This transcript is auto-generated and may contain inaccuracies.

Santiago Gonzales Irizarry:
All right, good afternoon, everyone. If you would like to get a little bit closer, I don’t think we’re going to have a full room today, considering that we’re competing with AI. But guess what? You need to have strong governance for the successful implementation of AI. So I’m really happy that you’re all here. Good afternoon again. My name is Santiago Gonzales Irizarry. I’m the data modernization lead at Washington, DC, and I’m very excited to be participating in this panel.

Today, we’re going to be tackling operationalizing governance frameworks, and for that, let me go ahead and introduce our distinguished panel. We have Chris Baumgartner. Chris serves as Deputy Chief Informatics Officer for the Washington State Department of Health, where he leads coordination of data exchange through the State Health Information Exchange System and oversees innovative interoperability projects.

From Kentucky, we have Brittany Saltzman Bell. Britney is the director of data modernization initiatives at the Kentucky Department of Public Health, bringing experience as both an environmental epidemiologist and COVID-19 reporting lead. She currently oversees PHIG A3 and agency-wide modernization projects. Then we have from Virginia, Anup Shrikumar. Anup directs the Center for Public Health Informatics at the Virginia Department of Health, bringing over 20 years of experience in using data to solve complex public health exchanges, public health challenges and develop innovative analytics solutions. So please help me give a warm welcome to our panel.

Before we start today’s agenda, we are going to be focusing on data governance in public health, and we’re going to tackle different experiences from three states in how they operationalize data governance and how, from different models, they’re achieving successful data governance in their public health agencies. I have a couple of questions that I would like to ask them at the end, but I will encourage all of you to please get your questions ready, because I think that at the end of the discussion, we’re going to have some ideas and also share some insights on how to bring actionable data governance.

As you know, data governance feels abstract, something we know, but sometimes struggle to implement practically. Today, we’re moving from theory to real-world examples from three states that have successfully operationalized governance frameworks. My goal is that at the end of this session, all of you feel like you get a taste of how to implement better governance in different settings, how jurisdictions are advancing enterprise-wide governance frameworks, develop key relationships, strategies for establishing governance, and real successes and challenges implementing governance from different state agency contexts.

And before we begin our panel discussion, to do some level heading, let’s explore the basics of data governance. When we first met to plan this session, we wanted to narrow down what data governance really is. What does that mean? And we all agreed that data is an asset. This is the most powerful tool that we use every day, and because it is an asset, we need to have a governance structure around it so that we can be effective. So data governance is a framework with rules, policies, and processes that guide how data is managed, shared, and used to ensure data quality, access, security, and compliance, and help us be more effective in improving decision-making, reducing duplication, and meeting program goals. As I said, establishes data as a powerful and strategic asset that informs decisions and demonstrates impact that drives innovation.

And when you begin to think about your governance bodies, there are different roles and responsibilities, beginning from the data owners. These are our team members who are accountable for data accuracy and compliance. They’re often program leads, and then the data stewards. They maximize the value of data by enabling its use, improving usability, reducing costs, and managing risk, often day to day. So these are the experts, the people who know that they have the best and the ones who know when something looks a little bit off. So they do play a very important role in our data governance.

It’s important to remember that there’s a critical distinction between data ownership and data stewardship. They cannot mean the same thing. They have specific purposes. Many of you already have these roles informally, but naming and formalizing who’s doing what clarifies and provides accountability and ultimately reduces silos. Then you have technical owners who manage big systems where the data lives, that could be EHRs, IIS systems, or surveillance systems. And finally, the enterprise governance body is a cross-functional group that makes decisions, improves standards, and resolves issues.

So those are the parts of the leadership that ultimately, if something is not granted, it’s not working our systems; they should be able to provide an answer. In thinking about the key elements of the governance framework, these four pillars supplement each other because they’re so closely related policies and standards, they need to be defined, with clear guidance on data definitions, access, privacy and quality, then processes and workflows, documented procedures of data entry, review, corrections and issue resolution, including continuous process and improvements. Tools and systems on the technical side share platforms for metadata, quality checks, and collaboration. And finally, I think one that is sometimes overlooked but very important: communication and training, regular stakeholder engagement, and capacity building.

One of the concerns we hear over and over is the lack of change management or clear communication with our teams. So all these are part of our key elements of the governance framework. What are the goals? Data quality and interoperability. Data Governance identifies a result root causes of poor for of poor, data quality, standardizes data definitions and coding, enables trustworthy data exchanges across systems, and internal interpretability across data sources. IT supports something that I’ve been hearing more often recently: TEFCA, ELR, or any other data-sharing standards, like FIRe implementation. So it’s critical that we keep governance to make sure that these goals are achieved.

