By Jonathan Pabalate, DNP, CRNA, APRN | Founder, JPCAIC — JPC Anesthesia Informatics Corp | Nurse Anesthesia Faculty, University of North Florida
Miguel typed four words into Google. CRNA. Data science. Intersection. My name came back near the top of the results, and he emailed me that night, sure he wouldn’t hear back.
He’s a brand-new ICU nurse, about two months into his first job, still adjusting to the way night shifts scramble your sleep. In college he did a research internship mining outcomes data, and it lit something up in him that nursing school hadn’t touched. He wants to be a CRNA. He also can’t stop wondering whether there’s a lane, somewhere, for a nurse who thinks the way he thinks. So he searched for someone doing both, found me, and got forty-four minutes on a video call.
Most of that call was ordinary mentorship. Don’t chase the charge-nurse badge before you’ve put in the bedside years. Two years of ICU experience is the floor for CRNA school, not the target. Ask a program whether its faculty still practice clinically before you decide it’s a good one. Useful, all of it, and none of it is the reason I’m writing this.
What stayed with me was the search result.
The Search Result Was the Point
For fifteen years, every time I’ve stood up to lecture, at grand rounds, at state meetings, at national and international conferences, I’ve built the title and the slide deck around the same handful of words: CRNA, anesthesia, artificial intelligence, data science, machine learning. Not to flatter an algorithm. Because somewhere out there was a nursing student who would eventually type those words into a search bar, looking for proof that someone like him existed in this profession, and I wanted my name sitting there when he did.
Nobody handed me a title for that. I still don’t have a clean one for it. But building a search footprint on purpose, year after year, so a stranger can find you at the exact moment he needs to, is a decision. It means the real work of mentorship in this profession doesn’t start when the email arrives. It starts years earlier, with someone willing to be findable.
The Problem With the Basement
Long before Miguel wanted to talk to me about data science, I wanted to talk to computers. I was the kid in the computer lab. I nearly became a computer science engineer, until I did the math on how many hours a day that meant staring at a screen, and chose nursing instead, because it turned out working with people mattered to me more than I expected. I got both anyway, by accident. One of my early clinical rotations happened to land at a hospital installing one of the first computerized anesthesia charting systems anywhere. Nobody asked me to help configure it. I just happened to understand both halves of the problem: the anesthesia, and the way the data had to move.
That combination is rarer than it should be, and the gap it leaves isn’t harmless. A hospital’s IT department can build a technically competent system and still get anesthesia wrong, because a generic data model and a real-time, life-and-death, hands-full clinical workflow are not speaking the same language. Ash, Berg, and Coiera described this exact failure mode in a now-classic 2004 paper in the Journal of the American Medical Informatics Association: patient care information systems generate errors not because the technology is defective, but because the system’s logic doesn’t match the fluid, improvisational reality of how clinicians actually work (Ash et al.). I lived that mismatch from the clinical side for years before anyone handed me the paper that named it. Anesthesia has almost no tolerance for it. You cannot ask someone mid-induction to fight software that was never built with anesthesia in mind.
So I learned to translate. I built that into a consulting business, installing anesthesia EMRs from the ground up, on three continents, for years, doing more of that work than actual anesthesia for a stretch.
How I Talked My Way Into a Data Warehouse
That same instinct is what got me into MPOG, the Multicenter Perioperative Outcomes Group, an academic consortium started in 2008 that now draws cases from more than fifty hospitals across twenty-one states and two countries, recording upward of thirteen million anesthetics (MPOG). Nobody invited me. I called Sachin Kheterpal, the University of Michigan anesthesiologist who leads it, and told him I wanted in. He said come on board. At the time I was the only CRNA with an office in a college of medicine at UF Health @ JAX.
What MPOG was actually solving was the basement problem, at scale: making sure a heart rate meant the same thing whether it came out of Epic or Cerner or Centricity, so that data from a hundred different hospitals could be compared at all. It took years of largely manual reconciliation before machine learning touched any of it. The hard part was never the algorithm. It was getting one hospital’s version of a drug name to mean the same thing as another’s.
