Alexis
Hi, everyone. Welcome to Data Dave Dives Deeper, your favorite bonus podcast, along with Talk Tech with Data Dave. I’m Alexis, your host of both podcasts, and I’m here today with Dave, Data Dave, as you know him, and Paula Reese to talk about Paula’s experience in the data world. Super pumped to have them both here on the show. But before we do any intros, as always, I want to remind our listeners that you can always reach out to us at talktech@d3clarity.com with any questions you have for Data Dave, or, or if you’d be interested in joining us on a Data Dave Dives Deeper. Okay, that was my intro. Appreciate you both. Hey, Dave. Hey, Paula. Nice to see you both today.
Data Dave
Good afternoon, Alexis, and good afternoon, Paula. Glad to have you here. Really excited to talk to you today and get into how you got here and various other things. I’m Data Dave, as Alexis just announced, and we are here to talk to Paula Reese today. So thank you very much. Thank you, Alexis. Thank you, Paula. And Paula, if you want to tell us a little bit about yourself and your journey into data, thank you so much.
Paula Reece
I appreciate both of you having me here today. I came through the data management landscape space by way of healthcare. I started out in healthcare. I’ve been in healthcare for almost 4 years. I can remember I started out probably in registration, and then I went from registration to legal, and from legal I went to something that nobody ever knows about. It’s called credentialing. And credentialing is… you kind of do your HR for physicians and mid-level professionals. You check their credentials, you make sure that they have the proper rights and they have the proper privileges to treat the type of area that they’re working in. Let’s say they’re surgery. They have training in surgery. They went to this particular school. You took their malpractice insurance, you took their background as far as where they practiced it. So, it’s a little-known area. It’s called credentialing. And I worked in credentialing for probably several years.
Data Dave
That sounds an awful lot like sort of, as you look back at that, a lot of data quality and a lot of data collection and data validation and data verification around that, which is very manual. Is this doctor who he presents himself to be? Does he have the right credentials to do what he says he’s going to do to you, and can we let him do it in this facility?
Paula Reece
You were spot on with the analogy, definitely what, you know, kind of what we did, and that kind of got me interested in data itself.
Data Dave
Okay.
Paula Reece
Lots of verification, lots of background checks. It was very manual at the time. I would say that.
Data Dave
How long ago was that? Was that five years, ten years, three years? Looking at you, two years?
Paula Reece
Yeah, it was probably ten to fifteen years because I started in the actual healthcare industry in 2012.
Alexis
Oh, wow. Okay.
Paula Reece
Yeah. So, I went from credentialing to working in supporting the back-end system of data management. So that’s kind of introducing myself to the healthcare it. I was a speed data in credentialing at this point, so I wanted to support the back end of a credentialing software for. It was a large corporate center in Michigan for hospitals, and I work for their corporate headquarters. We supported all their 92 hospitals at the time. And I worked on the back end credentialing system for this particular healthcare system.
Data Dave
So you went into it from credentialing. Did you have any technical background at that point, or was this sort of step over the threshold into the world of it? Sort of. Were you familiar with it? Were you, what was your background at that point?
Paula Reece
Yeah, I didn’t have much background itself. I did in between that time. I had graduated and got my bachelor’s degree in biochemistry.
Data Dave
Oh, biochemistry. Nice.
Paula Reece
I had some programming for biochem, very heavy in math and science. I went up to calculus three and physical chemistry two. So very heavy tech, or, let me rephrase. Not tech, but very heavy sciences and math. So.
Data Dave
So let me get this right, just to interrupt you. So, you’re working as a registration person within healthcare, working through credentialing, putting yourself through school to do a biochemistry degree.
Alexis
Just a degree in biochemistry…
Data Dave
Biochemistry degree, then into it.
Paula Reece
Yes.
Data Dave
That’s pretty good. I’m pretty impressed. That’s quite a story right there. Just working your way through college in biochemistry, and then it. Biochemistry. Obviously missing out at this point.
