#033: Navigating the Complexities of Data Integrity

Liz talks with public health expert Raina Sharma, who shares her journey from the Army Public Health Command to managing data platforms during the COVID-19 pandemic for the Maryland Department of Health. They explore the critical role of data in public health, the challenges of maintaining data integrity, and ethical approaches to flawed data. Raina's extensive experience in health IT, analytics, and epidemiology offers valuable insights into ensuring accurate and reliable public health data.

Get more insights about the podcast at www.atromitosconsulting.com/atromitos-way.

We've made some resources to continue your learning. Our articles and our whitepapers give you more information behind the scenes and to learn about our guests.

This episode of Atrómitos Way is brought to you by Atrómitos, a boutique consulting firm with the imperative mission of creating healthier, more resilient, and more equitable communities. Want to know more about the firm behind the podcast? Please visit our website to learn more about who we are.

Follow us on LinkedIn: www.linkedin.com/showcase/the-atrómitos-way

Subscribe to our emails: https://atromitosconsulting.com/atromitos-way/#subscriber

I encourage all of you to reach out and get to know us!
- Liz Church, Host + Producer of The Atrómitos Way

Transcript

00;00;00;00 - 00;00;21;20
Liz Church
Welcome to the Atrómitos Way Podcast, where we have meaningful discussions on the challenges in healthcare and the solutions behind them. I am your host, Liz Church. Each episode, we dive into the complexities of our health and social system, gaining the experiences and insights of the guests that shape our lives and our communities.  

00;00;21;22 - 00;00;46;05
Liz Church
Data integrity. So earlier this season, we talked about data quality, data literacy. But we are going to dive deeper into what defines data integrity. So ask your data is the foundation of effective public health policies and interventions. Now whether intentional or accidental, data manipulation can lead to misguided policies, harm communities and cause confusion, and lead to resource misallocation.

00;00;46;08 - 00;01;15;11
Liz Church
So we're going to explore handling role data precisely, ensuring accurate connections and the ethical approaches to necessary, correct, flawed data. My guest today is Rana Sharma, a senior advisor to a journalist and data strategist with extensive experience in public health analytics, serving database management and health I.T. Raina has assisted organizations including Maryland Department of Health's contact tracing unit during the Covid 19 pandemic and managing contact tracing technology.

00;01;15;14 - 00;01;39;15
Liz Church
Her expertise in data collection, strategy analysis and product management of data platforms, combined with her background in epidemiology and analytics, provides valuable insights into the importance of data integrity and the processes necessary to ensure accurate and reliable public health data. Without further ado, Raina, welcome to the show. This is fantastic to have you here.

00;01;39;18 - 00;01;41;26
Raina Sharma, MPH
Thank you. I'm excited to be here.

00;01;41;28 - 00;01;48;11
Liz Church
So to start our conversation, can you share your journey into public health and data analytics with this?

00;01;48;14 - 00;02;08;08
Raina Sharma, MPH
Yeah, I mean, I was thinking about this question earlier, earlier this week. And I mean, I've been in public health, I think, for forever. My dad was a professor at the University of Pittsburgh School of Public Health. So, you know, I've been learning about it for forever. and that's kind of how I decided to go into it was because of him.

00;02;08;11 - 00;02;31;03
Raina Sharma, MPH
so I got my MPH and epidemiology, a while ago, and I started out my career at the Army Public Health Command, doing behavioral health work, which was like coming out of a master's program and going in the Shah was probably one of the coolest things ever. because we basically took the outbreak investigation format for infectious disease and applied it to Behavior Health.

00;02;31;06 - 00;02;54;15
Raina Sharma, MPH
so, you know, I got to travel, which, you know, at that age was like the dream. but I also got to do every type of data, collection and analysis that I think, like, you can really do. so we did surveys. we looked at all the types of administrative data we could get, which included healthcare claims, but all the drug testing and any other surveys that soldiers get during their time.

