.
00:16:50.080 --> 00:16:51.440
So that's fantastic.
00:16:51.600 --> 00:16:55.519
And I want to thank all of you who participated in that assessment.
00:16:55.679 --> 00:16:59.600
And if you happen to miss it this year, it is available next year.
00:16:59.759 --> 00:17:07.920
And if you're concerned you might miss the window for opportunity next year, go out to the ASCQC website, which is ascquality.org.
00:17:08.000 --> 00:17:16.160
And on that website, you will be able to put in your contact information so you can receive an email next year and be reminded of that.
00:17:16.480 --> 00:17:29.839
Also, with respect to the ASC Quality Collaboration, we're excited about our annual membership only quality conference that's going to occur later this month, which is really a great way for our membership to collaborate and network with one another.
00:17:29.920 --> 00:17:36.640
And we'll actually have the opportunity to connect with some key regulatory agencies during the conference like CMS and CDC.
00:17:36.799 --> 00:17:41.119
And then, Mandy, to answer the second part of your question, how can AI work with quality?
00:17:41.440 --> 00:17:44.319
Think AI can really shine in this area.
00:17:44.480 --> 00:17:47.440
And I think about it in a number of perspectives.
00:17:47.680 --> 00:17:50.319
The first is pattern spotting.
00:17:50.559 --> 00:18:00.960
Instead of just staring at a spreadsheet of incident reports or infection surveillance logs, you can really ask AI to help you look for patterns that might be surfacing.
00:18:01.119 --> 00:18:04.079
Patterns related to time of days, staff combination.
00:18:04.240 --> 00:18:07.759
But be careful what you put out there because you have to make sure it's protected.
00:18:08.000 --> 00:18:11.200
It can also help you with documenting your PI project.
00:18:11.359 --> 00:18:17.759
It seems like whenever I'm ready to start documenting a project, I have a blank page, I can't think about where do I begin.
00:18:18.000 --> 00:18:21.519
So use a chat to just put your notes in and get you started.
00:18:21.839 --> 00:18:25.519
I recently had an experience where I was doing this myself.
00:18:25.680 --> 00:18:30.480
I was struggling with something that I was trying to write up, and it was a great PI project.
00:18:30.640 --> 00:18:35.039
So I put some notes in on what I had measured, what were some of my findings.
00:18:35.279 --> 00:18:39.759
I was surprised to see how it came out with a great first draft for me.
00:18:39.920 --> 00:18:43.680
Once again, 80% there that I just had to turn into a study.
00:18:43.920 --> 00:18:46.160
You can also use it for a quick literature scan.
00:18:46.319 --> 00:18:52.000
Let's say you have a committee that says to you, I want you to check into unplanned transfers or same-day cancellations.
00:18:52.240 --> 00:19:01.839
It can help pull together what research has been done on this and list you with references and resources that you can tap into to make sure that it's valid information.
00:19:02.480 --> 00:19:10.079
So I would emphasize again that it's going to get you 80 to 90% of the way there, but it always requires that human in the loop.
00:19:10.160 --> 00:19:11.680
And that human is you.
00:19:11.920 --> 00:19:16.000
So you have to make the first draft into the final draft, whatever you're working in.
00:19:16.240 --> 00:19:20.319
So I'm going to conclude this all, Mandy, with saying, why not try out an AI tool?
00:19:20.640 --> 00:19:24.160
Dip your toes in and see what it can do for you.
00:19:24.319 --> 00:19:30.400
Because I think you'll be pleasantly surprised to see how it can help you with your efficiency.
00:19:30.640 --> 00:19:31.920
I agree with you 100%.
00:19:32.400 --> 00:19:38.480
I challenge our members, give yourself five to 10 minutes a day and try something new.
00:19:38.720 --> 00:19:41.759
And Becky, I want to thank you for spending time with me today.
00:19:41.920 --> 00:19:43.599
I love catching up with you.
00:19:43.839 --> 00:19:47.200
You are a true legend in our industry, and we're lucky to have you.