That’s the whole instinct, if I’m honest about it: not asking where the puck is right now, but where the ball is going to be after I hit it. Miguel got a piece of that when he asked me directly, later in the call, why I’d built any of this in the first place.
If You Can’t Measure It, You Designed It Wrong
Every student who works with me gets the same instruction before anything else: decide what the outcome looks like inside the electronic record before you build the project, and confirm you can actually pull it back out. If you can’t, the project is designed wrong. Not the analysis. The design.
That sounds obvious until you’ve watched someone spend a year on a well-intentioned quality improvement project that turns out to be unmeasurable, because nobody checked whether the field they needed was ever reliably populated in the first place. Weiskopf and Weng, in a 2013 review in the same journal, laid out the dimensions that determine whether data pulled from an EHR can actually answer the question being asked of it: completeness, correctness, concordance across sources, plausibility, and currency (Weiskopf and Weng 144). I didn’t have that vocabulary when I started teaching this. I only had years of watching well-meaning clinicians drown in data they couldn’t use. Now I have a paper to hand my students when they ask why I’m so insistent about it.
I don’t gatekeep any of it, either. When a university needs its entire curriculum cross-walked against a new accreditation standard, a job that runs several hundred hours by hand, I can do it in about six hours, and I teach anyone who asks exactly how. People assume that’s generous. It’s mostly self-interested. Whatever I hand someone today will be obsolete in six months anyway. The advantage was never the technique. It’s staying current enough to keep building the next thing, and nobody can take that away from me by teaching them what I already know.
The Piece of Paper Doesn’t Matter. The Mind Does.
Miguel’s last question was the one that mattered most. Is there a lane for a CRNA clinician-scientist in data science?
Yes. Just not the lane he probably assumed. The Doctor of Nursing Practice, the degree that trains CRNAs, is explicitly a practice-focused doctorate, not a research doctorate. The American Association of Colleges of Nursing draws that line on purpose, positioning DNP graduates to apply the science that PhD-prepared researchers produce, not to generate it themselves (AACN). You do not need a PhD to participate in data science. What I told Miguel was blunter than that: “I don’t think you need a PhD to do that. You do need to have a mentor who teaches you how to do that.”
That’s not sentiment. It’s measurable. A 2017 systematic review in the International Journal of Nursing Studies found that structured mentoring improves research career development and scholarly output among nurses, beyond what the degree alone provides (Hafsteinsdóttir et al. 21). The credential gets you in the room. The mentor is what teaches you to do something worth doing once you’re standing in it.
Anesthesia education doesn’t have a formal bridge to computer science departments. It should. What got Miguel on that call wasn’t a program. It was a Google search, an old habit of putting my name where the right people would eventually find it, and forty-four minutes I was willing to give a stranger with the nerve to ask.
That’s not a scalable system. It’s a gap. I’m one CRNA filling a very small piece of it, one cold email at a time.
I told Miguel what I’d tell anyone standing where he’s standing: the door’s open, all you have to do is reach out. I’d like that to stop being unusual enough to write about.
References:
American Association of Colleges of Nursing. “DNP Fact Sheet.” AACN, June 2025, www.aacnnursing.org/
Ash, Joan S., et al. “Some Unintended Consequences of Information Technology in Health Care: The Nature of Patient Care Information System-Related Errors.” Journal of the American Medical Informatics Association, vol. 11, no. 2, 2004, pp. 104-12, doi:10.1197/jamia.M1471.
Hafsteinsdóttir, Thóra B., et al. “Leadership Mentoring in Nursing Research, Career Development and Scholarly Productivity: A Systematic Review.” International Journal of Nursing Studies, vol. 75, 2017, pp. 21-34, doi:10.1016/j.ijnurstu.2017.07.004.
Multicenter Perioperative Outcomes Group. “Who We Are.” MPOG, mpog.org/whoweare/. Accessed 13 Sept. 2026.
Weiskopf, Nicole Gray, and Chunhua Weng. “Methods and Dimensions of Electronic Health Record Data Quality Assessment: Enabling Reuse for Clinical Research.” Journal of the American Medical Informatics Association, vol. 20, no. 1, 2013, pp. 144-51, doi:10.1136/amiajnl-2011-000681.