00;02;54;17 - 00;03;19;06
Raina Sharma, MPH
and we also worked with a qualitative team to get interviews and focus groups. So it was next methodology. and it was just like such a great intro to the data world. And, I think it actually made me like my favorite job to date. but, you know, once I was done with that, I didn't really know what to do, but I thought maybe getting into the IT world was kind of the next great step.

00;03;19;06 - 00;03;42;16
Raina Sharma, MPH
And so I ended up at a health ag company, who was working for or with a health information exchange. And, I've been more or less in that world, since in some way or form I will say, but I did take a short break from the public health world right before the pandemic. And I was working for a video game company.

00;03;42;23 - 00;04;04;11
Raina Sharma, MPH
But data and analytics focus the whole time. we collect the data and basically everything a player does in the game and use that to make decisions. and that was a great place to learn about distilling data and like analyzing large amounts of data and also making sure you're collecting the correct things about the players to actually answer the questions you want.

00;04;04;12 - 00;04;38;03
Raina Sharma, MPH
So I learned a lot there that I was still able to apply to public health. and then when the pandemic happened, I was epidemiologist. I just had to get back into public health. and so I came back to manage, the technology platform that the Maryland Department of Health used for contact tracing. And once again, like it was, you know, not only a platform used by contact tracers to collect information, but also like it was a data platform, it was collecting, storing and then using the data to, so once again, data was at the center of all that.

00;04;38;06 - 00;04;55;14
Raina Sharma, MPH
and when that was over, I, decided to kind of, you know, use all the skills I've built that been canceled on my own. and I focus on analyzing data and building analytics platform. So it's just been a constant. I think data is just been constant in my public health journey.

00;04;55;16 - 00;05;20;23
Liz Church
And it's amazing that you've you've gone through different avenues because, well, you're I don't know if I really need to share your resume with everybody when we go live with this, but you have it's it's amazing how you jumped from being in public health, doing different things with data analytics and measuring. And it was just and see how you did jump from public health to the video game world and then back into public health.

00;05;20;23 - 00;05;33;09
Liz Church
It was kind of like you found your calling to come back. It was like, yeah, the global pandemic was like, okay, they need me. I got to go with other superheroes and take some stuff down.

00;05;33;11 - 00;05;54;15
Raina Sharma, MPH
Yeah. You know, unfortunately, like, I feel like, as an epidemiologist, even though, you know, I never really worked on infectious diseases a lot, I feel like. But when you kind of train for, unfortunately, is something like a pandemic. And so it I just there was no way I, I couldn't not go back into it. You're right. It was really just like a calling.

00;05;54;18 - 00;06;13;22
Liz Church
Yeah, yeah. And once we find our calling then it's like, okay, we got to find our meaning to stay back. Stay in it, if that makes sense. Yeah. So I feel like we're seeing, like, with what we're seeing with a lot of people in, navigating their career paths, the workforce shortages, the problems that we're having, it is really hard to stay in it.

00;06;13;22 - 00;06;31;01
Liz Church
So I have to commend you for staying. And I know you're not, like at the frontlines as a as a practitioner or, or anything like that, but like, still there is a toll that it does take. And yeah, there's a lot of strength in being able to stay in and understanding of the work, what you're doing and why you're doing it.

00;06;31;03 - 00;06;51;17
Raina Sharma, MPH
Yeah, I know I agree with you. I think I think too, we also just assume that, like, your career trajectory is always going to be linear and you're always going to be, you know, kind of it's going to seem to make sense. but I think we also forget that we can take positions and build different skills and then come back and put them all together.

00;06;51;19 - 00;07;14;07
Raina Sharma, MPH
you know, you know, I think each position and job that we have has a lot of skills and a lot of things that we can learn from it, doesn't mean we always need to be like, you know, working kind of the corporate ladder or something like that. But we can always keep learning. And, you know, I think taking a little break and then again and, you know, remembering that like, well, I love that you games, I still do.

