top of page
Horiz_red_podcast.png

#459 - Which AI tools are actually ready for the NICU right now?



Hello friends 👋

Is your EMR actually built for neonatology, or just retrofitted to survive it? Ben sits down with Dr. Lindsey Knake, clinical assistant professor and associate chief health information officer at the University of Iowa, to talk AI in the NICU. They cover the real difference between Epic's homegrown tools and third-party options like Evidently and NABLA, why a six-month hospital stay breaks most chart summarization tools, and how ambient AI is starting to change documentation and family counseling. Lindsey also shares practical paths into clinical informatics for clinicians without a coding background. A grounded look at where NICU technology actually stands right now.


Link to episode on youtube: https://youtu.be/tGiwGXfiZIM


----


Short Bio: Dr. Lindsey Knake is a Clinical Assistant Professor of Pediatrics-Neonatology at the University of Iowa, where she also serves as the Associate Chief Medical Information Officer. Leveraging her background in biomedical engineering and a Master of Science in Biomedical Informatics from Vanderbilt University, Dr. Knake specializes in bridging the gap between clinical care and technology. Her professional focus includes both operational and research-driven initiatives aimed at optimizing electronic health records to improve patient outcomes and enhance clinician satisfaction. Dr. Knake completed her medical degree at the Carver College of Medicine, her pediatric residency at Baylor College of Medicine, and her Neonatal-Perinatal Medicine fellowship at Vanderbilt University Medical Center.


----


Links:






----


The transcript of today's episode can be found below 👇


Ben Courchia (00:00.824): Hello everybody, welcome back to the Incubator podcast. We're back today for a much-anticipated interview. We have the pleasure of having in the studio Dr. Lindsey Knake. Lindsey, what's going on?


Lindsey Knake (00:12.140): Hey Ben, I'm so excited to be back. I've been a fan of this podcast since the beginning.


Ben Courchia (00:17.496): For people who follow along, Lindsey has hosted the podcast as a guest host in the past — you'll find that on the website. We've always connected with Lindsey because of her innovative nature. She comes to us from the University of Iowa, where she's a clinical assistant professor and associate chief health information officer for the Stead Family Department of Pediatrics. You did your undergraduate training in biomedical engineering at Iowa, earned your MD at the Carver College of Medicine, completed pediatric residency at Baylor, and went on to fellowship training in neonatal-perinatal medicine alongside a master's degree in biomedical informatics at Vanderbilt University Medical Center. Your work focuses on AI (Artificial Intelligence) and the analysis of continuous NICU (Neonatal Intensive Care Unit) data, and you're an active member of the NeoMind AI Collaborative. Welcome back to the podcast — we're excited to talk about AI, electronic medical records, and what's currently going on in that space. My first question: you have a relatively traditional path through college and medical school. How did you become an expert in biomedical informatics, and what does a more traditional path into this kind of role look like?


Lindsey Knake (01:57.302): Honestly, like everybody, when I was young I didn't think this was my pathway — now when I'm writing grants, I say, "of course I knew this was my pathway." In undergrad at the University of Iowa, I studied biomedical engineering, more as a backup plan to medicine. I didn't want to just be a biology teacher if I didn't get into medical school, but technology interested me, so I thought I'd try biomedical engineering. During that time, some of my summer research jobs involved computer programming — I was part of a digital human modeling lab, creating predictive algorithms for digital humans, like designing a tractor or Caterpillar cab and predicting the best posture and comfort for a person sitting in it. That got me into programming and building predictive models early on, which really helped develop my research career and skill set. It also helped me realize I don't love sitting in front of a computer all day — the extrovert in me needed more people to talk to — but it was a great setup for medical school at Iowa. There, I stayed connected to my engineering roots, because at the time you could collaborate with engineering students on their senior design projects and file patents together. As a medical student, I collaborated with engineers on a couple of patents — none became companies, but that's a whole other conversation for another day.