00:19:47.440 --> 00:19:50.559
Thank you for sharing all of your knowledge and insights on AI.
00:19:50.720 --> 00:19:52.880
And we will continue to pick your brain.
00:19:53.039 --> 00:19:58.559
And thank you for all the hard work that you and Nina Gwenes do at the ASC Quality Collaboration.
00:19:58.799 --> 00:20:02.000
It's such important work in support of the ASD community.
00:20:02.319 --> 00:20:04.559
Mandy, I thank you for having me today.
00:20:04.720 --> 00:20:10.079
And I so appreciate working with you and the fact that you reached out to me for my insights.
00:20:10.160 --> 00:20:22.799
And I also want to thank Ask and the work group for all the work that is being done on these AI initiatives to provide the guidance and the guardrails that we as the ASC industry absolutely need going forward.
00:20:22.960 --> 00:20:23.680
Thank you.
00:20:24.000 --> 00:20:36.880
Now, before concluding, I would like to acknowledge the support of our podcast sponsor, Surgery Partners, a leading operator of surgical facilities and ancillary services with more than 200 locations in 33 states.
00:20:37.119 --> 00:20:44.960
Surgery Partners offers multiple types of healthcare services that deliver high quality health care in a convenient and cost effective manner.
00:20:45.200 --> 00:20:48.319
To learn more, visit surgerypartners dot com.
00:00:01.760 --> 00:00:08.720
Welcome to the Advancing Surgical Care Podcast, brought to you by ASCA, the Ambulatory Surgery Setter Association.
00:00:08.960 --> 00:00:19.039
ASCA represents the interests of outpatient surgery setters of every specialty and provides advocacy and resources to assist them in providing safe, high-quality, cost-effective patient care.
00:00:19.199 --> 00:00:26.399
As with all of ASCA's communications, please check to make sure you are listening to or viewing our most up-to-date podcasts and announcements.
00:00:26.800 --> 00:00:28.800
Hello, and welcome to the podcast.
00:00:28.879 --> 00:00:35.840
I'm Mandy Hawken, CEO of Three Hawks Advisors, past president of ASCA, and the host of this recording.
00:00:36.079 --> 00:00:50.399
Earlier this year, the ASCA Board of Directors created a working group to study and explore how advances in artificial intelligence, or AI, could improve the efficiencies of both the administrative and clinical operations of surgery centers today.
00:00:50.640 --> 00:00:56.399
This is a subject that fascinates me, and I immediately said yes when I was asked to chair the new working group.
00:00:56.640 --> 00:01:17.599
I was particularly excited when I learned the group would include several very distinguished ASCA colleagues, including our new president Todrier, past president Mark Weiner, and ASCA board members Dare Meeks, Jessica Rodriguez, Alfonso Delgranado, and Lisa Cooper, along with ASCA CEO Bill Prentiss and his outstanding team.
00:01:17.920 --> 00:01:28.319
As I mentioned a minute ago, we intend to explore both administrative and clinical parts of AI that I imagine will require a number of educational tools for our members.
00:01:28.560 --> 00:01:35.760
For this introductory podcast, I want to focus on some of the ways that AI can provide administrator support at their surgery centers.
00:01:36.000 --> 00:01:48.159
To start this conversation, I've asked my really good friend and colleague, Becky Ziegler Otis, the assistant executive director of the ASC Quality Collaboration, to join me on the podcast.
00:01:48.480 --> 00:01:58.640
Back in May, Becky was part of an outstanding panel on AI at ASCA's annual meeting, and she too is enthusiastic and an early adopter of several AI tools.
00:01:58.799 --> 00:02:05.519
But before I bring Becky into the conversation, I want to say a few more words about the startup ASCA's AI Working Group.
00:02:05.680 --> 00:02:12.080
In our inaugural meeting, the consensus among the group was that AI indeed holds great promise for ASCs.
00:02:12.479 --> 00:02:16.560
But many of us are also discovering that there is a need to proceed with caution.
00:02:16.719 --> 00:02:22.879
We started by aligning on an understanding of the foundational elements of AI and healthcare for our working group.