00;07;14;10 - 00;07;23;13
Raina Sharma, MPH
I think coming back was, was that opportunity to bring everything I've been learning into one space and be like, okay, like, this is, I think, where I'm supposed to be.

00;07;23;15 - 00;07;48;12
Liz Church
It's a very interesting dynamic because, when you speak, when you talk, when you were talking about climbing the corporate ladder, I realized, like, I didn't I didn't want to do that. I wanted to be able to do something that gives back and where I feel like my place that it's for, for the mission that I'm doing, not for the, potential growth from it, from like, based off, like income or title or prestige or anything like that.

00;07;48;12 - 00;08;06;06
Raina Sharma, MPH
So yes. Yeah. Yeah. I you know, it's funny you mentioned title. I was someone just sent a document around this morning and asking if everyone's title is correct. And like my title was totally wrong, but I was like, I don't know what it should be because I like I'm one of the few words like, I don't care about the title, I care about what I'm actually doing.

00;08;06;08 - 00;08;19;10
Raina Sharma, MPH
I think, like, you know, sometimes we do get tied up on that because there's a societal pressure and like, the contracts around, you know, what you do and your title and what that means. So it's easy to get lost in that sometimes.

00;08;19;13 - 00;08;25;18
Liz Church
Absolutely. And I agree with that. We can totally do like a segment on that if you want to. After we have this discussion, when we come back, we could talk.

00;08;25;18 - 00;08;27;28
Raina Sharma, MPH
About the difference. Didn't mean to get off track, but I.

00;08;28;01 - 00;08;49;29
Liz Church
Mean so easy to do. But it's so true. But, moving onward. And earlier this season, I had the pleasure of talking with Andrew Grover. Grover with Whitopia about the benefits of having data at your fingertips. And when you and I were talking for this covers, to build this conversation, we were talking about how accurate data is often described as the backbone of effective public health initiatives.

00;08;50;02 - 00;08;56;14
Liz Church
So, Rana, can you elaborate on why good data is crucial for public health?

00;08;56;17 - 00;09;20;17
Raina Sharma, MPH
Yeah, I mean, and I'll just say, like, good data is crucial to every industry. you know, because everybody is using data to make decisions. but I would say in particular for bringing it back to public health and health care, it's incredibly important because the decisions we're making affect patient outcomes. You know, it can affect how patients, are receiving care or, you know, maybe funding to a certain program or something like that.

00;09;20;17 - 00;09;41;14
Raina Sharma, MPH
So it's incredibly important to have good data. but I'd say before think about why it's like crucial and critical. I'd also just like take a second to define what good data is. and so, you know, like some of the things that pop into my head and this isn't an exhaustive list, I think it's just some of the things you know, to mention is, you know, is it data complete?

00;09;41;15 - 00;10;02;05
Raina Sharma, MPH
Is it consistent? and how it's collected and what it's collecting is. especially, you know, if you have the same fields from across the system, you know, maybe there's multiple, multiple hospitals sending the same fields. Are they all filling it out the same way? what does that look like? Is it accurate? Is it valid? Is it relevant?

00;10;02;07 - 00;10;28;18
Raina Sharma, MPH
and, you know, like, is it representative of the population? and is your sample large enough? So, like, I'm just thinking, you know, like, well, narrative or definitions of good. And I think, you know, that, like, really depends on what you're doing. but I would, I would just say, like, you know, keeping your mind of some of the things that make data good if we don't have good data, then we could be making conclusions about things that aren't necessarily true.

00;10;28;20 - 00;10;48;11
Raina Sharma, MPH
So, you know, for example, if I ask name biggest question in a survey, like do you always check your cell phone? I'm going to guess people are going to say no, no sometime because they're like, oh, not always on my phone. so maybe you're not really getting out what you want. So maybe the question needs to be something more along the lines of often, do you check your phone?