Ben Courchia (03:49.260): For people listening who are already attendings — trained the traditional way, but who love technology and innovation and want to be involved in how their center uses the EMR (Electronic Medical Record) and new modules — is there still an opportunity to gain that kind of expertise in a less formal, but still professional way?


Lindsey Knake (04:30.163): Definitely. That's part of the point of this conversation — AI isn't going away, it's becoming more integrated into clinical practice and daily life. There's training everybody can do. Part of what NeoMind AI does is put out education — we have a monthly newsletter with training bits. But it really comes down to what kind of learner you are. My first recommendation is to just start using the tools — ChatGPT is free, jump on it, use it for planning vacations or daily life, and get familiar with it. Then find out what tools are available at your own institution to play around with and learn. There are also lots of online modules, or if you'd rather sign up for an online class, some universities — Iowa included — run learning sessions people can attend. There are a lot of opportunities out there; people just need to be aware of them and look.


Ben Courchia (05:43.878): We're going to talk a lot about Epic, the EMR you're familiar with and innovating on. For people in a similar situation, does Epic offer training to help someone become an "Epic superuser"?


Lindsey Knake (05:59.296): Yes, great question. There are Epic Super User courses, mostly online, usually with set timings so you can ask questions and get feedback. There are also smaller things — I think they're called Epic SmartBytes — short 10-minute videos. And your hospital's Epic training team can usually connect you with more resources. At Iowa, we have clinical informatics specialists who meet with people every so often to keep them updated on new features, since things keep improving and you can keep enhancing your skills.


Ben Courchia (06:45.986): We see a lot of innovation happening in the public sphere — you already mentioned ChatGPT and similar tools, which we're not going to dive into today. It's developing so rapidly out there. Can you give us a cursory sense of where Epic stands today? Is it moving as fast as entities like Claude and ChatGPT? If yes, what does that look like — and if not, where are we?


Lindsey Knake (07:27.273): As much as I want to say healthcare is moving fast, we're moving fast for healthcare, but not as fast as Anthropic and others. My husband codes, and he uses Claude constantly — what he can build so quickly with what people are calling "vibe coding," or really just knowing code and telling AI what to do, is remarkable. His comment is that it can type faster than he can, so it codes faster than he can. It's really exploding, and I think every industry will expand this way. I'm proud that healthcare hasn't put its head in the sand the way it has for other things over the years, and that we're starting to adopt these technologies — at our own pace, but in the right direction. Physician burnout and documentation burden have been huge issues for years, and that's really where this has taken off: ambient AI documentation, ambient AI for coding and billing, writing notes, and putting in orders in the EHR (Electronic Health Record). Epic's vision over the next couple of years is that everything could become voice-activated — so you're not rounding in front of a computer, and outpatient physicians aren't multitasking with their nose in a screen. Instead, AI listens and helps queue up orders and diagnoses, drafting your note so you can review and sign it much faster later.


Ben Courchia (09:04.941): In terms of currently available modules — I have a feeling a company as big as Epic is developing tools they hope will apply across every medical specialty and every part of the hospital. We could see that positively, or we could say, that's not great, because I'm not an ambulatory rheumatologist, I'm an inpatient neonatal critical care physician. Can you tell us about these system-wide tools you've had experience with, and your thoughts on them?


Lindsey Knake (09:50.902): For people who don't know, I've only ever used Epic — so I talk about it a lot, though they don't pay me, I don't hold stock. I do sit on the Epic Neonatology Steering Board, because I think it's important we give Epic feedback on what the neonatology community actually needs. That said, there are other third-party companies that might actually be better at customization and optimization for specific areas, because when Epic releases something, they try to make it work for everyone — but it's not necessarily trained on neonatal data. For example, Epic released a chart summarization tool meant to give a quick summary for a patient you haven't met yet. Our neonates, who can be in the hospital for six months, broke that very quickly, because the first release only looked back 30 days — not nearly enough. They've heard that feedback, and future iterations will improve, but there's a real balance in how much these tools can truly summarize. At the University of Iowa, I think we've struck a nice balance — implementing new Epic technologies where they work for us, and also bringing in third-party tools with better or different customization.