00:02:22.960 --> 00:02:33.919
And to confirm we are on the right track, we asked ASCA members to complete a short survey about their AI usage and where they thought ASCA could be most helpful in terms of education and support.
00:02:34.080 --> 00:02:41.599
And convincingly, the respondents in the survey said AI, compliance, and privacy guidance were a top priority for them.
00:02:41.840 --> 00:02:53.599
From there, we invited a couple of leading healthcare attorneys to join us for a briefing, and work is already underway to develop some useful tools for ASCA members to remain safe and compliant when adopting AI.
00:02:53.840 --> 00:02:56.159
There will be more about this in the near future.
00:02:56.400 --> 00:02:58.719
So let's bring Becky into the conversation.
00:02:58.960 --> 00:03:01.039
Becky, welcome to the podcast.
00:03:01.360 --> 00:03:04.240
Thank you, Mandy, and thank you so much for having me today.
00:03:04.319 --> 00:03:07.680
I really appreciate being able to be a part of the podcast with you today.
00:03:07.919 --> 00:03:09.039
We're so glad you're here.
00:03:09.120 --> 00:03:18.400
And I know our members are thankful that you are here and for getting us kicked off at ASCA this year at the AI panel that you were a part of.
00:03:18.639 --> 00:03:27.520
I notice AI can be a little intimidating for some, and for others, it almost gives them a sense that they're going to have to learn something complicated and new.
00:03:27.599 --> 00:03:29.039
And who has time for that?
00:03:29.360 --> 00:03:37.759
Would you talk about a couple of practical administrative applications that are available to surgery centers today that don't have steep learning curves?
00:03:38.080 --> 00:03:39.280
Mandy, absolutely.
00:03:39.360 --> 00:03:40.159
I would love to.
00:03:40.319 --> 00:03:46.240
And I want to start by saying that the practical starting points aren't really all that flashy.
00:03:46.479 --> 00:03:49.919
They're really things that you're probably already halfway using.
00:03:50.000 --> 00:03:52.000
So I'm just going to point out a few of those.
00:03:52.240 --> 00:03:59.120
I suspect many centers are already using Microsoft Office Suite, which has Copilot that's built right into it.
00:03:59.280 --> 00:04:03.919
And you might already be leaning on Copilot to help you draft some emails.
00:04:04.080 --> 00:04:08.319
Because as you write your email at the end of it, you'll see a little pencil that shows up.
00:04:08.400 --> 00:04:16.319
And if you click on that pencil, it can give you some options to help you rewrite the content of the email, maybe even allowing you to change the tone.
00:04:16.480 --> 00:04:21.600
You can say you want it to be more casual or more professional, and it helps rewrite it for you.
00:04:21.759 --> 00:04:27.439
And I think what's important to know is that you have the ability to replace it or ignore it.
00:04:27.680 --> 00:04:30.959
And that's the piece that I think many people really get nervous about.
00:04:31.120 --> 00:04:38.000
They think that AI is just going to do the answer for them and they won't be able to help out with the response, but that's not how it works.
00:04:38.240 --> 00:04:48.879
Once Copilot does the draft, you absolutely have the ability to look it over and make the decision, have the human in the loop to make sure that you decide what you wanted to say.
00:04:49.120 --> 00:04:52.079
And speaking of Copilot, I have another practical way.
00:04:52.240 --> 00:04:55.360
It can also help you outside of drafting emails.
00:04:55.519 --> 00:04:57.199
Maybe it can help organize your day.
00:04:57.360 --> 00:05:06.959
For example, let's say you've been off for a bit, maybe it's over the weekend, and you come in on Monday morning and you want to find out what's been happening over the weekend in all of your emails.
00:05:07.040 --> 00:05:07.839
What's your priorities?
00:05:07.920 --> 00:05:09.120
How do you get your day started?
00:05:09.279 --> 00:05:16.560
So a simple thing you can do is just go into Copilot and say, let me know what I've missed over the past 48, 72 hours.