Paula Reece
Yes. I actually graduated. I got a job offer to work for. It wasn’t the CDC. I wanted to work for the CDC. It was. I can’t remember the name of the actual healthcare sector. It’s eluding me at the moment. But what I was supposed to do is I was supposed to work in their meth lab and… and kind of make sure… it was the border.
Alexis
I’m sorry, Paula, I have to stop you for a second. Did you say math as in mathematics lab, or did you say meth as in methamphetamine lab?
Paula Reece
Methamphetamine lab. Okay.
Alexis
You did say methamphetamine lab?
Paula Reece
Methamphetamine lab. Because it was in San Diego. I was going to work on the border and testing different things.
Alexis
Okay. Okay.
Data Dave
So not a production lab, Alexis. This is a testing lab. Not a production lab.
Paula Reece
Yes. I was going to work for the metaverse lab doing testing, but it was during the time that where the stock market crash and all that.
Alexis
Oh, right. In the Great Recession.
Paula Reece
Great Recession. And so, I didn’t know if I was going to have the money to move from Michigan to California because it was going to give me the money for that. At the same time, I got accepted into medical school. But again, I didn’t know if my family could afford it. So that’s how I got into healthcare because I needed to make more money. I was no longer the credentialing coordinator. I needed to take the next level up, and that was my next level up.
Alexis
All right.
Data Dave
Okay.
Paula Reece
Who I was going to work for?
Alexis
The DEA.
Data Dave
Yeah, DEA. So you kind of fell in love with data when you were doing credentialing, and it all just kind of worked out to your favor. And now you’re a published data author that works out really well.
Paula Reece
Find a job that’s going to pay me about the same as I would have made if I was a scientist or a physician. So I found my niche.
Alexis
That’s awesome.
Paula Reece
Yeah.
Data Dave
So now you’re clearly, I’m going to use this title “The Queen of Data Management” for pharmaceutical, because just having that biochemistry degree and working through it gives you that title. I think that’s well deserved. So, working on that data, where do you find yourself now? What kind of data are you working with? What are the problems that you face? What’s taking you through the next steps of your career from that point?
Paula Reece
Great question. So what I’m working on now is that I work for a large pharma and I pretty much work on any of their commercial data management landscape, meaning that anything that is going to help a field rep do their job, from analytics, from leads from their CRM system, from their healthcare reporting.
Data Dave
So let me stop you again.
Paula Reece
Okay.
Data Dave
Sorry, I’m interrupting you a lot.
Paula Reece
No, you’re good.
Data Dave
So now you’ve made another leap, is what I’m hearing. Because you’ve gone from clinical biochemistry type data testing, testing labs, presumably studies and that sort of thing, into the commercial side. Commercial type data supporting commercial operations versus supporting medical research or medical type testing.
Paula Reece
That is correct.
Data Dave
Did I hear that correctly? Wow. So how did you make that leap? Because that’s another leap, because you went from the scientific side to more the commercial side?
Paula Reece
You know, that’s a good question. I made the leap by way of just training through the data landscape itself. So once I was able to understand how data management worked, let me back up a little bit. So I went from the healthcare, it supporting that credentialing, back end systems to I started working for a data warehouse within that same hospital, the healthcare system in Michigan. So, I started their project. They were setting up a data, a patient data warehouse. Of course, I knew the patient information because I worked for healthcare, right? So, I knew how to master provider data.
Data Dave
And you know, all the providers personally because you just created it.
Paula Reece
I knew everything about provider data. So, they needed somebody to help them set up this data warehouse, and they needed to meet because they didn’t know anything about the data itself. You teach me about data warehousing, I’ll teach you about the data. And that was a relationship that we had. I learned SQL querying within that transaction for that particular position. And from there, once I learned sort of how a data warehouse worked, how SDLC works, then I was able to take that and kind of translate it to any space that I went into the commercial pharmaceutical.
Alexis
I gotta ask about SDLC.
Paula Reece
Yeah.
Alexis
What is that?
Paula Reece
It’s a, it’s a standard way that you move data into your system. From concept to adoption within a data landscape space. It is really just a methodology.