00;10;48;12 - 00;11;10;07
Raina Sharma, MPH
Is it every hour, every ten minutes, you know, whatever that is to really get at the question that you're you're answering. Because if you don't, then you're not going to be able to actually answer whatever the question is from the data. and so, you know, I think making sure we have the correct data is and is critical to answering the question.

00;11;10;09 - 00;11;33;27
Raina Sharma, MPH
you know, I'm just saying it's something I was working on recently. but let's say a Department of Health wants you to do a study on all the hospitals in the state, but only 60% of the hospitals are sending data. Can we actually answer what all the hospital to learn state? The answer is no. because we may be missing data from a critical hospital system, or maybe some of the rural hospitals that say we're missing all the data from them.

00;11;33;29 - 00;11;48;11
Raina Sharma, MPH
we can't really give you we can't really give a full picture of the situation. so I do think, you know, having good data ensures that we can really answer a question of hope. Really?

00;11;48;14 - 00;12;14;08
Liz Church
yeah. Yeah. And there's so many things that I think of when you were talking about how granular you are in your data collection. And, I, for a class project, when I was in college, we had to do, it was, it was for marketing to generation Z. And like how we had to frame our questions that MIT data was always getting on students saying, like, you need to me granular because everyone's going to say yes to this.

00;12;14;08 - 00;12;31;17
Liz Church
And then there's the other facet of when you're collecting data, if you make your questions very, vague, you're going to get the same answer for each one. You're not going to actually get genuine responses at that point. And, the interest is going to be there at all. So there's that as well.

00;12;31;19 - 00;12;50;25
Raina Sharma, MPH
Yeah. Yeah. I think in surveys, wording your questions is it's critical because, you know, you can also ask leading questions. Right. You can also accidentally kind of bias your question. And then you're going to bias or answer. And so you're not going to be able to use that information. So maybe something like I don't know don't you think smoking is bad for you.

00;12;50;28 - 00;13;13;11
Raina Sharma, MPH
I don't think anyone ever use that. But that's a leading question. Right. Because you're already kind of implying that smoking is bad. And so I'm like, yeah, I want you to say I want you to agree with that question. Yeah. so you're putting bias into the question. and so that's just not not going to give you the answer, but the truth of how that person feels, because they're probably going to just want to go ahead and answer it how they think you want them to answer it.

00;13;13;14 - 00;13;33;26
Liz Church
It's amazing when you look at how handling data can be done right. Handling raw data obviously can be pretty challenging, especially when ensuring it's integrity. So yeah, what are some of the biggest challenges that you face ensuring data integrity during your projects. And you don't have to name anybody in particular, but just references maybe.

00;13;33;29 - 00;14;01;27
Raina Sharma, MPH
Yeah. I mean, I think one of the biggest challenges that I face is not having control over how the data is collected or sent to me and kind of having to figure it out. But I do also enjoy the challenge of figuring it all out. but it's definitely hard when, you know, you're kind of at that space of, here's all this data, and you haven't actually been the one to, let's say, let's say like when the survey, I didn't actually get to set it up, I had no control over the questions.

00;14;01;27 - 00;14;26;15
Raina Sharma, MPH
And, you know, it sounds like it may be a kind of an intrusive questionnaire, like they're asking a lot of detailed questions, and sometimes people don't like that. So they could skip entire sections. and so I think there's sometimes some, you know, frustration with that. But I do find that if, you know, you can just be flexible, and do your best to use the data to answer the question.

00;14;26;18 - 00;14;56;10
Raina Sharma, MPH
and so, you know, like one of the examples I can think of right now, is one of the I work a lot with ETS which are admit discharge and transfer data, which are sent from health care organizations like hospitals, from their EHRs, to us. And while they're our standards, they're not always followed. but there's also, you know, maybe potential there's a difference in how data is collected, and put into the air.