00:05:16.720 --> 00:05:23.120
And it will scan those emails, come up with what's the high priorities, what are the things that you really need to focus on.
00:05:23.279 --> 00:05:26.160
So it helps you with your efficiency.
00:05:26.399 --> 00:05:32.480
And if you haven't tried this, I have another pearl, and that has to do with agenda or minute preparation.
00:05:32.720 --> 00:05:35.600
And if you haven't tried it, I would say give it a try.
00:05:35.920 --> 00:05:42.879
As we all know, we have to be responsible for committee meeting minutes, whether it be the governing body, quality meetings, staff meetings.
00:05:43.040 --> 00:05:46.000
We're always putting together agendas and then we have to do the minutes.
00:05:46.160 --> 00:05:56.000
So the next time we're preparing for an agenda, for example, for your quality meeting, for your governing body, just say to Copilot, put my agenda together for me.
00:05:56.240 --> 00:06:03.600
I think you're going to be pleasantly surprised because you're going to come up with an agenda that is probably about 80% complete.
00:06:03.759 --> 00:06:16.639
And it's really going to help you with your efficiency because you're going to be able to go through it, tweak what it is that you want to change, have that human in the loop, making sure that you make the decision on what's in that agenda.
00:06:16.879 --> 00:06:20.079
But it's going to get you there so much faster.
00:06:20.319 --> 00:06:22.560
And I've really focused a lot on Copilot.
00:06:22.639 --> 00:06:29.600
And I just want to bring up that you can also use some other AI tools such as chat, GPT, or Claude that are available.
00:06:29.839 --> 00:06:32.079
And let's say you have to draft a document.
00:06:32.160 --> 00:06:33.680
And this is very helpful.
00:06:33.839 --> 00:06:45.120
Whether that document be a policy, whether it be a job description, a competency checklist, an orientation checklist, for example, think about using that AI tool to assist you.
00:06:45.279 --> 00:06:50.720
So, Mandy, all you really have to do to start the chat is ask the tool to draft a document.
00:06:50.879 --> 00:06:59.360
For example, I want to create a job description for a surgical tech for an ASC environment, and it's an ortho-specific center.
00:06:59.519 --> 00:07:05.279
The more detail you can give in the prompts, the better on how it's going to come up with the first draft.
00:07:05.519 --> 00:07:10.240
But I even learned recently that sometimes I'm not sure what I want to make for my prompts.
00:07:10.399 --> 00:07:16.079
So I will just say, draft me a job description and provide me with some clarifying questions.
00:07:16.319 --> 00:07:20.000
And then the tool will give me the clarifying questions to answer.
00:07:20.240 --> 00:07:25.360
I can come up with a draft, once again, is about 80 to 90% there.
00:07:25.600 --> 00:07:31.439
I can tweak that document, use my human in the loop, my decision making on how I want it to look.
00:07:31.519 --> 00:07:36.160
And it works much faster than if I have to just sit down with a blank piece of paper.
00:07:36.399 --> 00:07:40.720
So I hope those are some really practical tips for an administrator, Mandy.
00:07:40.959 --> 00:07:42.800
I think those are all great tips.
00:07:42.879 --> 00:07:45.279
And I love that you called out the human in the loop.
00:07:45.439 --> 00:07:53.759
The human in the loop gives you that intelligence and judgment and oversight while working with AI, and it also improves accuracy.
00:07:54.000 --> 00:07:56.800
It makes me think of self-driving car like Waymo.
00:07:57.040 --> 00:07:59.519
It just didn't just start driving on its own.
00:07:59.680 --> 00:08:09.040
A human was in the loop providing data that was used to train and guide the models, teaching it what it was supposed to do so it would have that safe driving behavior.
00:08:09.279 --> 00:08:16.480
I hope some of our members are using or beginning to try these AI tools you talked about in their everyday work or even home life.
00:08:16.800 --> 00:08:18.959
They absolutely can save you time, Andy.
00:08:19.040 --> 00:08:23.680
And as a leader, one of the things we can certainly benefit from is saving time.