Alexis
Okay.
Paula Reece
It’s a fancy term that we like to see when you come to data management. We have a acronym for everything, which is a fancy way of saying methodology of adapting data into your system.
Alexis
Okay, so you just ended at, you’ve learned about a data warehouse, and now you can apply it to anything. And then I stopped you. Please finish your story, because I’m really.
Paula Reece
So that’s kind of how I make the leap from, you know, healthcare, hospital to pharmaceutical. And I’ve been working in that ever since. But here recently, I transitioned into a new position to where I’m learning about innovation and I’m learning about how to innovate within a therapeutic area, within pharma, particularly the one that I’m supporting right now. It’s plasma derived therapies. And, you know, that’s actually, you asked me about the successes and were the challenges. I see now with the introduction of AI and machine learning, I’m finding that how we do data management is going to change. Yes. And in order to do that, we’re going to have to learn how to innovate. So this new position that I kind of been working in is a blessing in disguise. And one of the things that I think I’m going to have not, I think one of the things that I actually have put into the works is that I’m going to have to go back to school and get a master’s degree in data science, human design or emerging technologies that I’m playing with.
Data Dave
Okay. So I’ve got a lot of questions buried in the last little piece. So the first one, I’m going to take you back a little bit. So you went from sort of the biochemistry, mathematical engineering side of data into the commercial side of data, into data warehousing. You made this trade where you would teach them about the data and they would teach you about the technology of managing data, which I think is excellent.
Paula Reece
Yeah.
Data Dave
I’m going to ask you, how necessary is it in data management to have domain knowledge of the data? You’re managing everything.
Paula Reece
It makes you be more well rounded. It helps you to facilitate the conversation. When you’re in data management, you literally are liaison between the business and it. And you really, in order to be a great data manager, you have to be able to speak both languages. You have to be able to help the business person understand what they need. Then you have to be able to translate what the business need into what it is going to develop and design. So you have to.
Data Dave
So I couldn’t agree with you more. I think it’s absolutely vital to really understand what is in the data and to have people that can explain the data to you and walk through the data and get that intuitive knowledge of the data to know that the data is describing what it’s supposed to be describing.
Paula Reece
Correct.
Data Dave
Because I’m sure that you can look at from your days credentialing, I’m sure you can look at a provider and in a heartbeat you can say whether he’s true or false, real or whatever, just sort of.
Paula Reece
Yeah, you can.
Data Dave
Right. Because you know what the data is trying to describe and you know the context of what it’s doing. Correct?
Paula Reece
Yeah, you’re 100% correct. I can look at a data set and immediately tell you what’s off in that data set.
Data Dave
Yeah.
Paula Reece
How do you do that? I’m like, because you train your brain to quickly scan something. That’s the one thing that’s off.
Data Dave
People ask me the same. People ask me the same thing. I look at a tremendous amount of datasets every day, and people say, how did you find that? Yeah, it was obvious. It was just jumped out at me. I ran a sequel query and it just jumped out at me that that was wrong.
Paula Reece
Yes, because we’ve trained our brain to look for the one anomaly. Everything else is consistent, but that one anomaly screams at us, and like, that’s it right there.
Data Dave
The next follow-up question to that is you’ve got this great history, and now you’re moving into the innovation space and how to bring AI and ML to therapeutic studies and to therapeutics within a pharmaceutical organization. And you made a really provocative statement, which you said, data management is changing and it has to change in order to support this new world. I want you to drill into that. And don’t worry about how deep you go; I just want you to drill into that because that’s a really provocative and interesting statement.
Paula Reece
Oh, for sure. I manage a large data management team. Okay, 20 plus data stewards. Now, how efficient would we be if we can run a algorithm in the background to understand how many times we’re touching a file? Many times we’re touching the same file? How do we correct the data that we are bringing in so that we don’t have to continue to correct it? And how do we publish our data out of our system efficiently? Okay, so if I just worked on those four answers, think about the number of stewards I can reduce on my team just immediately after that, right?