And now is the part where we begin to put all of these concepts into action. I think it’s important to normalize data governance. The way to do that is by including it at the beginning of any of the data project life cycles, not at the end or in the middle, which could delay the implementation of a project or cause issues with vendors and internal teams. So, having that feedback loop present the whole time is important, assigning roles in kickoff meetings and decision logs, so there’s good documentation, and including data governance activities in work plans. It’s important to remember that data governance is not just about. Control is about bringing control in the midst of something that could turn very easily into chaos. So by having your data governance defined under this framework that works for your own agency, you get closer to being successful at making sure that those data projects can be implemented.

All right, so we have a little exercise. Wanted to do a poll with all of you, looking at the positive first. So please take this QR code poll and share with us, what governance areas do you feel like you have the most success? We’re going to give it a minute to allow everyone to give their responses. So far, strategy and leadership are very high, and stakeholder engagement is also very high. Excellent, so far, we’ve been having a lot of success in building strategy and leadership closely after that stakeholder engagement.

Now let’s get real. What governance areas do you feel like you have the most challenges? Very clear that roles and responsibilities are a challenge. All right. Thank you so much for participating in the poll. I think this is very important because that way we can begin to channel the conversation and pick up some strategies to begin to think differently about data governance. Let’s go ahead and begin our discussion. We’re going to hear from three states that have put these concepts into practice. Each speaker will share a real-world experience in about 10 to 12 minutes, and then they will focus on implementation, practical implementation relationships, and lessons learned. We are going to start with Chris from Washington state, then Brittany from Kentucky, and finally, Anup from Virginia.

Chris Baumgartner:
All right, thank you. I’d like to talk a little bit today about our actual implementation center project, something we call tracks, the transformational repository and analytics exchange. The idea started actually, as what we affectionately called a spark tank. It was based off of the show Shark Tank, but instead, public health folks within our agency could bring ideas to the executive leadership team, innovative ideas that they wanted to see move forward. One of them was the idea of modernizing chronic disease surveillance. And so this actually was chosen in our first round of the Spark Tank, and we were given some initial resources to explore this idea. Where we went with it was kind of innovative in the sense that, unlike infectious disease and chronic disease, we do not have any mandate in our state to report things like hypertension or diabetes, and so we really approached our healthcare payers and providers with a different kind of approach to governance in this space.

This wasn’t a system that the Department of Health owned or operated, or that we just had people dump data into, and it disappeared, and we did whatever we wanted with it. This would be a shared environment. And so to really do that, we needed to think about how to create shared governance externally, with payers and providers and other partners within our space? And so that was something we have been working on for some time now. And overall, this is our business context act. Gram, or our idea around how we’re going to operationalize this, a lot of the work started out of the multi-state EHR network for disease surveillance. There are a few states participating in that. It’s a way that CDC has been looking to modernize chronic disease surveillance, moving away from things like burfish surveys and towards the use of electronic health record systems. And so we’ve been participating in that, and our IC project is hoping to kind of build on that.

So you can see we have sponsorship, we have oversight, data contributors, partnerships, our users, and our initiatives that are all feeding into this concept that we’re moving forward right now. We have two tracks within tracks, one is the actual proof of concept, technology, and exchange, and the other is the formation of this unique governance between public health and our private partners to develop a shared space where we can all leverage this for our different use cases. Just real quickly, our proof of concept, what we’re doing with the Implementation Center Project, is kind of exciting. It’s reusing the ECR now fire app for chronic disease surveillance. It has been reprogrammed to trigger on hypertension and diabetes, but it produces the electronic clinical quality measures that our healthcare partner needs to send to CMS. The underlying data for the numerator and denominator can come into our platform for chronic disease surveillance.

And so at the end of the day, the app kind of does three things for them. It produces eCQM. It can still report notifiable conditions for infectious disease, but it can also provide the underlying line-level data we need for chronic disease surveillance. So my joke is we don’t have a stick in this case, we needed more than a carrot, we’re trying to create carrot cake, my preference was cream cheese frosting. So this is our concept for how this is going to work. Multicare is a health care organization in our state. They’ve just installed and configured the modified app, and we’re hoping soon to see if they can produce the data payloads we’re interested in. And we’re working through E-health exchange and our health information exchange to create this new shared, again platform with our partners.