00:08:24.000 --> 00:08:24.800
Saving time.
00:08:24.879 --> 00:08:26.000
That's what we all need.
00:08:26.240 --> 00:08:30.720
Becky, we have to take a short break to hear a few words from our podcast sponsor.
00:08:30.879 --> 00:08:32.960
Stand by and we'll be right back.
00:08:33.200 --> 00:08:42.080
This recording is being brought to you by Surgery Partners, a leading owner and operator of ambulatory surgery centers with more than 200 locations nationwide.
00:08:42.320 --> 00:08:54.960
Surgery Partners collaborates with ASC leaders and physicians to acquire, develop, and optimize surgery center operations and helps ASCs scale efficiently, strengthen performance, and drive long-term growth?
00:08:55.200 --> 00:08:58.159
To learn more, visit surgerypartners.com.
00:08:59.440 --> 00:09:15.039
Becky, let's talk about business associate agreements or BAAs, which should be well known to ASC administrators, but take on an added level of sensitivity when using AI tools that are coming in contact with PHI.
00:09:15.279 --> 00:09:23.440
Do administrators need to know about BAAs, even with some of these free AI tools that don't require an enterprise agreement?
00:09:23.759 --> 00:09:25.039
That's a great question.
00:09:25.200 --> 00:09:37.200
And here's the one thing I'd really want the listeners to walk away with, and that is paying for an AI tool does not mean automatically that you have a business associate agreement in place.
00:09:37.360 --> 00:09:39.360
They're really two different things.
00:09:39.519 --> 00:09:42.159
And you should never assume that you have one.
00:09:42.320 --> 00:09:52.320
You should always verify that it's in place before putting in any proprietary information or patient personal health information in any AI tool or chat.
00:09:52.559 --> 00:09:58.320
Absolutely, someone has to confirm that there's a BAA in place for that specific product.
00:09:58.559 --> 00:10:00.720
And it can't be just that we have an account.
00:10:00.799 --> 00:10:03.440
It's for what account and what plan.
00:10:03.600 --> 00:10:05.440
It has to be very specific.
00:10:05.759 --> 00:10:16.399
Certainly when it comes to a free Chat GPT or personal Claude or even your own personal co-pilot account, none of those come with a business associate agreement.
00:10:16.559 --> 00:10:20.399
Even a paid personal subscription does not count.
00:10:20.639 --> 00:10:26.240
It has to be a business or an enterprise account, specifically one that is configured for HIPAA.
00:10:26.720 --> 00:10:32.159
And where I can really see this going wrong for an organization isn't really one of those dramatic breaches.
00:10:32.480 --> 00:10:41.360
To me, it's someone on your team is just trying to be efficient and ultimately ends up doing something that gives your organization some exposure.
00:10:41.519 --> 00:10:45.360
For example, let's say they're drafting a response to a patient complaint.
00:10:45.519 --> 00:11:03.200
Maybe a family member has complained about their patient care experience, and they just want to paste that information very quickly into Chat GPT to help reframe what they're saying, but they accidentally forget to remove the patient's name or some other portion of the demographics.
00:11:03.360 --> 00:11:09.840
That information is now part of the large language model and in that overall chat, and you have some exposure.
00:11:09.919 --> 00:11:19.679
So you can see it's really just a simple mistake, which is absolutely why you need to have some guardrails in place, such as policy depicting AI governments.
00:11:19.919 --> 00:11:28.480
So my most practical advice is to make sure you always strip out any identifying or proprietary information before entering into a chat.
00:11:28.639 --> 00:11:34.080
Because once it's in the chat, it's part of that large language model and it's out there.
00:11:34.320 --> 00:11:37.759
Becky, I really like that you said confirm it's in place.
00:11:38.000 --> 00:11:41.440
That's such an important piece for all of our members to understand.
00:11:41.600 --> 00:11:49.840
When you're getting a new contract, as well as with vendors that have now added automation, review and make sure it is there.
00:11:50.159 --> 00:11:58.559
Hopefully, as we continue to build on our AI working group, we'll be able to provide some more education and resources for our members to get these guardrails in place.