00;14;56;11 - 00;15;19;13
Raina Sharma, MPH
And so, for example, is RAF an ethnicity? Ethnicity. we can assume that's always filled up by the patient. But there have been cases where, you know, like maybe like a nurse fills it out on behalf of the patient. They for whatever reason or sometimes, I mean, sometimes people hate answering questions about this. And so they may fill it out differently every time.

00;15;19;16 - 00;15;40;25
Raina Sharma, MPH
and so you could have one person with, let's say ten visits and they for each ten visits, they have a different race ethnicity that they fill out or that's been sent over. And so I, you know, how do you deal with that? because ideally you'd want to have one race and ethnicity on a patient level instead of using a different one every time.

00;15;40;27 - 00;16;02;09
Raina Sharma, MPH
and so I think, you know, you kind of have to take a step back and understand that, okay. This happened. how do I deal with it? What type of decisions do we need to make around it? is it affecting a large percent of people? Maybe this only happens 2% of the time. So it's, you know, we just need to figure out how we want to deal with it and make a decision on it and move forward.

00;16;02;12 - 00;16;21;19
Raina Sharma, MPH
But if it's affecting, I don't know, 50% of the data. And let's say most patients aren't consistently answering that question or have a consistent answer for the question. what do we do? and, you know, that's just the situation where it's out of your hands, but because you're not in control of what happens at the hospital level.

00;16;21;22 - 00;16;44;05
Raina Sharma, MPH
So it's okay. Well, what do I do next? how do I handle the data? how do I make sure, you know, I stay true to what the person put in there in the best way? But also, you know, for an analysis, you can't have ten different responses. you know, so it's just a matter of figuring out what your decision is and how you handle the how you handle the data.

00;16;44;08 - 00;16;45;14
Raina Sharma, MPH
On the analysis side.

00;16;45;16 - 00;17;10;19
Liz Church
Whenever you're looking at how data is collected, I mean, there are obviously there's so many ways that it can be done through papers or through like a third party software, like, I mean, I'm not saying everybody should be using Google Forms, but that's like a platform that comes to mind, you know, how do you navigate? I feel like that's a big challenge, navigating the data coming from different avenues or different, streams.

00;17;10;26 - 00;17;13;17
Liz Church
Now, filtering through them seems difficult to me.

00;17;13;22 - 00;17;38;11
Raina Sharma, MPH
Yeah, it is, because I think unless there is, let's say, like a mandate or maybe like a financial reason to, or some say something compelling that's going to get people, let's say, to fill in data in a standard way they may not be compelled to. And I go back to this hospital example because I think it's also really important to see it from the people that work at the hospital.

00;17;38;14 - 00;18;06;13
Raina Sharma, MPH
They're super busy. They may not have time to fill out everything in a way that someone like an analyst would like to see it, right? Like they're going from patient to patient. They have little time, you know, to fill out what they need to fill out. And so they're also doing the best that they can. so I think it's, you know, kind of this, this balance of I like going back to the hospital temple of trying to make it as easy as possible for them to fill out whatever it is that they need to.

00;18;06;15 - 00;18;35;23
Raina Sharma, MPH
but also on a level that's going to help out patient care down the line, too. so I think there's a balance, but I also do think, like, you know, in some cases, there also needs to be a compelling reason for them to do it. especially if, let's say you're sending out a survey to, I don't know, all of the public health organizations in your state, and you want to just learn a little bit about how they're using.

00;18;35;23 - 00;19;00;27
Raina Sharma, MPH
I don't know, a particular resource or like where just maybe where they're at or something like that, just getting us like an idea, like, of how they're doing. they may just not have even the resources to fill it out or just, you know, be like, who needs to fill this out? I think, like, we we forget that we may have this great idea on the research end or the data collection, but we also have to consider the people taking it.