Ben Courchia (11:32.686): How do people find out which tools are actually well-designed for neonatology versus completely inept for a preterm baby or an ICU (Intensive Care Unit) patient? Using your chart summarization example — how long would it take someone to even realize it only looks back 30 days?


Lindsey Knake (11:54.784): It's interesting — each hospital system handles this differently. Some do a lot of education upfront; in others, tools just turn on and people stumble onto them, and you have to actually read your email to catch the education they send out, which isn't always easy. A lot of the time, tools just appear and people play around with them. That's part of why I wanted to talk about this on the podcast — it's useful to have an IT or clinical informatics person within neonatology, or your department, to go to with questions like, "I've played around with this, I like it for my one-liner, maybe it doesn't work for the whole summary, but could it help with handoff?" Having those conversations and vetting tools before using them blindly is always important. I try to be that champion who's willing to read the fine print and figure out when a tool is actually ready for use in neonatology.


Ben Courchia (12:58.498): The sticky question is always whether these modules are something you can unlock just by clicking a button — sometimes that's literally all it takes in the hospital system — or whether it requires financial investment. A lot of people worry, "we don't have the budget for this," but maybe some things are free and some aren't, or it's a mix. What have you gathered?


Lindsey Knake (13:27.041): You're exactly right, it's a new space, and different companies and even different Epic modules vary a lot. Epic is moving toward offering a full AI suite — you pay a large check upfront and get all their AI technologies. But some institutions might just want to trial one or two tools, and pay differently for that, sometimes releasing access to pilot users. If you're the clinical informatics champion in your NICU, you might become one of those pilot users deciding whether it's worth turning something on now or waiting for future versions. It's similar with ambient AI technology — where a device listens to you and helps draft a clinical note. That started with outpatient notes, since that's where the initial need was greatest, but it's moving into the inpatient world too, capturing what's said on rounds and drafting notes from that. Ambient AI still isn't great for highly templated notes — if your notes are already heavily templated, that might actually be faster than using AI-drafted text. At Iowa, we use NABLA, a third-party company. We went live essentially all at once — we paid for the license and gave access to everybody, so any inpatient doctor could use it and try it out. I've tried it in the NICU, and it's really helpful for things like complicated care conferences with families — even with interpreters, it can summarize everything discussed. It's also useful for transport calls, since I don't have templated notes for those, or for prenatal consults where I'm giving recommendations to a mom — instead of typing everything out afterward and possibly forgetting details, I get a generated summary of what I actually said, which I can then edit.


Ben Courchia (15:52.952): So what you're describing is essentially taking a tool with one intended use, as defined by its developers, and finding immediate, ready-out-of-the-box value for something else entirely. It comes back to your earlier point — you just have to play around with these tools. Even if an ambient AI tool doesn't summarize your note well, it might be great for something else entirely, improving your workflow in an unexpected way. It sounds like that principle will apply to every tool we're offered through the EMR going forward.


Lindsey Knake (16:40.032): I definitely agree. As with anything, the first version isn't the best version — tools need feedback, and you need to play around and figure out what you can customize. It's exciting to see all these third-party tools coming out that are genuinely useful. And if your institution hasn't turned something on for inpatient use, or hasn't even considered piloting it, you need to be the advocate — ask for a single pilot license and see if it works for your unit. Sometimes you have to be the person who asks to turn it on.


Ben Courchia (17:19.438): Without putting you on the spot — you're one of the people I know who's tested and worked with this the most. Of everything you've tried that's genuinely market-ready, not just some prototype years away, what are one to three favorites — tools you'd call a real game-changer for you? With the caveat that this is biased toward your specific workflow, which might differ from someone else's.