When we worked with our partners, what we really ended up looking at was the Civitas network for Health’s model for a health data utility, and this is their minimum necessary use cases. And this is what really, I think, intrigued us about what we could do again with this platform. Because while chronic disease prevalence is really the primary use case we care about in public health, there’s a lot of other use cases our payer and provider partners are very interested in, as well, around Value Based Payment and care models, knowing how to best distribute their limited resources, knowing where they need to position, maybe new clinics or other resources within their spaces. And so we’re really, again, looking at the minimum use cases there in that space.

So some of our key governance features are really, again, having high-level organizational leadership buy-in. We are very fortunate because we had that from our executive leadership team. But we also have, again, this entity in our state called the health care forum, which is made up of all the large primary health care providers and payers in our state. And we were able to get buy-in from their leaders as well. Again, because the idea is that this is not owned or operated by the Department of Health, it was also very key to bring in a neutral, non-governmental facilitator for this work. And so for that, we have partnered with the Foundation for Healthcare Quality. They do a ton of work in this space already in our state, in helping facilitate and provide a kind of neutral facilitation between public and private, as well as payers and healthcare. And so they’ve been partnering with us on creating this in this space. And so again, this came through our Spark Tank Project, and then we had the idea to have a shared ownership of this data utility.

So, as you can imagine, the governance for this is really critical. Data submitters and users have a say in how their data is governed. It’s a neutral convening. There’s no mandate. There’s no requirement to report that the use cases are aligned. We feel well with Tefca and the Civitas HDU concept, and we’ve developed many types of agreements that work together for this purpose. We’ve already developed our executive governance charter, and what we’re focused on next is an MOU that would be between a data contributor and the governance team. So that’s kind of what we’re working through right now. We have a draft out that multi care again, our pilot partner is using and working with our legal team on. We also hope to re-leverage the trading partner agreement that they already have in place with one health port, as well as the DURSAs that they have in place with the health exchange, to facilitate some of this as well.

So the MOU, again, serves as an agreement, specifically between two or more parties. It’s not legally binding, but it’s really between the executive governing team. And the trading partner, these are not mutually exclusive, and it really is. The idea is to try not to set specific standards for interoperability, but at least talk about what types of standards are acceptable. And we’re trying to really reduce administrative burden by relying on, again, standards that they already have to have in place for interoperability and trying to reuse data transport that’s already being used as well. And so it really talks about the responsibilities of the governing committee as well as the participating organization.

Where we’re trying to go next is kind of the other end of the spectrum, which is how we handle new use cases, and so we’ve developed a Data Use Case request form, and we’re actually piloting that as well with the University of Washington. We’ve partnered through our IC program with the School of Public Health to actually evaluate what we get out of this. So the goal is to see what tracks produce using the ECR now fire payload, compared to men’s, compared to burfish, for example. And trying to understand, do we feel like we’re getting a better sense of chronic disease surveillance and prevalence through the use of this method? So we don’t want to just do something cool and say, Wow, we used fire. That was neat. We actually want to see did it produced a better outcome for chronic disease surveillance.

And so at our upcoming meeting, actually just next, this next week, we will be reviewing the data requests from the University of Washington School of Public Health and hopefully formalizing at least our initial request form. And again, there must be a unanimous approval from those members of the committee, and this will be ongoing, so if another member of the committee or a third party approaches the executive governing team and wants to leverage the shared analytic platforms, data repository for something. That will be the framework that we use in place. I think the most challenging part of this, which is usually the case in public health work, is sustainability. And that’s another key focus of governance that we have yet to work on is to figure out how we sustain this platform once the Implementation Center funding is over.

Department of Health, again, we don’t own or operate it, but we’ve agreed to provide the seed money, if you will, to stand this up. And I think that’s where we’re excited to be partnering, though, with providers and payers. Because I think collectively, that’s how these types of things are going to have to happen moving forward. Public health is not going to be able to sustain things like this, but if we can build a shared governance framework with our partners, and they can find value in a product like this. I think we can all find ways to collectively sustain and pay for something like this to continue moving forward.

So just to wrap up, this is kind of our partners in a nutshell for what we’re doing right now. Again, we’ve partnered with Multicare specifically, the Association of Public Health Labs has updated the ECR now fire app again to trigger on hypertension and diabetes. For us, they do have it available through a GitHub repository, and we’re hoping, and again, it’s based on the 2024 eCQM measures from CMS. We are using our State Health Information Exchange and have the contract in place with them now to develop the shared analytic platform, and they’re working through their agreement with the E-Health Exchange, which is our public health proxy that allows them not to have to be responsible for all responding to queries over the TEFCA network. They can remain kind of safe in that public health space and just be our third-party proxy in this space for receiving the data into the shared repository.