00:11:58.799 --> 00:12:08.720
Well, Mandy, I think a lot of ASCs are struggling with knowing what is the right thing to do and not even knowing where to begin from the governance perspective.
00:12:08.879 --> 00:12:12.960
So the effort on behalf of this work group will certainly be well appreciated.
00:12:13.120 --> 00:12:14.159
Thank you so much.
00:12:14.399 --> 00:12:19.600
We've talked about document prep and protecting against HIPAA violations.
00:12:19.840 --> 00:12:29.360
I know that you've also had some personal experience with AI agentics that have become lifelike, intuitive, and engaging in conversations with patients.
00:12:29.600 --> 00:12:34.080
Can you share some personal interaction that you've had with your AI agent?
00:12:34.399 --> 00:12:36.960
This one was very, very exciting for me.
00:12:37.120 --> 00:12:49.840
I had a trained AI agent that was being used to complete the pre-operative phone call that, as we know, is typically done by the pre-op nurse prior to a procedure or a phone call in the ASC environment.
00:12:50.000 --> 00:12:53.519
Or in some cases, we go out to a portal and we complete that information.
00:12:53.759 --> 00:12:57.840
Well, I decided to give it a try using the trained AI agent.
00:12:58.000 --> 00:13:03.919
And when I received the call, it sounded very lifelike and it was very engaging.
00:13:04.080 --> 00:13:06.240
And in my case, it was a woman's voice.
00:13:06.399 --> 00:13:10.399
She asked me all of the required questions of your typical pre-op phone call.
00:13:10.559 --> 00:13:15.919
And my goal really was to try to stump this woman, to stump the AI agent.
00:13:16.080 --> 00:13:19.600
I wanted to see what would happen if I asked a lot of questions.
00:13:19.840 --> 00:13:21.840
So I really had a lot of fun with this.
00:13:22.000 --> 00:13:25.679
For example, I told her, geez, I know I'm on some medications.
00:13:25.840 --> 00:13:30.000
One is for my blood pressure, and I'm not sure if it starts with a P or a T.
00:13:30.240 --> 00:13:36.320
And the agent responded back with me on some very clarifying questions, and I was not able to answer them again.
00:13:36.399 --> 00:13:38.320
And she said, You know what, Becky?
00:13:38.480 --> 00:13:39.440
And she used my name.
00:13:39.600 --> 00:13:41.279
She said, You know what, Becky?
00:13:41.440 --> 00:13:45.039
I'm going to make a note of this and I'll have the nurse call you back and discuss it.
00:13:45.200 --> 00:13:47.919
And I thought, oh, that's a great answer.
00:13:48.159 --> 00:13:51.039
And she went out again and I kept trying to stump her in other ways.
00:13:51.120 --> 00:13:56.879
And I also remember I asked her about MH when she asked about having a history of that.
00:13:57.120 --> 00:13:58.720
I said, I don't understand what you mean.
00:13:58.879 --> 00:14:00.000
Can you help me understand?
00:14:00.159 --> 00:14:03.840
And she responded very appropriately with a definition.
00:14:04.080 --> 00:14:11.360
And when I said, I still don't understand what you're talking about, at that point, she said, You know, Becky, I think we're going to have the nurse call you back and help answer that.
00:14:11.600 --> 00:14:12.960
And we finished the call.
00:14:13.039 --> 00:14:22.000
And right towards the end of the call, where I thought the agent really went above and beyond, is started to frame for me all of the questions that I had asked.
00:14:22.159 --> 00:14:26.240
And she said, These are the questions that the nurse is going to give you a call back on.
00:14:26.399 --> 00:14:31.200
And she hit every one of them, even the ones that I had forgotten.
00:14:31.440 --> 00:14:39.919
So I felt very comfortable that the person that I was talking to, which was a trained agent, was a caring person, felt like it was a real person.
00:14:40.320 --> 00:14:46.799
And this isn't an actual ASC practice management software that we have available to all of us if we want to use it.