Data Dave
Absolutely. Just the mechanism of data steward. But there’s still that intuition that the data steward has that you came into it from credentialing, there’s still that intuition that you have that says, this smells wrong, this isn’t quite right. Do you think ML and ML algorithms can take on some of that? Are there enough patterns there that you can train an algorithm around that you can?
Paula Reece
So, what I think the, you know, the AI and ML is going to do is going to help the data steward do their job faster and more efficient. Whereas I may have needed five data stewards, they can run that program and now I need two or three.
Data Dave
Why?
Paula Reece
Because they can quickly, they can use AI to understand a pattern quickly, quickly, but they still have to go in and change that data set. I can’t rely on a computer to do that for me.
Data Dave
Yep. No, I agree with that.
Alexis
The upfront efficiency is going to be more, just, just a more efficient overall.
Data Dave
You want the final word to be the data steward, but you want the data steward to be informed by the machine, learning to be able to say the machine thinks that this is true. Now you’ve got to correct and validate.
Paula Reece
Correct or of the data that we have potentially profiled from X source, what is the computer telling us that we need to go in and pay attention to, we no longer have to go line by line. We no longer have to run various SQL queries to pull back the data. AI is going to kick that output for us themselves.
Data Dave
Right. So let me ask you another provocative question based off of that. Are you generally excited by AI and ML or intimidated by it or scared by it?
Paula Reece
You know, I’m probably the one person that’s very excited about it. AI and, you know, ML, everybody’s sort of scared about, oh, what it’s going to do, what it’s going to do. I just think it’s going to make us be more efficient. I think it’s something we should embrace.
Data Dave
Yep.
Paula Reece
Because it’s not going to replace what we’re doing. It’s just going to aid us in what we’re doing every day. It’s going to help reduce that manual error, but also help us quickly identify errors themselves. We still need to do our jobs. We just will do our job a little bit more efficiently.
Data Dave
I couldn’t agree more.
Alexis
Yeah, I’m with you 100%, Paula. I don’t work in data. I work on the soft side of d three clarity, hence why I post a podcast.
Data Dave
But she does work in beta.
Alexis
Okay. Dave thinks I do. Okay. So I, but I do. I work on the soft side of it, and I use AI now to help me write better emails. After editing the podcast, I have it transcribe it for me, and then I have to go do a manual check. But I know what I’m looking for, and it’s much easier than listening to it and typing it all out, which I did one time and I will never do again. Yes, I have those opportunities to use AI right now. That a year ago when I didn’t have Chad GPT at my fingertips, I didn’t have an opportunity to do, and my efficiency has shot up because I’ve been able to kind of embrace those opportunities. Now, one of the things that we just launched big chat GPT with our whole team, and one of the things that we reminded them was that it’s going to remember everything you put in it.
Paula Reece
Yeah.
Alexis
So be mindful of what you put in it. How are you fighting that within your organization as you’re starting to embrace AI and machine learning more and more?
Paula Reece
It’s like any other data management system. If you put garbage in, getting garbage out. We say that in data management all the time. Garbage in, garbage out. So you have to make sure whatever you’re feeding that computer, who’s going to learn from you? You’re feeding it good information. So we still have to do our work upfront to make sure that we get the best out of AI and machine learning.
Data Dave
So I couldn’t agree with that more. You’re absolutely speaking my language. Now, the way I look at it, Alexis heard me say this a number of times, right? I look at it and say data is the evidence of history; it’s what we’ve done. And then mathematics is the language of prediction.
Paula Reece
Yes.
Data Dave
So now you’re starting to predict it. You’re starting to predict things. What I find absolutely fascinating is applying those two rules to unusual use cases. Yours I find actually more fascinating because now we’re talking about how we capture the data surrounding the actions of stewards in a way that we can analyze it from the features of those decisions. We can track the history of those decisions and build the mathematical patterning to build a model to predict the answer going forward.
Paula Reece
Perfect. Yes.