00;19;00;29 - 00;19;22;29
Raina Sharma, MPH
and I think that's, that's like, I think a really tough part because we can give them all the information of how we'd like to see it, but are they capable of filling out the data in the way that we would like it? do they have the resources? Is it something they can even handle on a regular basis, especially if it's sending regular data or whatever project.

00;19;23;02 - 00;19;33;15
Raina Sharma, MPH
I don't know. So I think about it from that angle too. It's like not just on setting up the standards, but also setting up standards that are reasonable for the people who need to be doing the, the thing.

00;19;33;17 - 00;19;50;27
Liz Church
Doing well obviously. Yeah. Doing the thing the right way for sure. yeah. But one thing that I'm thinking of that this where it leads into kind of my, my next question on this is it's, it seems like, you know, you got to have a lot of procedures in place to ensure that that that data holds its integrity.

00;19;50;27 - 00;20;15;02
Liz Church
And it's true. So flawed data is something that it's like it's constant challenge and obviously flawed data can lead to, severe consequences. Can you discuss any ethical approaches to dealing with flawed data, like how public health professionals, I mean, and also other industries can navigate the complexities of correcting data without compromising the integrity of their analysis.

00;20;15;04 - 00;20;42;12
Raina Sharma, MPH
Yeah. So I think if we're thinking about, like, you know, at the beginning of a project, I think, or beginning of setting up whatever data structure it is, I think data governance is incredibly important. so, you know, ensuring you have processes around data and you understand what those are. you know, as an example, as an example, like let's say you're doing interviews, just thinking about the interview process, how many interviewers do you have where they all train the same way?

00;20;42;14 - 00;21;07;02
Raina Sharma, MPH
do they have really clear instructions or are they kind of just allowed to go off the cuff? because of, you know, they're allowed to go off the cuff and you just got to know that there could be some differences based on how they do that. So I think just like, you know, data governance and no matter how you're collecting data, if you're doing it, you know, in a technology, it sounds like you're collecting data from, let's say that, you know, video game example from a video game.

00;21;07;05 - 00;21;44;18
Raina Sharma, MPH
understand how you're collecting that data, what it all means, how you're storing it, how often it's spent, that sort of thing. So just making sure, like, you understand what's going on, you know, duplicates is also important missing data. and I also think cross-checking, if you can data with another source is really important as well. So going back to the hospital example, if I, you know, I'm getting data from hospitals, and let's say I'm getting like admission data, maybe double check that, you know, if I get 100 admits every day, is that matching up for that hospital?

00;21;44;18 - 00;22;01;06
Raina Sharma, MPH
Also sees on their side because maybe they're seeing 120 and it's okay. Well what happened? It was 20 people. Are we just not getting it? What, you know, is something happening during the process where it's just not maybe firing those 20 people? did the system go down for a couple hours? And that's why we didn't get done.

00;22;01;06 - 00;22;22;25
Raina Sharma, MPH
So just like, you know, I think any cross-checking you can do to make sure your data is, is matching, maybe is also really important. and then, you know, once you do understand any issues, being open and honest about it is really important. so making sure that you're documenting all that and you're sharing it with stakeholders.

00;22;22;25 - 00;22;46;08
Raina Sharma, MPH
So, you know, go back to the race example. So let's say we decide to roll it up and just say people who select multiple races, we consider them in a category of multiple races or something like that. Then we need to tell people that's what we're doing. not assume that they understand that, you know, let them know it's only happened in 1% of the people.

00;22;46;08 - 00;23;05;04
Raina Sharma, MPH
So it's a very small amount that we need to just make this decision for. But we needed to do it. so I think knowing and letting, making sure your stakeholders know what you did. So the interpretation of this is correct is also really important. I say often get a good statistician. I laugh about that, but I'm also very serious.

00;23;05;07 - 00;23;33;14
Raina Sharma, MPH
They, you know, they can't fix everything, but I definitely think they can. help apply statistical techniques to ensure your analysis is still, you know, still remains valid. So something like weighting, to account for any incorrect proportions of groups who maybe responded to something, or imputing missing data, something like that. So I do think, if that's the phase of analysis you're in and you know, your data is not looking as you expected.