00:14:47.039 --> 00:14:51.039
Wow, this can really help with efficiency and effective use of staff time.
00:14:51.200 --> 00:14:54.399
I thoroughly enjoyed learning about this and being a part of the call.
00:14:54.480 --> 00:14:57.360
And to frame it all up, it was exciting.
00:14:57.679 --> 00:14:59.840
I also have had the same experience.
00:15:00.000 --> 00:15:12.320
I find it fascinating, the quality of their responses, how they change their rhythm and tone, almost how graceful they are when they're addressing you and not interrupting.
00:15:12.480 --> 00:15:19.440
So it's nice to see that you can really train that agent to do exactly what you want it to do, keeping that human in the loop.
00:15:19.679 --> 00:15:34.000
When we kicked off the working group, we had an executive speak to us from Google Health, and they had a lot of experience in this with pre-calling patients and how they've trained their agent to have those conversations, and it's been really successful for them.
00:15:34.159 --> 00:15:40.960
So it's nice to hear that people outside of the ASC space are using this and that's extra support for us to do the same.
00:15:41.200 --> 00:15:43.120
Those are some great insights, Mandy.
00:15:43.200 --> 00:15:53.679
And actually, when you were talking about it, it's amazing how when you listen to those trained agents, you have a preconceived notion that it's going to sound like a robot, that there's going to be uncomfortable gaps.
00:15:53.919 --> 00:15:56.080
That's not how it really works out.
00:15:56.399 --> 00:15:56.960
It isn't.
00:15:57.120 --> 00:15:58.559
It is not how it works out.
00:15:58.879 --> 00:16:07.360
Well, Becky, I would be remiss if I had you on the podcast without giving you a chance to tell us to what's the latest at the ASC Quality Collaboration.
00:16:07.600 --> 00:16:09.840
I know you have a new assessment tool.
00:16:10.159 --> 00:16:12.320
What are your responses looking like?
00:16:12.399 --> 00:16:15.519
And how are you thinking about AI and ASC Quality?
00:16:15.759 --> 00:16:16.720
Well, thank you, Mandy.
00:16:16.879 --> 00:16:24.639
The ASC Quality Collaboration has recently just finished their annual ASC Safety and Quality Assessment, which is complimentary.
00:16:24.879 --> 00:16:26.639
It's free, complimentary.
00:16:26.799 --> 00:16:30.559
So any ASC across the country can help participate in this.
00:16:30.720 --> 00:16:38.240
And what you're doing is you're benchmarking your center against national performance on quality and safety best practices.
00:16:38.480 --> 00:16:49.919
And this year we had it open for a six-week period of time and we had over 700 submissions, which compares very positively to a little under the 400 submissions we had the year prior.
00:16:50.080 --> 00:16:51.440
So that's fantastic.
00:16:51.600 --> 00:16:55.519
And I want to thank all of you who participated in that assessment.
00:16:55.679 --> 00:16:59.600
And if you happen to miss it this year, it is available next year.
00:16:59.759 --> 00:17:07.920
And if you're concerned you might miss the window for opportunity next year, go out to the ASCQC website, which is ascquality.org.
00:17:08.000 --> 00:17:16.160
And on that website, you will be able to put in your contact information so you can receive an email next year and be reminded of that.
00:17:16.480 --> 00:17:29.839
Also, with respect to the ASC Quality Collaboration, we're excited about our annual membership only quality conference that's going to occur later this month, which is really a great way for our membership to collaborate and network with one another.
00:17:29.920 --> 00:17:36.640
And we'll actually have the opportunity to connect with some key regulatory agencies during the conference like CMS and CDC.
00:17:36.799 --> 00:17:41.119
And then, Mandy, to answer the second part of your question, how can AI work with quality?
00:17:41.440 --> 00:17:44.319
Think AI can really shine in this area.
00:17:44.480 --> 00:17:47.440
And I think about it in a number of perspectives.
00:17:47.680 --> 00:17:50.319
The first is pattern spotting.