Data Dave
Now you’ve got something that you can glean the history from, and you’re in a great place because you’re actually putting humans in there. You’re not expecting it to come up with an answer itself because the answer is actually really, really important. So you put in humans in there to verify and actually make the change. Now, what you’re doing is verifying every answer that the ML gives you, which makes your training even better because every answer you verify becomes a training set to feedback in. So you’ve accelerated your machine learning.
Paula Reece
Exactly. It’s… think about, you know, our data quality is going to be so much better after we apply this learning.
Data Dave
Yes.
Paula Reece
If I have a million data sets in my system, only maybe 10% is a problem. Well, if I’m constantly correcting that 10% because I can’t see the 10% that I’m correcting. What happens when I correct that 10%? Then, I can add more data to the system without adding more problems. That’s going to mess up what I’ve been fixing all the time.
Data Dave
You can go back to the source and stop those recurring problems happening and you can start to do a lot of sort of analytics and research behind that to train these models and get to a really exciting place. So, I was just talking to a friend of mine the other day, this is another exciting one that I had, and my friend, she’s an archaeologist, and she’s been digging up copper plates for years, her whole career, and she says, I’ve got this theory that my copper plates aren’t randomly placed. I said, okay, that’s interesting. So, what do you do? She said, well, I just dig random holes. I said, what if you could give me all the data that describes where your copper plates are and let me build a mathematical model that can predict where to find the next one?
Paula Reece
Yes.
Data Dave
Now we can start asking, why were people burying copper now? They weren’t probably burying them, they were just dropping them or whatever. But was this a migration path? What was happening? What was causing plates to be in this pattern? And that’s just another exciting, random use of the same kind of machine learning path that I find really exciting. And that’s one of the things we do at D3Clarity. Help with people preparing their data, cleaning the data, and getting their data ready for feeding into these kinds of models and then helping build them, models, etcetera. That, to me, is just fun and exciting. Anyway, back to your story.
Paula Reece
Yeah, no, I agree with you. It really is because we can now unlock patterns that we just couldn’t do as humans. Right. There was that piece that we just could never tap into because we didn’t have the capabilities to do so. AI and machine learning and various models that we’re building can do that because they can process these large data sets for us that we can never process. Because, again, we as humans have limitations, which is why, you know, AI and machine learning are an aid; it’s not a replacement.
Data Dave
I couldn’t agree more. It’s another tool in our toolkit for want of a better friend. One of the things that frustrates me more than anything else is people working on the problem. Rather than solving the problem, I’m going to throw resources at it and add another ten people to just work more on the problem.
Paula Reece
Right.
Data Dave
Why don’t we step back and solve the problem rather than work on the problem? That would be more intelligent here. And I think we can use AI in spades and ML in spades around this kind of thing for data stewards and other areas where there are endemic patterns. Because there are always patterns in data.
Paula Reece
There are always patterns. And I like your copper plate story because I’m pretty sure once you run that algorithm, you’re going to know that it’s not random at all; there’s some sort of pattern that you can follow. Again, it’s going to make her more efficient when she goes out to dig up those copper plates.
Data Dave
Right? Exactly, exactly.
Alexis
Paula, I’m sure that no matter what path you take when you go back to graduate school, you’re going to be successful. You have so much, like excitement and passion about what you talk about, no matter what you do. I know it’s going to be amazing. So, this has been such a good conversation. This has been great. I loved hearing all these perspectives and all these ideas.
Paula Reece
I appreciate you guys having me. It’s been a pleasure. Truly. It’s my pleasure. I appreciate both of you. Thanks Data Dave. Thanks Alexis.
Data Dave
Thank you.
Paula Reece
And I look forward to talking to you guys soon.
Alexis
Thanks, Paula, for being with us. If anyone else, like I said at the beginning of the show, is interested in joining us for a Data Dave Dives Deeper, just reach out to us at talktech@d3clarity.com. Thank you, Data Dave. Thank you, Paula, for being with me today. I really appreciated our time.
Paula Reece
All right, have a great one.
Data Dave
And thank you, Alexis. And thank you, Paula. Thank you.
Paula Reece
Absolutely. Have a good one, Dave.