Lindsey Knake (18:03.550): Great question. I'll start with clinical tools, then get into the rest of the work we do, since a lot of us have non-clinical work too.


Ben Courchia (18:13.622): Let's just do clinical for now.


Lindsey Knake (18:17.598): One of my favorite clinical tools is called Evidently. It's not Epic — it's a third-party company with SMART on FHIR capability, meaning it lives alongside Epic. It's built off the concept of chart summarization, which, as I mentioned, Epic's own version has fallen a bit short on. What's nice about Evidently is that it lets me build all the custom prompts I want — it searches a NICU patient's entire hospital stay, and I can build custom tables. I showed some examples at Delphi — for instance, a table showing every antibiotic course a baby got over a six-month stay and the documented indication for each, pulled straight from clinician notes (assuming the notes were well-documented). Or pulling every positive blood culture, instead of me digging through all the results myself. It searches Care Everywhere, PDFs, anything in the EHR the patient might have brought from another institution. I've found it really useful for babies transferred to us after six months elsewhere — trying to get to know them, catch what I might be missing from the history. Even as humans, we forget things in handoff, like an abdominal surgery from six months ago. It's brought up things I'd forgotten or never knew.


Ben Courchia (19:55.360): It's interesting that you're going straight to a third-party module that integrates with Epic rather than a homegrown tool. This connects to something we've discussed before — the conventional note and hospital-course system doesn't lend itself well to how we think in neonatology, because babies stay in the ICU for so long. Take a baby with, say, acute kidney injury early on — we might consider that resolved after some time, and for billing purposes it would be incorrect to keep it listed as active. So it gets buried in the hospital course and drops out of the note. Same with infections — has this baby grown Klebsiella before in an endotracheal tube culture? That can get buried too, and often it's not until someone who was on service months ago says, "oh, he's grown that before," that anyone realizes it was missed among hundreds of cultures sent. The application of what you're describing seems quite powerful, and it sounds like it's worked well for doing exactly that.


Lindsey Knake (21:23.604): It has. Part of the benefit of this being a small third-party company we partnered with early, while they were still developing, is that they're very open to feedback. I can right-click on anything and flag, "this didn't give me the result I expected" or "it's not pulling in the genetics labs I need," and I'll get an email by end of day saying, "thank you, we're working on it." Sometimes it's user error on my end, sometimes it's a genuine gap they're planning to fix. I've been really happy working with them, and that responsiveness is one of the benefits of partnering with smaller third-party companies.


Ben Courchia (22:06.146): Give us one or two more.


Lindsey Knake (22:08.416): I have to plug ambient AI broadly — there are something like 90 different ambient AI companies out there. We use NABLA at Iowa; Epic is partnering with DAX. It's worth trying for cases like transport notes and consults.


Ben Courchia (22:26.200): How do you spell that — NABLA?


Lindsey Knake (22:30.974): N-A-B-L-A. It might actually be a French company, so maybe I'm mispronouncing it and you can teach me.


Ben Courchia (22:38.562): No, you're saying it pretty well. Do you have one more, or is that it?


Lindsey Knake (23:03.680): The last one isn't really AI yet, but I've done all the Epic Physician Builder courses, which is why our notes are fairly templated — I try to pull in as much automation as I can for things we need for billing. That's actually why I don't use ambient AI for my own progress notes yet — it's still faster for me to fill out my templates and keep moving. But I'm curious where AI-integrated notes will take us, because we may need to start rethinking what notes of the future should even look like. Epic's ambient technology vision is that on rounds, if I say, "we have a new diagnosis of culture-negative sepsis," it could go straight onto the problem list, or queue up orders — assuming we can get billing, compliance, and everyone else on board. That's what's going to be really exciting: easing the burden of being on the computer so the clinical workflow gets more seamless, and we can get back to the bedside, back to talking with patients, teaching our learners, and spending less time documenting.


