 Good afternoon. I'd like to talk to you today about Project NEL. A Tool for Educating Nurse Scientists for the Age of Information. My name is Vicky Hertzberg. I'm a statistician and I direct the Center for Data Science at the Nell Hodson Woodruff School of Nursing of Emory University in Atlanta, Georgia. I came to the School of Nursing in 2015. And a few months after I got here, there was a story on NPR about medical students at NYU using big data sources to better understand healthcare trends. And I thought if medical students can do this, why not nurses? So today I'm going to talk about why we're doing Project NEL, what it is, what our history is, the challenges, the usage to date, and what we see in our future. Well, it seems to me, not let me advance. In full disclosure, I have part of the ownership of Project NEL, intellectual property. In the Center for Data Science, we firmly believe that if you can change nursing, you can change healthcare. And we're all about developing the tools and the talent to help change nursing. Nurses are the largest segment of the healthcare workforce and they spend the most time with patients. And that's why we believe that if we can change nursing, we can change healthcare. And if you want an example, I'll just say Florence Nightingale, a nurse who changed healthcare, also a statistician. So how can we support nurses in changing healthcare? We can help them in implementing evidence-based practices as they accumulate. We can help them develop evidence-based practices where none exists. And we can enable nurses to develop data-driven evidence-based practices. And that's where Project NEL comes in. Yes? We're actually seeing your presenter view. Can you hit that swap displays button up there in the left-hand corner? Oh my goodness. I'm sorry. No worries. Perfect. Thank you so much. It's my camera's on one screen, my presenter views on the other. So my apologies. So we want to put nurses on the data information knowledge wisdom spectrum, where we're going from data, which is the naming, collecting and organize of simple measurements of objects, and to information where we organize and interpret them to knowledge, where we interpret and integrate and understand to wisdom where we understand we apply. And in nursing, we apply with compassion. The framework was developed by Doctors Patricia Brennan and Sue Bakken. Dr. Brennan is the current director of the National Library of Medicine, and Dr. Bakken is the editor-in-chief of the Germany American Medical Informatics Association, and both are renowned data science, renowned nurse scientists. To ask the question, what happens? What should a nurse do when they have a hunch? And that's where Project NEL comes in. Project NEL stands for Nursing Electronic Learning Laboratory, and it's comprised of the de-identified electronic health records taken from our clinical data warehouse for a randomly selected sample of every healthcare patients across the care continuum, so we get both outpatient and inpatient. This was done across four hospitals. We've recently expanded to more hospitals, and we'll be getting those in too. We started off in response to this NPR story. Let's get a small data set to help nursing students learn how to use an EHR data set to support their hunches in a clinical setting. And what is that involved? Well, we need to get the patients and their demographics, but then we need to categorize all of their encounters with every healthcare. And so every records indexed by the patient ID, the encounter ID, and those encounters comprise every hospital and clinic visit, the diagnostic codes, the administer medications, procedures, orders, medication and medical history, lab values, and other test measurements. We started off with 20,000 patients, which in nursing is an incredibly large data set. We were able to accomplish this easily using an access database with an Oracle back end. And we accomplished this so easily, and we thought, well, that was so easy. Let's go to 100,000 patients. And that necessitated a change in our back end to Postgres SQL and necessitated a change in our front end as well. And I'll talk about that in a minute. And that took a little longer. And then we were challenged to go to a million patients, necessitating another change in our back end. And so we're now with the MongoDB NoSQL back end. And we've tagged all of our variables to an OMOP standard. And we're really glad we did for another reason I'll talk about in a minute. So we developed our database, and it's now residing on the Amazon Cloud. Project Now Today is in what I call the version one stage, comprised of over a million patients. It's all electronic health records for these patients going back to 2012, up to about 2020. And right now we're adding in the records from 2021. And we'll soon be tackling 2022 as well. But that's kind of another wrinkle. The data have been thoroughly cleaned. And there's a data dictionary that we've developed that even every healthcare doesn't have. The database also includes clinical notes. Our front end is shiny, hence my presentation here. And it's a shiny amp with a series of drop down menus that allowed nurses or other users to select what patient populations, what tables they want to see, and has various blow up points where they can do finer selections. Nella's also comprises short