Our IC partner is Chris shared services, and they’ve been providing a lot of great expertise, because they run an HDU in several different states already, and have really helped us, kind of keep things moving, and are helping us kind of herd all the cats, because there’s a lot of pieces to this when you think about a PHL multi care, e health exchange, state, HIE, and then our governance body, which is made up of the department and our partners. And so that’s really what I just wanted to share with you today, quickly, in a nutshell, and with that, I will turn it over to Brittany.

Brittany Saltzman Bell:
Can you hear me? Okay, perfect. Good afternoon. Good afternoon, everyone. I’m Brittany Saltzman Bell from the Kentucky Department for Public Health. Today, I’m going to share a bit about our department’s structure, because I think we’re a little unique, and a couple of ideas for how you can get started on operationalizing governance at your public health agency. So the Kentucky Department for Public Health sits within the Cabinet for Health and Family Services in the Kentucky state government. It. So I’ve highlighted two sections here, so you see on the right, the Department for Public Health, and then on the left, the Office of Data Analytics; both of our programs sit underneath the Office of the Secretary for the cabinet.

However, most of our data governance work has recently been turned over to the Office of Data Analytics via a state statute that passed in 2022. So, a lot of the activities we would like to address with data governance within the Department for Public Health, we have to coordinate very closely with our office of data analytics. The role of the data governance program at our office of data analytics within the cabinet, they oversee data sharing both within our cabinet. If the Department for Public Health would like to exchange data with our behavioral health department, they would oversee the agreement and exchange. They also do the same thing with external requesters. So if a university or other researcher would like to exchange data with the health department, that will go through our office of data analytics.

Now, our data stewards and data owners within the Department for Public Health are involved in that process, but the ultimate, final decision on the exchange of that data is from the Office of Data Analytics. They’ve also been responsible for some policies and procedures that are overarching and apply to the entire cabinet. For example, our data release and suppression policy, we currently have a really good relationship with this policy, where they have let DPH maintain our existing policy, but I don’t know how long that will last. We may have to transition to following theirs. The office of data analytics also oversees a Data Governance Committee for the entire Cabinet for Health and Family Services. This data governance committee had representation from all of our departments across the cabinet and our larger programs.

However, this data governance committee is no longer meeting. Instead, it’s been replaced with a monthly check-in between the office of data analytics and each department to review only outstanding data sharing and data use agreements. So this is a big issue for our agency that we are trying to work on resurrecting that data governance committee.

So, given this structure within our cabinet at KDPH, we spend a lot of time thinking about what role we can actually play within this cabinet structure. So we’ve kind of identified these key points here. We have the ability to survey and prioritize areas that we would like to address with data governance. We’re able to collaborate with others within our cabinet or within our department. We can create some policies and procedures that apply specifically to our agency and only our agency. We can educate our staff and external requesters that we work with frequently, such as our universities, on existing cabinet policies. We’re able to distribute existing policies to those within our department who are impacted by them, as well as maintain some of the existing policies that we have within KDPH.

Some of the key questions we worked through as a data modernization team are listed here. This was used to guide our work within the constraints of operating within our cabinet. So what are our current data governance gaps? And the next one is a big one for us. Are existing governance documents accessible to staff? And do they know where to find them? This was a huge issue for us. Staff would spend a lot of time Googling, searching through our department’s internet page, unable to find a policy that they were supposed to be following. Another one that we focused on was, How do we prioritize efforts with limited resources? We decided to focus on things that had the biggest impact and were quick wins for us to build up momentum.

And lastly, kind of the overarching theme here, what can we as an agency do within our cabinet, state, government, structure? A couple of things came out of reviewing those questions. The first thing our team decided to focus on was governance for existing systems and processes. So we decided we wanted to work on creating documentation to go with a lot of our processes, policies, and expectations. We identified a list of five key systems or processes that were widely used across the agency and had no existing documentation or had insufficient existing documentation. A big red cap is used by the public health agency. So that was one of the primary systems that we wanted to focus on.

We now have a data governance committee focused specifically on REDCap, and have made a lot of improvements. For example, recently transitioned to single sign-on for our REDCap instance, which was a big improvement for us. Also, recently started adopting monday.com for project management across our agency, and you can guess who got assigned to be the manager of monday.com, and the first thing I said was, Oh my god. We don’t have any agreements for any users on this. There’s no documented policy on expectations, how you give a user an account, or how you remove their access. So we have about 400 users across our department, so that’s been a big one we’ve been focusing on.