Ben Courchia (24:06.008): That's the utopian dream for all of us. You've mentioned several tools that aren't developed by Epic itself. Do you think Epic will eventually offer a suite built from these third parties? And what's your take on the fact that most of the AI tools you've actually enjoyed and found effective aren't homegrown at Epic for neonatology?


Lindsey Knake (24:44.170): Epic's stance has generally been not to acquire companies — they try to design and develop the technology themselves, which is why they built their own chart summarization. But as you said, it's hard for a company that size to start immediately optimized for neonatology; they roll things out that work broadly for everyone. I have seen on Epic's roadmap that they plan to work on longer hospital courses and improve things specifically for neonatology, restructuring their systems to work better for us in the future. That's why being early adopters of these startup companies has been useful for us in the meantime.


Ben Courchia (25:23.180): That's helpful for people to know — for neonatology specifically, you might get more bang for your buck right now with third-party developers than with the homegrown Epic modules.


Lindsey Knake (25:38.497): And that's something hospital leadership might not even know — that these companies exist — until you bring it to them and make the case: "I know Epic might eventually build something similar, but it's not there yet, why don't we look at a demo from this other company and consider moving forward with them first?" Epic may eventually catch up, but we don't know how quickly, or what the future holds.


Ben Courchia (26:03.470): This interview is flying by, so let me get to a few more questions. One of the big worries with everything you've described is that we're always a bit reluctant to upend a workflow that already works for people. Even with the most terrible EMRs, people get accustomed to bending over backward to make them work, and become reluctant to change. Once you start adopting these tools, does it require a radical mindset shift from day one, or can it be integrated slowly, with progressive adaptation — since no team is 100% early adopters, some people are more traditional and prefer things as they are. How do you navigate that?


Lindsey Knake (27:05.344): I love that question, Ben. Part of formal clinical informatics training actually includes whole courses on change management, because we're the cheerleaders saying, "I know this is hard, I don't love change either, but it's okay because it'll hopefully be better down the road — there will be growing pains, but we're trying to get everyone on board." When I started at Iowa, one of the first things I did was overhaul the entire note-writing process and build new NICU templates. Did I create the perfect, ideal template I want for the future? No — perfection is the enemy of good, and chasing it would take forever. I did a hybrid: something similar to the old notes people were used to, so they could adjust gradually, and we keep iterating from there. I also have to shout out our APP (Advanced Practice Provider) colleagues in the NICU, who help enormously with our notes and have been very open to learning new ways of working. We've been pushing, step by step, to get off paper — historically we take information from the EMR, write it on paper, present on paper during rounds, then put it back into the EMR for notes afterward.


Ben Courchia (28:44.962): Just to make sure people understand — you're not talking about paper documentation as a system, you mean what we all do: printing the sign-out, writing on it in pen during rounds, presenting from that paper, and then — if you're unlucky enough to lose that paper between finishing rounds and reaching a computer — you're stuck reconstructing your notes and sign-out from memory. You're saying this is way too many steps, and that it could largely be avoided by moving to a more digitized process, correct?


Lindsey Knake (29:14.912): Correct. And we're still not at the ideal state, obviously — none of us are. Everyone at Iowa knows I hate paper, I don't even have a printer connected to my laptop, I actively try not to use it. But that means sometimes on rounds I'm pushing a computer around and typing, which means I'm not making as much eye contact as I'd like with my APPs and the team. I'm hopeful we'll eventually reach that better future state, but taking baby steps to reduce inefficiencies — cutting out the step of transcribing to paper and back into the EHR — helps our overall workload burden.


Ben Courchia (29:54.670): So if I understand correctly, there almost has to be a bit of an "electroshock" when this change happens — it can't really be distilled one tiny piece at a time. It requires a bit of a culture shift, in your opinion?


Lindsey Knake (30:03.434): Yes, yes.


