Skip to content
TrackPodcasts
scienceJan 12, 202631:36

Under Pressure - CI, Robot: Computer Vision for Sterility Assurance

About this episode

The dreaded missing indicator—you've trained on it, created systems for it, and yet trays still reach the OR without them. What if technology could catch that mistake before the tray ever leaves your department? On this episode of Beyond Clean, Gregory Warino, Director of Central Sterile Processing at Mercy Health, shares how his team used AI to tackle missing CIs across three facilities. From getting skeptical staff on board to watching the technology actually catch missing indicators before trays leave assembly, Gregory breaks down the real story behind implementing AI in SPD—from what it cost, how long it took, and whether the tech lived up to the hype. If missing indicators are a pain point in your department or you're curious about what AI means for your SPD's workflows, this episode has the answers you need!

After finishing this podcast episode, earn your 1 CE credit immediately by passing the short quiz linked here: https://www.flexiquiz.com/SC/N/episode31-02

Visit our CE Credit Hub at https://www.beyondcleanmedia.com/ce-credit-hub to access this quiz and over 350 other free CE credits.

#BeyondClean #SterileProcessing #Podcast #Season31 #UnderPressure #Indicator #CIs #Workflows #AI #Technology

Get every episode summarized

Each time Beyond Clean Podcast publishes, we email you a written briefing from the transcript — the topics, who appeared, and any specific claims, with the ad reads skipped.

Email me new episodes

Free for 3 shows. No card needed.

Transcript ready

334 searchable segments. Every word is indexed and playable.

Under Pressure - CI, Robot: Computer Vision for Sterility Assurance

Beyond Clean Podcast

0:00
31:36

Full transcript

Beyond Clean PodcastUnder Pressure - CI, Robot: Computer Vision for Sterility Assurance. Machine-transcribed; use the interactive transcript above to jump the player to any line.

and what they wanted us to do was look at every single trade and at the time I was really down on stuff and I said well wait a minute we don't have to open all these trays we could go back and look at all of the pictures that we took and see if the correct spoon is there or not and so that saved us a ton of work I mean able to go back and look at the pictures versus opening all these trays and reprocessing everything reassembling also yeah there's many more benefits with having that final picture or that indicator picture Rising above the buzz of ultrasonic cleaners and the clanking of stainless steel are the ideas and voices that are changing an industry you're listening to the Beyond Clean podcast the central nexus for the people processes and products that are improving our sterile processing world each week we speak with frontline technicians CEOs engineers and entrepreneurs with a common

goal to help you fight dirty every instrument every time whether you're tuning in for education or inspiration we're glad you did now turn on those washers and turn up the volume it's time to go beyond clean all right clean freaks welcome back to another episode of the Beyond Clean podcast this is season 31 under pressure all about steam sterilization I'm your host Bobby Parker and joining me in studio for this whole season is our guest co-host Melissa Morgan Melissa welcome back to the studio thank you so much Bobby so glad to be here yeah I'm glad to be here too you know I'm I'm in Louisville Kentucky and we get a little bit of snow each year usually just one or two snowstorms and we just have one this morning and so my my legs are freezing I've been like go sledding down the hill with the kids I'm ready to sit down and lock in for sterile processing

and talk about some SPD stuff but you're in central Florida is that right absolutely making sure we were in the pool just yesterday afternoon enjoying enjoying the sunshine that that sounds so good right now I'm freezing well we've got a really fun topic you know today's episode is one of those conversations where we get connected with somebody in the industry who's doing something cool and immediately know that we've got to bring this person on the podcast and share this story a little bit broader we we cheekily titled the the episode CI robot using computer vision for sterility assurance because you know AI is eventually going to take over the whole world you know if you've ever seen that Will Smith movie but but no there are some really cool things out there of how computer vision and AI is being used in departments and this one about using computer vision to help with chemical indicators as a topic I'd never thought about