We are also in the process of adopting GitHub, potentially DevOps, and GitHub is still a little up in the air, but since that is one that is just now getting off the ground, we wanted to use that as an opportunity to start from the beginning of the implementation of a new system or process and building the governance to go along with it. And then the last two I’ll talk about in some additional slides. And that is more focused on policy, our data retention policy, and our data release and suppression policy. So this is a little phrase that I came up with after working in state government for almost 10 years, and that’s my word-of-mouth policy. I’m sure you all have experienced these, where there is institutional knowledge and history, but no written policy to go along with something. For example, our data release and suppression policy was the perfect example of this one. I’ve actually had the quote that was said to me when I started in 2015, John told me I should suppress all case counts less than five, and that was the data suppression policy.

So we also had heard there’s a draft floating around from 2007, it’s never been signed, but everyone just knows that’s what you’re supposed to do. Well, that is not very acceptable for data governance. So we decided to create a work group to start moving some of these verbal historical policies to written policy, especially now that we have a lot of staff getting ready to retire, you’re going to lose that historical knowledge. Found that this was really easy for our staff to get on board with, because all it is taking is an existing known policy and putting it in writing; we didn’t run into a lot of concerns from staff about this not aligning with the way I’ve always been doing things. So that would be one recommendation for you all to maybe focus on some of these policies that already exist, and you just don’t have the documentation to support them.

One of my other word-of-mouth policies that we’ve been focusing on recently was our data retention policy and how it relates to outbreak data. This one was not said to me, but to a co-worker. When I joined in 2002, I was told I had to keep all data related to outbreaks indefinitely. And I don’t know if you all were around in 2020, but there’s something pretty big that happened, which meant our disease surveillance system is pretty overwhelmed with records at this point. So we set out to figure out what we have to keep and how long we have to keep it. What’s in writing? How can we change what’s in writing if we want to change it?

We actually had the opportunity to do a technical assistance project with our friends at ASTHO to look at our data retention policy. So they did a lot of fantastic research for us. Turns out there’s nothing in Kentucky regulation or statute that says you have to keep outbreak data until the end of time. So that was a nice win, but it does say you have to keep all disease-related data until the end of time. So even worse, but we were able to come up with a process that aligns with what’s written in statute on how we can change that, so that is a project for the upcoming months. How can we decide as a team what is a reasonable amount of time to keep our reportable disease data, based on what our neighboring states are doing, looking at you, Tennessee, Ohio, Indiana, what you all have been doing? And then how can we change our process to be more in line with what other states are doing?

Since our MBS is on the cloud, there is no way, from a sustainability point of view, that I can continue to pay for Cloud Storage for the amount of records that we have. So this is one of our attempts to get that cost down. So, wrapping up on some lessons learned here. We have had some really exciting wins at KDPH. They seem small, but I love documentation. I love to dig through old policies and procedures because I am weird, and so I was very excited about these. We were finally able to get the KDPH data release and suppression policy, and writing building off of a draft that had been floating around since I was in high school. So that was really exciting. We’ve also made a lot of headway on our departmental collaboration. So within our seven divisions at KDPH, most of our work groups working on these topics have representation from each of those divisions. So it’s not just Brittany Bell and the Commissioner’s Office telling every division what you’re going to do instead. It’s a collaborative effort. I feel like that really helps the buy-in with all of our divisions to implement these policies.

Now we do still have some areas for improvement, probably not a surprise, but document, document, document, if you don’t have it in writing, do you even have it? That has been a big issue for us, making that transition from historical word-of-mouth policies to having something in writing that could outlast my computer or me in my role. We have also begun conversations with our partners at the Office of Data Analytics about resurrecting the cabinet-wide data governance committee. We feel within KDPH, there is a huge need for this committee to start meeting again with representation from our sister agencies and from KDPH, and for it to be a collaborative effort, not so much the office of data analytics say you must do these things, but instead allowing our agencies to be involved in drafting those policies and how they are implemented, we’ve also talked briefly about a KDPH Data Governance Committee for those areas, from our initial kind of questions and survey that we felt we could impact outside of our role in the cabinet.