courses. About quantitative tools for large data sources that are available through the every nursing experience, our continuing education arm as online courses. We've been distributing it to other schools of nursing for use in courses for doctoral students, both PhD and DMP. And we're also developing a series of case studies about its use so that we can use it in our curricula at all levels, which will help us. When we undergo re accreditation because we've got new accreditation standards involving increased attention to data and information technology tools. We've got had a few PhD dissertations that have come out of this. A couple dozen DMP quality improvement projects that's their capstone experience in that program. Several students in our BSN honors program have based their thesis on data from project now. First, we've had a variety of faculty research project and presentations that have arisen from using the data available and now our current challenge is that in October 1st of 2022, every healthcare changed electronic health vendor from Surner to Epic. Totally different system. So we're looking to that to change that that's being tagged to the OMAP standard so we can hopefully integrate that seamlessly into version into now the operative words being hopefully and seamlessly looking to incorporate data from Children's Healthcare of Atlanta. Every health care is exclusively adult and pediatric data is a real missing element, especially for nursing. And so we really want to get that in. And we also want to integrate data from the Grady Health System that being the public medical system of Fulton and cap counties where every is located. Those are both on Epic. And so it's really critical that we get the integration of every health care data into our system right before we go about seeking data from these other sources. With this, I'd like to leave you with this thought. When things go wrong because they sometimes will. When the funds are low and the debts are high, and you want to smile, we have to sigh. When care is pressing you down a bit, rest if you must, but don't you quit. Life is queer with its twists and turns. As every one of us sometimes learns. And many a failure turns about when he might have one feet stuck it out. Don't give up to the base seems slow. You might succeed with another blow. Often the struggle has given up. When he might have captured the victor's cup. And he learned too late and the night slipped down. How close it was to the golden crown. Success is failure turned inside out the silver tint of clouds of doubt. And you never can tell how close you are, maybe near and it seems a farm. So stick to the fight when your heart is hit. It's when things seem worse, you mustn't quit. And with that, since I'm the only one standing between you and your dinner, I'd like to close and thank you for your attention and ask what questions you have. That was an awesome talk, Dr. Hitchcock. Thank you so much. I think Dr. Higgins have one question, just around the size of the data saying that ended up being larger than RAM data. I mean, I don't have any issue with R and I was just wondering as well. I'm guessing guys maybe have like a positive connect instance. I know that you're, it's in the cloud. So you haven't had any issues with that. We haven't had any issues that I know of, but then I am, I mean, that's, I don't get down in the weeds with this. That's why I have a team that does this. So, So I'm sure it probably does, but not at this point. There's a question about OMOP. Supposedly the OMOP mapping should help with the EHR change, because instead of going, you know, variable A over here to variable C over here, it's going label to label. Label here, label there, put them together. That's my understanding of it at a very simplistic level. We'll really see what happens when we start to tackle that. We've recently introduced clinical notes into this, which is a really valuable resource. And that took a long time to happen. It's been a real process because it has to be de-identified. Even though the record took de-identified, you know, then there's still stuff that pops up in notes that is identifiable. So we had to de-identify that. But then we had to make sure that it's not over-reacted because there are things like walker or Simpson forceps that might be considered identifiable data, but they're actually medical terms as well. Yeah, do you guys have like a natural language processing team that's helping with that or how are you guys attacking that? That's being done by an entirely different group at our institution. And so we just, there's a lot of politics involved in this because we're outside of what's considered the covered entity. And so we have to go to the trusted whatever and they do all of that for us. And then we get the data and then, you know, we started this in 2017. And by 2018, 2019 was when we got the challenge to go to a million. And it hasn't been until recently that we've really maybe the last year that we've really gotten to that million because of first with the pandemic and the same team had to build all of these dashboards and everything. And then we then the same team got charged with the epic conversion. So it's been a real journey. Yep. I bet. Okay. Well, thank you so much. And I think I see any other questions. I really appreciate that was a great talk.