00;23;33;16 - 00;23;59;02
Raina Sharma, MPH
I do think having a good decision is very important. and when you're making those conclusions about the data, make sure it matches in the data that's collected. So, you know, if you do have some issues and you can't 100% maybe that, like, answer the initial question, what part of that question can you answer? you know, if you miss maybe a portion of a sample.

00;23;59;02 - 00;24;19;13
Raina Sharma, MPH
So let's say like you are trying to sample all adult women, but you miss a group of them. Say that, you know, I think and try to understand like why. You know, why that may have happened. So like, if you had a I don't know, let's say a mobile survey that went out, women over the age of 18, but it went up to the age of 100.

00;24;19;16 - 00;24;45;11
Raina Sharma, MPH
There's a chance that, you know, maybe some of the older population who may not know what to do with a mobile survey, they may not answer it. So you may not get responses from some of the older population. So it's just the kind of knowing like why that may happen. Accounting for that, understanding it, making sure your conclusion doesn't include, people who haven't responded.

00;24;45;13 - 00;25;13;26
Raina Sharma, MPH
I would yeah. So I think and I think like going back to just when you're, thinking about the conclusions, just also note that there's limitations in the data to, and then assessing the impact of what those could be as well. so if you, you know, going back to, you know, thinking about your entire group of women that you wanted to survey, not responding.

00;25;13;29 - 00;25;33;01
Raina Sharma, MPH
Okay, well, how does that impact, not only the conclusion we can make, but like, the decision we're also going to make. so I think, you know, just taking a step back and understanding that and, and just being once again, being open, honest about it. and then I think ongoing wise, I think there's a great opportunity to learn from it.

00;25;33;03 - 00;25;42;04
Raina Sharma, MPH
And so when you're working on the next project or the next data collection, the next survey, you can kind of make those changes. So you're not maybe seeing the same issue over and over again.

00;25;42;05 - 00;25;58;07
Liz Church
It's also a defining like what's the purpose? Why's it. And so it's so important to gather this information like what are we going. It comes down to that decision making, like you said like what are we going to do with this information once we have the data collected, are we going to be able to move forward with this particular thing that we're trying to do?

00;25;58;09 - 00;26;05;28
Liz Church
Or are we just collecting data to collect data? I mean, I don't want to say that's really what's happening in, but it's just the thing.

00;26;06;01 - 00;26;33;09
Raina Sharma, MPH
Yeah. Well, and I think too, like, you know, a part of it is, you know, okay, what do you do if you're just handed a flawed data set? How do you handle it? You know, you do the best you can, but if you are able to kind of be a part of the the data collection process and setting it up, I think to the strategy around being able to collect the right data to answer the right questions is also super important.

00;26;33;11 - 00;26;49;14
Raina Sharma, MPH
because like you said, if you're not really sure what you need and you're just collecting data, you end up doing that. You're just getting data for the sake of having data. Because when it comes time to answer the question and you hand it over to the person who's going to use the data, they're like, this doesn't answer the question.

00;26;49;16 - 00;26;56;25
Raina Sharma, MPH
So I do think, like being really intentional upfront is if you're part of the process is super critical as well.

00;26;56;27 - 00;27;06;04
Liz Church
What kind of call to action would you suggest for our listeners, especially those in public health, to ensure they prioritize data integrity in their daily work?

00;27;06;06 - 00;27;27;04
Raina Sharma, MPH
Yeah, I mean, I just think it's on a daily basis, you know, it's something that you pay attention to and all this, all the work that you do, it's data integrity is like like, you know, we were just talking about part of every step of the process. It's not just when you collect the data or when you're analyzing the data, it's when you're setting up collecting the data, when you're thinking about what you need to do.