And lastly, this is one I’ve been working on right now. It is a repository of existing governance documentation. We have a very good resource called our intranet page, where all staff have access to lots of reference documentation, but it’s not complete. It will take quite a bit of time to go back and dig through the paper and electronic copies of policies that we do have and make those available for staff, as well as training staff on where they can find these tools, there’s also an added issue where the internet page is only available to state employees, so our local health department friends are not able to access it. So it’s not a perfect solution, but it is an attempt to start to make those resources available for everyone. So I hope that maybe gives you a couple of ideas on some small wins, so your word of mouth policies and existing systems, and hopefully you can go back to your state health or local health department and start making some progress on data governance. Thank you. Over to you, Anup.

Anup Srikumar:
Good afternoon. Everybody hear me. Okay, if you can’t let my wife know, she thinks I talk too loudly. So, interestingly, a couple of things before I start off on this is for folks who said that stakeholder engagement is a top win that you have, please let me know, I’ll apply for a job at your office. Because in my experience, that’s one of the most complicated things to talk about, right? Because it comes down to, I’ll record something I learned probably 40 years back, is when you have two people standing at a bus stop, they have four opinions, right? Same thing here. You have two stakeholders in the room. For one, calling themselves stakeholders itself is a challenge. On top of that, they’ll have more opinions about what is needed. Typically, it’s something like, I’ve been signing vital records like this for longer than you have been alive, and that was factually true, right?

Why would I modernize 40-plus years? My state register had been signing it the same way, right? You want access to data, you have to go look at their actual record. You have to go into the vault, pull up a record. No obvious reason, things like that. So you also have something that is sort of toddler behavior, it says, Who right? If I ask you to do something, most you ask me do something, most likely I’m not going to do it right. Why should I change habits? Why should we do this? I like the success stories in Washington, especially because they. It is a single-minded drive in getting something done, a luxury for us, easier for them, and exactly like what Brittany mentioned about policies, right? Everybody loves hiding behind policies, right? The four-letter bad word called data. Nobody wants to talk about data. They assume that they’re losing control, giving something over, and things like that.

So something that we will talk about, from the Virginia Department of Health’s viewpoint, is we had multiple tracks of work going on, right, pre-covid, during covid, and now post-covid, right, where we are talking about data initiatives, some of them are business stakeholder driven, right? Some of them are pure programmatic-driven. A lot of them are CDC-driven, right? Plus, we also have technology initiatives going on in play, right? You want to modernize your Oracle Applications. You want to modernize from Excel to anything else, or you won’t even do something like now, cloud-based solutions, or AI-based solutions, right? Everything that we talk about is going to boil down to data. And last year, they did identify that more than 80% of the data that we have collected across the board is post-COVID. So imagine when you say you can extrapolate that as many times as you want, but what are you going to do with that data? And again, says, Who? Right? People love me at my organization, but not that much, right?

The first question is going to be, why would we have a cloud repository talk to most EPIs, most of whom are like, ” If I put data into a cloud repository, I lose control. Right? Contextual value of the information is gone. I have more footnotes and actual data. So all those questions have to be answered, and it’s incredibly frustrating for technicians like us. Informatics is more of a technology blend, where we have to go back to this and say, Let’s try a slightly different approach, right? You have these initiatives in play, and again, truly, data modernization means different things to different people, right? You also have to worry about sustainability grants; a lot of other conversations are in play. And then you have infrastructure agencies, which say, We will love that you can play in your own sandbox, but if you make it an enterprise solution, you cannot do this right? So then that’s where it sort of slowly moves along at a very slow pace. State governments are all about efficiency. I’m sure you all know this better than I do.

It’s always an efficiency problem. It’s not a funding problem, it’s not a vision problem. It’s efficiencies always, so what we looked at is how to change this narrative. How do I get actual work done? And we say, let’s bring in a council and make sure that your data modernization again, more labels, right? Your Data Modernization Council Advisory Council, as such, has a stake in this. Why? Because if they want, if somebody wants, let’s say I want to get a project approved. I want to get an activity and initiative approved, and I propose that it goes up the food chain. There is a formal governance process, there is an intake process, and all that’s being established. So that’s why we have to make sure that governance can be implemented only through the lens of your advisory council. You need to buy in. You need people to understand this is not COVID-specific. It’s not condition-specific. The Google Cloud Platform that I love to implement needs to be alive, needs to be sustainable, no matter what happens. So that’s why we want to make sure that the data modernization Advisory Council has full skin in the game, right, and things like that.

So what we looked at is, and again, it’s only a label that everybody has in every jurisdiction. Every entity has its own ways of labeling. Advisory Council. We used to have a data governance council in the beginning. Again, more flashy words like governance have negative connotations, which means they imply you’re taking something away or you’re policing stuff. Nobody likes that. So what we sort of did is look across the board, and again, we are a democracy, but we want an autocracy in the middle. Everybody has an opinion, but you have to make sure that that can be collated across the agencies, across stakeholders, identify what key projects exist, key initiative exists, and make sure that that overlap can be identified and taken up the food chain.