and yeah we've got a really cool success story to to be bringing today so we will be right back after this short break with Gregory Warino to talk to us about his pretty cool AI project are your SPD and OR department struggling to keep up with demands learn how the Stereq multi-trace sterilization system can improve patient safety and throughput by eliminating many steps in your workflow reduce costs OR turnover time IUSS and surgical delays and increase confidence in sterility and staff satisfaction the Stereq is the only container FDA cleared for sterilizing lumens as narrow as 0.7 millimeters and as long as 850 millimeters with no limitation on quantity to find out how you can save significant time and cost by using the Stereq go to www.thestereq.com

that's T-A-G-S-T-E-R-I-C-U-B-E dot com in a world of AI automation robots and advanced technologies the future is now and it's time for every sterile processing department around the country to start considering DSI's new high density robotic storage system for those hard to handle DaVinci instruments this new system is a game changer designed specifically for peelback or blue wrap robotic instruments in a manner that exceeds industry guidelines honor the days of using chrome wire cards made for the food and beverage industry 50 years ago these DaVinci tips are expensive and extremely valuable for every hospital it's time to protect those instruments in a specialized system offered exclusively by DSI the instruments lay on their back in individual compartments running lengthwise in the tray

and can store up to 108 arms within one cart DSI's new system will protect DaVinci tips like no other system in the marketplace today in a world controlled by robots and change coming at us in lightning speed now is the time to take action and connect with DSI contact DSI at dsidirect.com to find out more from the studios of healthcare HQ you're listening to Beyond Cleanse the global voice of sterile processing and joining us now in studio is Gregory Moreno director

of central sterile processing at mercy health in youngstown Ohio Gregory we're happy to have you welcome to the show thank you for having me absolutely and Gregory we we had a conversation about a month ago it feels like about something cool that your department has been doing with your chemical indicator problems and I just knew that we had to have that have that conversation on the show here so I'm excited to get into it but before we do that I want to give people an opportunity to to get to know you you want to introduce yourself and you know give us a little thumbnail of your career yeah my name is Greg War you know on the director for central sterile processing for mercy health the youngstown in Ohio I've been in career op the services since 2008 I've been working as a sterile processing leader since 2013 and I'm really excited of where we're going with sterile processing especially using the tracking systems and incorporating AI within

the tracking systems to get better outcomes with the trays and products that we produce yeah it's moving so fast isn't it it's crazy so so many opportunities we have in front of us you you're interested introduced yourself as Greg is that is that how we're going to roll with the rest of the episode today okay that's fine all right you'll be you'll be you'll be Greg to us first name basis here Greg you said your title is director of central sterile processing so if you got a couple of facilities that you're overseeing or is that a single facility director roll what's that look like hi I have multiple facilities I oversee operations at three hospitals in one surgery center okay when you started out in sterile processing leadership back in I think you said 2013 did you did you have a multi-site director role in your sites as like the mountain to climb or has this been a big surprise for you no when I started in in store processing I started as a manager and I reported to a pair of services director at that time I've since then

have switched organizations when I started with the my current organization I went in as a supply chain manager who also oversaw the operations of sterile processing at that time I got hit with joint commission two weeks in on the job he followed prior to me ignored sterile processing all together so of course I got hit with joint commission big time oh we turned around the department over it away and the president of mercy young sound who is also over all of perioptive services saw what I did and then offered me the position to be the director of central surprocessing for the market I never thought it was gonna turn out to that but here I am yeah yeah well good things happen when you apply yourself and do a good job and and folks notice so joint commission is definitely an opportunity to shine right that's right well congratulations on your success I'd love to hear success stories like that and that you spread in some of that

success to the rest of your to the rest of your market there one more personal question before we get into the technical bits so the interview today is about an AI solution that your team is implementing any background you have in you know doing computer science stuff or just kind of a personal hobby because it usually takes a special kind of innovative thinker to really be one of the first guys through the wall implementing a new technology like this I'm just curious what what inspired you I really don't have any background with it and more or less just working with the census team and trying to come up with ideas how to better their system we're doing some work with AI with instruments and it led to them going to grapes yeah I love that well we love talking to innovators and shakers and movers and the industry and so yeah even even if it's not your solution sometimes just being the one brave enough to to try something new and say hey we could make this better that's I'm sure really valuable for the team that you're working with from your