00:17:50.559 --> 00:18:00.960
Instead of just staring at a spreadsheet of incident reports or infection surveillance logs, you can really ask AI to help you look for patterns that might be surfacing.
00:18:01.119 --> 00:18:04.079
Patterns related to time of days, staff combination.
00:18:04.240 --> 00:18:07.759
But be careful what you put out there because you have to make sure it's protected.
00:18:08.000 --> 00:18:11.200
It can also help you with documenting your PI project.
00:18:11.359 --> 00:18:17.759
It seems like whenever I'm ready to start documenting a project, I have a blank page, I can't think about where do I begin.
00:18:18.000 --> 00:18:21.519
So use a chat to just put your notes in and get you started.
00:18:21.839 --> 00:18:25.519
I recently had an experience where I was doing this myself.
00:18:25.680 --> 00:18:30.480
I was struggling with something that I was trying to write up, and it was a great PI project.
00:18:30.640 --> 00:18:35.039
So I put some notes in on what I had measured, what were some of my findings.
00:18:35.279 --> 00:18:39.759
I was surprised to see how it came out with a great first draft for me.
00:18:39.920 --> 00:18:43.680
Once again, 80% there that I just had to turn into a study.
00:18:43.920 --> 00:18:46.160
You can also use it for a quick literature scan.
00:18:46.319 --> 00:18:52.000
Let's say you have a committee that says to you, I want you to check into unplanned transfers or same-day cancellations.
00:18:52.240 --> 00:19:01.839
It can help pull together what research has been done on this and list you with references and resources that you can tap into to make sure that it's valid information.
00:19:02.480 --> 00:19:10.079
So I would emphasize again that it's going to get you 80 to 90% of the way there, but it always requires that human in the loop.
00:19:10.160 --> 00:19:11.680
And that human is you.
00:19:11.920 --> 00:19:16.000
So you have to make the first draft into the final draft, whatever you're working in.
00:19:16.240 --> 00:19:20.319
So I'm going to conclude this all, Mandy, with saying, why not try out an AI tool?
00:19:20.640 --> 00:19:24.160
Dip your toes in and see what it can do for you.
00:19:24.319 --> 00:19:30.400
Because I think you'll be pleasantly surprised to see how it can help you with your efficiency.
00:19:30.640 --> 00:19:31.920
I agree with you 100%.
00:19:32.400 --> 00:19:38.480
I challenge our members, give yourself five to 10 minutes a day and try something new.
00:19:38.720 --> 00:19:41.759
And Becky, I want to thank you for spending time with me today.
00:19:41.920 --> 00:19:43.599
I love catching up with you.
00:19:43.839 --> 00:19:47.200
You are a true legend in our industry, and we're lucky to have you.
00:19:47.440 --> 00:19:50.559
Thank you for sharing all of your knowledge and insights on AI.
00:19:50.720 --> 00:19:52.880
And we will continue to pick your brain.
00:19:53.039 --> 00:19:58.559
And thank you for all the hard work that you and Nina Gwenes do at the ASC Quality Collaboration.
00:19:58.799 --> 00:20:02.000
It's such important work in support of the ASD community.
00:20:02.319 --> 00:20:04.559
Mandy, I thank you for having me today.
00:20:04.720 --> 00:20:10.079
And I so appreciate working with you and the fact that you reached out to me for my insights.
00:20:10.160 --> 00:20:22.799
And I also want to thank Ask and the work group for all the work that is being done on these AI initiatives to provide the guidance and the guardrails that we as the ASC industry absolutely need going forward.
00:20:22.960 --> 00:20:23.680
Thank you.
00:20:24.000 --> 00:20:36.880
Now, before concluding, I would like to acknowledge the support of our podcast sponsor, Surgery Partners, a leading operator of surgical facilities and ancillary services with more than 200 locations in 33 states.
00:20:37.119 --> 00:20:44.960
Surgery Partners offers multiple types of healthcare services that deliver high quality health care in a convenient and cost effective manner.
00:20:45.200 --> 00:20:48.319
To learn more, visit surgerypartners dot com.