So I don’t want my boss, the CIO, or the state epidemiologist, who is our key stake business stakeholder, having to worry about every single project, have to worry about every single activity. What’s the goal? Maintenance. What’s the data? Is there data sharing? All those processes? So the way we did this was, how do you make sure that you have consultation and input from multiple stakeholders? And you’d be surprised if you actually had a stakeholder listing, it would be extremely long. You’d need a governance process for that. We suddenly identified that we have projects. We have a Commonwealth state agency that does infrastructure support. So like, remember the sandbox I mentioned about the Google Cloud? I can put it on my desktop, but if I put it in the Commonwealth infrastructure for scale and sustainability, they have different rules. You can’t do the same things again. You cannot govern your own systems. You cannot give people access to lots of things, like they become a key stakeholder, which nobody talks about. They’re the most expensive liability there. But even then.

So, making sure that we can help folks understand that this decision-making needs to be done. What goes up the food chain? We have clear decisions in identifying if it’s go, no go, or if it’s going to be, and we are sort of separating the funding streams from the actual activity. You see how I didn’t talk about governance at all, right? You say governance, people are like, I don’t want to do this. So we focused on actually getting operational stuff done. The Advisory Council is focused on initiatives. It’s easier to talk to them about what is actually getting done and what impacts we have, what overlaps we have. I’m a data person. I understand Venn diagrams all the time, so helping folks understand what they can get done is key. So that’s why you said, raise the awareness, make sure we have adequate participation, and you have cross-agency representation. We have, like, we have key stakeholders in local districts, or we have independent members and things like that. So that was extremely key, setting up those expectations and goals.

Now, as I said, bring a group of people together, if you don’t tell them what you need done, or at least have a nice voice and say, you know, we would like you to get us to this particular stage. We clearly identified again, the buck stops with me, literally, right? I have a co-chair, of course, but we are like I have, if you pay me enough, I’ll approve your projects. But essentially, the idea is, float the projects up. And we have found quite often that we have lots of solutions, lots of initiatives, processes which are very different from Central Office versus operational districts and divisions, right? So make sure that everything you want an AI initiative that’s running next door, they talk about that, that is brought up with the DREAM Act, right?

Because you have the modernization label that you can attach to it. And then we sort of look at the pros and cons, make those strategic, technical, and program decisions, and bring them there. That way, we can make those collective decisions, and then tap into a variety of funding mechanisms that we use because either you have no money in one place or you have too much money in one place. So make sure that we can have that overlap being done. So identify that respect people’s time, and make sure that you tell them what you need, what the coordination is going to be, what the information flow is going to be, and what strategic decisions we can make. So that was key in terms of building that foundation. Again, nothing new. There’s no brand-new Kool-Aid that I’m selling. It’s always going to be stuff that’s all work that has already been done. How do you put that together? Repackage it so that it becomes a lot easier.

The lesson learned: start small, show value. I don’t 100% agree with that, because Kevin’s low-hanging fruit typically gets me into trouble too many times, because once you have an initiative, and then you go back with the bill and say, it’s not sustainable, but you need a million dollars to do this. But still it is. It is how we frame that conversation. It is how we talk about it because some of these stakeholders come up with projects. We had a data modernization project, which was modernizing Xerox printers. We procured $75,000 to get new secure Xerox printers for vital records. The side benefit of doing that was that it is a secure print service; all the data that we are manually digitizing after that, instead of getting printed alone, you get access to that data. It cost them only 20,000 to do it. But anyways, beyond the point, Xerox never gave us a bill, but the idea is, these are initiatives.

These are projects that have facts and kudos to our state registrar. He pushed that project through. He’s like, Anup, I have this project. Can we get money for it? That’s all they wanted. It just wanted to be a formal activity, but the extremely impactful quality of their data increased by 96%, quite simply put. I mean, I’m like, whatever, when you ask me for maybe a couple of extra dinners, that would be nice, but you know that is actual, factual value that you have there. Documentation is key. I think you already heard it multiple times from Brittany. Documentation is key. Why? Because it should not be, Anup said so all the time, or, you know, make sure that we understand those processes are, and then audit comes along. Everybody’s happy. Shared governance. I don’t want to use the word again, but again strengthens buy-in. You have to be able to bring opinions to the table.