instrument tracking perspective but also for your organizations enough teasing about the topic Greg why don't you walk us through what was the problem that you had with indicators in your facility and why you took the route of going to your tracking system and AI to try to to try to solve that problem yeah so just trying to get a chemical integrator into every trip we've done multiple times putting the indicator on the count sheets putting notes up trying to comment stuff to put the indicator in the trade as the very first thing or maybe as the very last thing um and it seems like no matter what we did there was always a missing indicator and it was the worst case is if it's a one-up kind trade it's up to the OR and they open the trade they need it and there's no indicator so now they can't use it so after a couple of incidents I was working with the sensor track and we're working with the instruments trying to identify instruments with AI and after

a couple of incidents I said instead of looking at the instruments right now I know a big one would be indicators because I just had a couple of cases that this happened like what if we could train the AI to look for an indicator so of course the team in the sense that said yeah we can make that happen so we would end up going that route and then young sound with my three hospitals one of the first to be test pilots and we kept testing the machine or the AI over and over getting stuff to use the feature and since it has since monster co-pilot program where it will recognize indicators within trace and say if it's missing or not yeah and you've got all the royalties on that idea right that's just kidding so that's really cool that's really cool so I can certainly resonate with that problem the issue of indicators being in trays you know been a technician Melissa I'm sure that you've experienced the same of like we've done all this work on this tray had everything

ready like you said Greg it may be the only one in the house but there's not an indicator in it and we've tried all the same solutions with the popups with putting it physically as an item on the count sheet that has to be physically checked off we've tried yeah the popup that has verifications that they have to type their initials into or scan their badge saying I verify that it's in there we've tried the double checkers of having the second person check for the indicator and you're right there is no there's no perfect system that will eliminate the indicator issue although I did visit a facility one time that had a triple check Melissa they had they had somebody set up the tray they had another person come back take the instruments apart and reset up the tray and then the third person was the one who would package it and confirm the indicators are in there and they hadn't had a missed indicator for a very long time in that facility but my goodness what an investment from a staffing perspective and then the efficiencies perspective right yeah how how many

trays are you processing yeah for sure all right so how did how does this how does this system how does it work you you mentioned kind of a computer vision and it's able to see the indicators is there like a special station that all the the trays have to move through or how does it how does it identify the indicators for you so we use web cams those are set up at every a workstation they're mounted right up by the lights and what happens is we assemble the tray mark off everything that's inside that tray and then once we get complete a camera will come up here on the street and then we have to hit the picture button bomb and at that time it takes about one second for the image to be recorded and then the AI recognizes this if a indicator is in the tray or if it's not now sometimes it will say that it's not there and it is there and we found it's more of a lighting issue because we'll have you know bright lights inside the department but if we just move that indicator slightly it will pick up that is there and then we have that

reported picture the beauty of it is we have a recorded picture so anytime that they are saying that we've missed the indicator we can go back and look at that and say you know was it there was it not and we have that picture proof yeah that's the technicians dream because I don't know if I don't know if either of y'all have ever been been talked to or written up for a missing indicator that you're sure is in there or been on the other side of the desk I'm sure and talk with folks about it and they're like no I'm sure it was in there and all you've got to go on is you know this is this is what the OR was telling me so you're having that picture would be would be cool is I've never been written up for it Bobby but I sure have popped a few locks in my days to make sure it was in there yeah yeah you're like laying down on the pillow at night and thinking oh man I can't remember if that whole stack of loners as indicators in it so yeah that's that you don't you don't want to be in that situation for sure Greg the AI workflow so you said there's there's webcams at the station are these like attached to the workstation like looking

down at the table or they on the computers like I'm just trying to get a mental picture of what your setup looks like yeah now I I ended up mounting them to the lights above the workstation we use the old stairs tables most tables are similar right the surface so the ones that go up and down with big light bar overhead exactly yeah yeah okay so you've got a camera this looking down at the at the whole workstation then and then how how have you integrated the computer vision piece into it is it just when you're finished with with the with the count sheet and click I'm finished it just pauses for a second to look for to look for an indicator is that is that how that goes so yeah once once you hit the finished button of assembling the trade within census then the the picture will come up on the screen it will take about a second and yeah the AI just looks for that indicator in in the trade and says it's here it's not there in the beginning uh they also had the option of a thumbs up a thumbs down so we're able to tell the census people