Ben Courchia (30:24.650): In terms of this kind of integration — we've talked a lot about clinician-centric tools. What's your take on tools currently available that help us be better with families? You touched on this indirectly — compressing documentation time frees you up to spend longer in the room counseling families, which is an obvious value-add for patients. Are there specific tools you use in conversations with families — not necessarily tied to the EMR, but AI more broadly, to support families better at the bedside?


Lindsey Knake (31:10.848): Great question. I think we're going to see more and more families typing their questions and worries into large language models — ChatGPT, Claude, Gemini, whatever they prefer — and coming to us saying, "this is what it told me, this is what I'm worried about." I've seen families do versions of this for years, whether it was the neonatology Facebook group or families asking me directly, "why are you choosing this steroid for extubation?" — and honestly, that's a long, convoluted answer that nobody fully has settled, but we can talk through it together. I think it'll actually help our relationship with families — sometimes we go through what they found online together and discuss the pros and cons, and it can help explain things at their level in ways we sometimes struggle to. I also use it myself sometimes — if I'm not sure how to explain a difficult situation, you can type into a large language model (without patient specifics or identifying information) something like, "how do I describe this difficult situation to a family," or "how do I respond to this difficult email." It's helped me in a lot of scenarios where I just needed something like mentor advice, and ChatGPT sometimes provides that.


Ben Courchia (32:38.754): Do you talk with other clinicians nationally or internationally about this kind of work — comparing notes on what's working, sharing use cases? Are those conversations actually happening?


Lindsey Knake (32:59.904): Yes, in a lot of different avenues. One Epic-specific one is the Epic Neonatology Steering Board, made up of people from academic and private institutions across the US — I use that as a sounding board for how others have solved a particular neonatology problem. The other is NeoMind AI, which — you may not know this, Ben — actually got started because I was on maternity leave listening to your podcast episode with Kristin Beam and Andy Beam, thought "they're amazing, I want to be their friends," and emailed Kristin and a few others. That's genuinely how NeoMind AI got started, just from meeting and talking about shared interests. So, thank you for that.


Ben Courchia (33:48.206): I did not know that — that's very cool. Shout out to the Beams, if they're listening.


Lindsey Knake (33:56.181): That's really the great value of this podcast — it brings like-minded people together. We've built NeoMind AI into a group with specific work groups — EHR, predictive analytics, research, and others, depending on your area of interest. 


Ben Courchia (35:11.428): It sounds like, if someone is a complete novice, NeoMind AI is the best port of entry for that initial conversation and to get the ball rolling.


Lindsey Knake (35:37.138): Our monthly newsletter, which we just started, has short, digestible pearls — like "have you tried NotebookLM, here's how you can use it" — so people can just read something quick in their email without committing to a long course.


Ben Courchia (35:53.250): As a privileged insider, what's one thing coming down the pike that you're genuinely excited about?


Lindsey Knake (36:17.216): I've talked about fixing rounds and documentation — I think there's still work to do there, years of it, but I'm confident it's coming. The next big area I'm personally excited about in my research is the predictive modeling piece — using continuous vital sign monitor data. I think that's genuinely untapped data that we as humans can't fully comprehend on our own. Our nurses are excellent at picking up trends, like noticing a heart rate creeping up, but there are subtleties in that data we simply can't detect ourselves. I think the next frontier is using machine learning on continuous vital sign data to predict things like decompensation, sepsis, and extubation success — knowing when a baby is truly ready to come off the ventilator.


Ben Courchia (37:17.802): Is anyone working on that at Epic, or elsewhere that you've seen?


Lindsey Knake (37:21.756): No — this is my own personal area of interest. Epic actually asked me once whether they should build an extubation success prediction tool, and I told them I worry that hourly vital signs don't have the granular data you'd need to truly understand it — things like intermittent hypoxemia, which you can only get from continuous vital sign data, are really important insight into whether a baby is truly ready to extubate. This is a really promising area for future trainees interested in research, because more institutions are starting to invest in third-party companies to store ventilator and vital sign data long-term. Historically, that data was only kept for 30 days before being wiped. Now we have cloud-based storage and other options, but it's still an investment institutions have to decide on — purchase a third-party solution, or build the capability themselves.