00;27;27;05 - 00;27;48;05
Raina Sharma, MPH
So I think no matter what position you're in, that has to do anything with data. So even if you're a data engineer, you're just maybe setting up pipelines or something like that, like you're like, oh, I'm not actually dealing with the analysis part of it. But even then, like, what you're doing is so important to make sure the data is coming through on a regular basis.

00;27;48;08 - 00;28;26;19
Raina Sharma, MPH
because if it isn't, do you have anything set up to notify yourself so that you can talk to the right people to say, hey, like you didn't send your feed in yesterday or today? What's going on? Is everything okay? Can you resend it, or are we just gonna have missing data for two days? so I think, like, you know, I think no matter what part of the process you're in, there's always something that you can do to make sure that you're doing the best with that part of the data process, to ensure that you're collecting the best data, you know, and really, prioritizing, prioritizing that in your work.

00;28;26;21 - 00;29;01;26
Raina Sharma, MPH
I really do think, like, you know, you never know what's going to like how what you do is going to help. I think, when I was one of my actually my first project when I was at, the video game company, I was asked to send out a daily report to track a bunch of metrics. like, including, like kind of the some, like the high level KPIs, like how many players come to the game and, you know, like, are people doing that old and kind of the high level stuff just to make sure that there were no large changes because there was a game designer who was going in and making

00;29;01;26 - 00;29;23;07
Raina Sharma, MPH
some combat changes, and so didn't want that to kill the game. and so he was like, can you just like, send out this high level report so we can make sure things are staying pretty, pretty level? and so I did. But not only did I report into being really helpful for him, it was helpful for the entire company because we were able to see, like on a daily basis what is going on.

00;29;23;09 - 00;29;42;07
Raina Sharma, MPH
because sometimes, like, there wouldn't be any data. And it was like, what? What have I up to something go down and sometimes like, they're an issue in the database. or sometimes there is an issue like on some other side or, you know, I think there are also times where if we put out a new build, there were things that maybe we didn't expect to see that change, that change.

00;29;42;08 - 00;30;08;04
Raina Sharma, MPH
So it's, you know, you just sometimes never know, like what you're working on to keep track of data, like how that, can, like, you know, just change the way that you like, track and look at things and understand, whatever it is that you're working on, I just think, like, there's so much opportunity when you're collecting analyzing data to use it to help, tell things you don't really know that are going to help you.

00;30;08;07 - 00;30;15;02
Liz Church
Do good work. Yeah. At the end of the day, it's to do good things, try to do good things with the data that you get.

00;30;15;02 - 00;30;38;16
Raina Sharma, MPH
Yeah. Yeah. I, you know, and I do it also I think being ethical about it too, I think there's so many ways to take data and manipulate it. So I think too, it's like on us, especially in public health, to make sure we're being true to what the data says and what it doesn't say. and, and, you know, sticking to that, I do think that's really important.

00;30;38;16 - 00;30;48;16
Raina Sharma, MPH
I know we saw that throughout Covid. How important it was to be to be open and honest in a stand behind it, even if it wasn't what people want to hear.

00;30;48;19 - 00;31;46;08
Liz Church
The Atrómitos Way is produced by me, Liz Church. Editorial assistance for this episode was by my fantastic team at Atrómitos. I would like to express our heartfelt appreciation to our guests who shared their expertise, stories, and insights with us on the podcast—finally, a big thank you to our listeners. Your support and engagement have meant the world to us at Atrómitos.  
 
We are a boutique consulting firm with the imperative mission of creating healthier, more resilient, more equitable communities. I encourage you to connect with us. Let’s continue these conversations and work together towards positive change. 
  
You can listen to all of our previous episodes on our website,  
atromitosconsulting.com/atromitos-way.  
That’s (a-r-t-o-m-i-t-o-s) 
 
We can also be found on Apple Podcasts, Spotify, Amazon Music, or wherever you get your podcasts.  
 
We’ll see you next time.