if a indicator was inside the trade and it wasn't recognizing we were able to give it a thumbs down and they had that to add it to go back and adjust the system uh to make it work a little bit better so I have this vision of big brother hanging over my head making sure I'm doing the right things I think we're all thinking in the list so Greg how did your staff did they have challenges or concerns about this technology resulting in other processes for inspection just curious yeah absolutely and in the very beginning when I said there you were going to do this we're going to be a test pilot when I told him that we're going to take additional steps to do this and then it may not work 100% at a time I got a lot of mean faces but the benefit was I said we would have the picture proof of anything that we did so if there was an incident of saying hey there's no indicator we would be able to go back and show the picture proof of that that got

total buy and then as we worked with the system given the thumbs up thumbs down and getting census to tweak their system a bit and we got better and better we went from a three-second picture down to a one-second picture and staff are loving it because we have that picture proof it only takes a second and it's it's closed them down for that one second to really make sure that indicators inside that tray I'm thinking through this value of having the picture piece and how that was a big cultural win I can certainly think of a couple of examples in my days of where if we had had a picture of what the tray looked like when we put it together that would have been a big win for us just in having a conversation and trying to figure out what had happened probably the craziest example of that is somebody lost their name badge and we couldn't figure out where it was and after enough searching we actually looked through some security video of what was like

what happened to be going on in that like corner of the department and found some time stamps of when they had it when they hadn't had it and the only thing that happened in between there is they wrapped some loners and we popped them all open and there was a melted name badge inside of one of those loners it's so thankful we caught it but like yeah having those images would be helpful has this process changed ended up helping in any of those kinds of unexpected ways that didn't have anything to do with indicators absolutely I've had multiple conferences where maybe a tray was looks like it was never assembled and I would get the picture from the manager we're paired up with directors saying hey well this is how it was sent to us and we were able to go back and pull that image and say well here's when we assembled it last this is what it looked like I also had just a very recent event where there was a joint commission and they found a spoon a peel pack spoon and it wasn't the correct type of spoon but it was something from like 2008 so then they started popping open trays and what they wanted us to do was look at every single tray

and at the time I was really down on stuff and I said well wait a minute we don't have to open all these trays we could go back and look at all of the pictures that we took and see if the correct spoon is there or not so that saved us a ton of work I'll be able to go back and look at the pictures versus opening all these trays and reprocessing everything reassembling so yeah there's many more benefits with having that final picture or that indicator picture yeah for sure so thinking through kind of the change management piece of it you had some some understandable opposition at the front and like what did that adoption look like for your team how long would you say it took for you to go from announcing a new idea to everybody being on board and finding some success with it where it's not getting in the way it worked well I took I think it took about a series of three months just really getting the system tweaked and getting total buy-in especially after a couple of conferences where we had the ORC there was no indicator

or if there was something wrong with the training we were able to go back and prove that now this is the way we assembled it the indicators there here's the picture proof and just showing that once we we did that a couple times I got total buy-in at all three sites with all staff taking those pictures and then also tweaking it from a three-second video or picture down to a one second but that also got a lot of buy-in yeah I would imagine three seconds felt much longer than three seconds really is as fast as they move sometimes in that closing that tray getting enough the table space so Greg I'm fascinated by this process and why didn't I think about it before they did and so when you think about you know your data your results where did you start and where are you currently and have you seen any back slides or things that continue to improve I haven't put any real data for this yet I can tell you that we haven't