Ben Courchia (38:36.588): We're wrapping up, and I think if someone just opened this episode and started listening, they might feel like the train has already passed them by — like, "my god, this is moving so fast." But it's a good reminder that we're still very early in this implementation, and what you're describing right now is really still about gaining control of the data — which is always the first step in any automation or AI development. There's still plenty of time to get in on the ground floor of this. I think that's a good note to close on, and to keep people motivated to get involved. We'll link all the resources.


Lindsey Knake (39:26.878): Before we wrap up completely, do you mind if I share a bit more, since we're talking about trainees and getting interested — some ideas for people who want formal training?


Ben Courchia (39:38.720): Yes, please, that would be very helpful.


Lindsey Knake (39:41.760): This matters for neonatologists who think, "I don't want formal training in this," because you're going to end up mentoring other people, so it's worth knowing what avenues exist. I was fortunate to end up at Vanderbilt for fellowship, which is where I discovered clinical informatics existed at all — I had no idea until I interviewed there. I chose to pursue a master's in biomedical informatics because I already had the technical background to jump into those classes. But there's also a newer clinical informatics fellowship pathway for people without a coding background — not having that technical background doesn't mean you can't do this. We just started one at the University of Iowa, and we may soon have our first combined neonatology and clinical informatics fellow. The clinical informatics fellowship is technically two years, but you can combine it with neonatology fellowship for around four years total — similar to how people add an extra year for hemodynamics training — to build in dedicated time for coding skills, learning to use large language models, or clinical informatics skills like becoming an Epic Physician Builder. I'm happy to talk to anyone interested in those formal opportunities. And additionally, for APPs, since I know a lot of APPs listen to this podcast too —


Ben Courchia (41:23.234): Before you move on — the Clinical Informatics Fellowship at the University of Iowa is housed under the Department of Anesthesia, is that correct?


Lindsey Knake (41:34.046): That's correct, because it's open to any trainee — anyone who's completed any residency can enter this fellowship, since it teaches informatics applicable to any subspecialty, whether pediatrics, neonatology, or elsewhere. That's how most universities set it up — it could be housed in pediatrics, internal medicine, anesthesia, really anywhere.


Ben Courchia (42:02.056): We'll put the link in the episode description, so people searching for it don't come up empty. I also think it's a neat opportunity, because as we discussed earlier, connecting with people from other specialties in that kind of environment is often the first step toward real collaboration on these tools. It's actually a plus that it isn't neonatal-specific — you get to engage with adult physicians, anesthesia, and beyond. You were mentioning APPs before I cut you off — sorry about that.


Lindsey Knake (42:35.655): No worries, thank you for circling back. I've actually found that the way I've gotten the most done in our NICU is by working closely with our APPs, because they understand the workflows and pain points of putting in orders and everything else in Epic. I want to give a shout-out and thank-you to our APPs, and encourage continuing to cultivate that partnership. Two APPs at Iowa, Claire and Cassie, have gone through the Epic Physician Builder training courses so they can help improve our note templates and build new dashboards and data visualizations. They've really helped push us toward getting off paper and getting the broader APP team on board with this kind of change. So to any APPs listening — there are real avenues here, and this isn't just a physician thing.


Ben Courchia (43:33.602): Love it. It's important to keep fostering that collaboration with APPs on this front rather than trying to go it alone. Lindsey, this was phenomenal — thank you so much for taking the time to share all this knowledge and experience. We'll put a lot of these links in the show description, and people can find you readily through neomindai.com or LinkedIn, since you're pretty available online. Thank you for the time.


Lindsey Knake (44:02.813): Thank you, Ben. This has been great.


Ben Courchia (44:04.876): Yeah, it was. Thank you.

 
 
 

Comments

bottom of page