There, sustainability, we talk about that in public health, space sustainability is a big shadow that we have. But again, efficiency, if you’re able to have cross-cutting approaches to this, and look at it from the perspective of saying initiatives and grant funding or other sources of funding can be separated, and the overlap can be identified, and then communication builds culture. We love talking, but making sure that that can actually translate into action is again, there is no template here, so only guidelines, best practices, and everybody can, sort of, you know, benefit from that. I don’t want to go into the slide. I don’t know how I managed to write it up, but we only have a few minutes left, yeah, so we will be sharing that also. But again, the message is, how do you make sure these initiatives can overlap with governance? And you know, how do you make sure that they are actually successful as such? In case you need to reach out. I believe slides will be shared if you need to reach out for any information that’s listed there. Thank you.

Santiago Gonzales Irizarry:
All right, so I think that this is really great, because we see three different states with different approaches, Washington having in their data, governance structure partners that are participating in these projects, and Ken talking more at the interagency government level. So it could get more complicated, and a big state like Virginia, including its locals in their data governance. And I think that that’s the lesson, that it’s going to look a little bit different depending on the needs of your PHA, and the best part is seeing how in these states data governance has matured, so we can begin to see it in action. And I know that we’re a little bit pressed with time, but I wanted to ask a question to the panel about when we get to the level that we can make decisions and put data governance in action, maybe a little bit more of a tactical question, but if you could share one or two things that you have implemented successfully, have you implemented specific tools or platforms to support, data governance, data catalogs, glossaries, dashboards and what would You recommend again, one or two, and I’m hoping that the mic, the microphones, will work.

Chris Baumgartner:
It’s on, all right. It is on. So when I talked to our data governance manager, the thing he recommended the most highly is the data catalog. He feels like the value of a good catalog isn’t, though, in collecting the information; it’s in giving you the means to integrate that information and the resource into the daily lives and workflows of the people, really making it human-centered. It’s less about the data itself and more about how we interact with it. We do have a few dashboards that we try to use to track some important things, though, like data sharing agreements. I don’t know all of you; we were in one of those situations where we didn’t even know how many data-sharing agreements we had. I mean, I will be transparent and honest. It was kind of ugly. We’re like, Yeah, we have them. We think we have most of them documented in this repository, but probably not. And so just even that simple step of having a SharePoint site and a dashboard for tracking data sharing agreements is a huge first step.

Brittany Saltzman Bell:
Glad to hear. I’m not alone in having no idea how many data-sharing agreements there are. For our agency, we have a couple of resources that serve as a data catalog. We have one that we maintain and share for all of our external partners. We call it the Data Resource Guide. It’s updated semiannually, bit of a delay during COVID, so it was a little longer than semiannually. But this list includes every data system that we maintain at KDPH or via one of our bona fide. Agents at our University of Kentucky partnership, where the public can request data. It also has information on the methodology used to collect that data, the types of data that are maintained in it, who the key point of contact is, etc. Then we have more of an IT side for a repository that lists how we rated it in a cybersecurity assessment, prioritizing what systems will come back online. We have a good partnership with our preparedness program to build that out, and that has our IT key points of contact, our public health point of contact. So we have two separate systems for unique uses, and both have been incredibly important.

Anup Srikumar:
We did look at the data catalog component a few times, and too much head scratching. Moments later, we are like everybody has a different version of it, and we found that implementing our data repository solution by itself, which is the Google Cloud Platform, made it was then easier for us to build on top of that. Why? Because everybody has a handbook. Some of them are historical. Some of them we should not use stuff like that. So everybody has a different word understanding of the word catalog. So we were like, Let’s bring it all in. We sort of started filling out and the under, building the understanding about the data, life cycle, right, left to right. I implemented a quality data, quality dashboard process. The commissioner’s office called me the next day. They are like, You cannot say this, right? You cannot say that. So things like that. That’s when we have to understand that quality means different things to different people. And then the cataloging, again, was a side benefit where you fill out metadata documentation, and that becomes your catalog.

Santiago Gonzales Irizarry:
Thank you. So I think that we have been hearing some common challenges. It’s good to know that you are not alone in that journey. In DC, it took us almost a year to put together our data governance team, and with the help of events like this, it was possible because that’s how you can share ideas, and you can learn from your peers. So I think that even though we didn’t have enough time to get more questions, let’s keep this conversation going and keep sharing that knowledge and those ideas. So please help me, thank our panel for participating this afternoon, and look forward to continuing to talk about data governance with all of you.