had many incidences of a missing indicator and usually when we do we have that picture proof that now it was there which makes me think as maybe as the OR moving too fast then it's flying out because we have that picture proof it was there during the assembly I did have a recent one incident where the picture was not taken so I didn't have anything to prove but also had the conversation with that technician saying hey this is saving us just make sure that we're taking that picture and I honestly don't know how he missed taking the picture because I have it built into the workflow it's a requirement and the only thing I think of is maybe you shut down the system at that time for some reason yeah that was going to be my question is is it is it a forcing function like is it a requirement to get the the count sheet you can't move forward until you've got that picture with the computer vision giving you the thumbs up on the indicator yeah now I built it into a workflow so that it is a requirement yeah well yeah we'd we'd love to to hear about the data and the results on that of you know what your what your trend graph looked

like from a number of reported missing indicators from from before and after but it sounds like at least kind of experientially it's plugged a few holes in the dam of a phone calls that you're getting from the OR so that's very cool so as we think about technology and the team like realize and how helpful this was has it spawned their thoughts around other ways to use technology in this kind of process improvement type work we've been brainstorming with it how can we keep going with with this type of technology one is you know the recognized instruments but that takes a lot of work programming I would imagine and trying to get every single instrument that's out there I don't want to say impossible but it's going to take a lot of time one way and I think what happened is with the with pdick total joint sets so we know that those typically come back to us on next step and maybe if we could use that AI to recognize just those instruments and maybe there's

something that says hey this instrument doesn't belong in this trade but it goes to that that would you know maybe prevent missing instruments in the future I don't know yeah that would be huge because most of the time those orthopedic ones are loners so we're not familiar with them we don't get trained on them I have similar thoughts about instrument recognition I think I think it's powerful for really strange rod instruments you're not used to seeing and nobody can figure out what it is and there are a couple of platforms out there they're doing a lot of really great work on that instrument identification piece of it for just you know building your major ortho tray or your general major tray it almost feels like it would slow you down more than help identifying yes that is a six inch curved cry like you know like not not the most helpful but but yeah things that you're not identifying are not trained to be able to identify in s pd like those orthopedic loner trays and being able to tie it back to who it belongs to I think that would be pretty powerful

I'm convinced most of the instruments that we lost from our orthopedic vendors went out of the hospital in somebody else's tray like we put the striker thing in a deputre and deputre their trays up and off it goes never to be seen again so yeah if we can figure out how to stop stop some of that I think that would be that would be pretty powerful thinking about your setup Greg with you know you've just you've got a webcam on your workstations and a button that you click to take a picture I'm sure from a development standpoint from a computer vision standpoint there's a lot of complexity that goes on behind the scenes that that your instrument tracking system team was able to help you with at the same time um and I don't think it would be terribly hard at all to Mickey Mouse this thing and just to have a webcam with you know with a camera button then she just Melissa is that the second time I said Mickey Mouse's eyes of this season I didn't come up with other illustrations because I'm in Orlando Bobby that might just destined for the magic kingdom

or something and it's true I look at you and I think yeah that Mickey Mouse is the right illustration but you could maghiver this thing and have pictures at your workstations and be able to succeed or am I simplifying this too much what do you think no I think it's very possible the issue is if you were to go back to pull that data I don't know how you may have a lot of pictures that look through to try to find the exact tree that you're looking for and then just get in a picture or how do you identify that tree is what it is yeah so it's where do you store the files and how do you name the files well it's findable yeah all right well yeah here's the gauntlet being thrown down for our instrument tracking systems even if it's not incorporated into some kind of AI computer vision make a way to attach a photo to account sheet so that it lives with that sterilization record like what you're talking about and you know if if we have to shortcut the computer vision piece of it for identifying a identifying a chemical indicator think I think at least having that

that photo yeah it would have it would have changed change some of the conversations culturally that we had I can definitely see that benefit hey Bobby I've a follow up question yeah go ahead Greg were they it didn't matter what kind of indicator you used right like it you know for low temp or steam I mean it it just sits as that indicator correct it just recognizes that indicator and I think it's really looking for that rectangle shape that's what the AI is looking for I think remember correctly back in the very beginning the indicators that had the long tail the extender piece but I think we had a little bit of issue with those but other than that yeah it's just looking for that rectangle yeah and it sounds like like all things computer vision it probably gets smarter the longer you use it I imagine yeah new weird indicators are gonna show up and it'll figure it out and then keep on moving with identifying it the right way have you run into any issues where somebody's stuck because the thing won't let them finish a tray like it just for whatever

reason can't see the indicator or has it been pretty smooth at this point so after in the very beginning yes that did happen but since we wait it no it doesn't happen anymore yeah crazy crazy it seems such an obvious solution of you know because we're all the time talking about having an extra set of eyes to make sure that we got the indicators in there why not train some computer eyes to be able to do that for us makes a lot of sense well Greg we're we're about out of time here but I want to give you the last word any final exhortations for our teams out there that are maybe trying to figure out how to innovate or how to improve their departments using technology well I would just say go for it we know census is already has this um but I'm and I'm not sure if there's other tracking systems out there that have used anything like this but it just keep going for it throw those ideas out to those tracking systems maybe they will work with you and

build it into your current tracking system awesome well Greg congratulations on getting this implemented in your facility and getting now to reap the benefits of some hard work early on and earning the trust of the team I am sure that you're going to think of some other clever ways to use AI and computer vision to help improve reliability in your departments I think this is a fantastic example of what we often talk about with AI and sterile processing you know not necessarily so concerned that AI is coming for all our jobs because there's a very human element to this but how can we use AI to help augment our technicians to make them more reliable to make the difficult and redundant tasks of the department more reliable and yeah you knocked it out of the park so thanks for sharing your story with us today Greg thank you for having me yeah absolutely hey if anybody wants to get in touch with you uh is linked in the best way for them to reach you yes all right well reach out to Greg on LinkedIn hey one more things again Greg for joining us on

the interview thank you and that was Gregory Moreno director of central sterile processing at Mercy Health in Youngstown Ohio and Melissa what a cool story of what they're doing out there I mentioned it earlier in the episode I really do love talking to the folks who who have an idea they find somebody who's got the ability to deliver on that idea and they just roll it out in their department and see what they can do with it and it sounds like it's been a big success for them yeah it's such a cool idea and you know continually ask myself why I didn't think of that what's wrong with us Melissa earlier in my career like seriously Bobby but we're going to do the podcast about it so we'll be good that's right that's too funny you know I did have an indicator idea a long time ago long before computer vision and not nearly as sophisticated or as sleek as this because I wanted some kind

of forcing function like that that you cannot click the the complete button until you were sure there was an indicator in the tray so when I had pictured was an indicator a chemical indicator like dispenser vending machine robot box you know just a little box on your table and when when you try to hit count sheet first you have to click giving my indicator and it like dispenses two of them out for you and you take them and you put it in the tray and then you hit okay but if you try to hit okay before it's dispense those or if you try to hit okay and they're still hanging out of the dispenser it won't let you keep going you know some kind of forcing function like that but somebody I'm sure would still manage to take them out of the machine and set them on the counter not put them in the box so this is yeah this is definitely the right answer so I hope that it has inspired some folks to ask some questions and you know go to Best Buy and get them some web cams and start talking to IT because I think yeah definitely definitely a good way to implement this technology for higher reliability well that's going to do it for today's episode

not the most related to steam sterilization but certainly yeah when we're talking about getting sterile trays to the OR having that chemical indicator in there is so important we will be back next week on this season 31 coming at you about underpressure steam sterilization but until we talk to you then we'll leave you as we always do here at Beyond Clean on behalf of Melissa myself and the rest of the team keep fighting dirty thank you for listening to this week's episode of Beyond Clean as a reminder you can help support us by subscribing to Beyond Clean on your favorite podcast app or by downloading the smartphone app on iPhone or Android simply search for Beyond Clean in the app store or Google Play the best thing about downloading the smartphone app is that you can access bonus content for certain episodes and view episodes in certain categories like articles on the go and vendor spotlights are you following us on LinkedIn or Facebook yet

if you are in the 11th episode or post then let your social network know about it like comment or share our posts along with your thoughts and keep the conversation going if you have any topics or guests that you would like to recommend for a future episode just send us an email to infoatbeyondclean.net we look forward to hearing from you

More episodes

More from Beyond Clean Podcast

View all episodes →