[music playing]

Oby Ukadike-oyer: From the campus of Harvard Medical School, this is ThinkResearch, a podcast devoted to the stories behind clinical research. I'm Oby, your host. ThinkResearch is brought to you by Harvard Catalyst, Harvard University's Clinical and Translational Science Center, and by NCATS, the National Center for Advancing Translational Sciences.

Karen Emmons: Hi. Welcome to this episode of the ThinkResearch podcast. I'm Karen Emmons.

I'm a professor at Harvard T.H. Chan School of Public Health, and I am the faculty lead of the Harvard Catalyst Community Engagement Program. And I'm delighted to be here with you today to be joined by two of my wonderful colleagues and community partners, Susan Adams, who is the Vice President of Health Informatics at the Massachusetts League of Community Health Centers, and Dr. Cheryl Clark, who's the Senior Vice President and Executive Director of the Institute for Health Equity Research, Evaluation, and Policy, also at the Mass League of Community Health Centers.

We're going to talk today about the NIH data sharing policy and what data sharing means in the context of community-engaged research. The current policy at NIH on data sharing is basically designed to facilitate knowledge sharing and discovery and learning. And that's a really great thing. When we do studies and we hold that data to ourselves, others can't learn from it. And so it is quite important to be thinking about how these large investments that the government makes in research can be leveraged in the best possible way.

At the same time, we have to recognize that not all data is the same. Not all data is acquired in the same way. And there are issues for the different kinds of data that are going to be important to consider.

So today, we're going to really be talking about community-level data. So I'm going to just kick it off by asking both Cheryl and Susan to talk a little bit about the role that you each play in the acquisition management and use of community-level data. Really, how do you think about data ownership in relation to the data generated in the context, particularly of health care and federally qualified health centers. So, Susan, can I ask you to kick it off for us?

Susan Adams: Yeah, so happy to kick it off, Karen. Thank you for having both of us here. I think I should start with the access to data that I have and the role that I play in that. My role at the Mass League is to oversee what's called a Health Center Control Network, or HCCN. And we help the community health centers in Massachusetts with their technology and data needs.

And along with that, the health centers participate-- not all of them, but many of them-- in feeding their data into a partner of ours, which is a large data warehouse population health tool. They have access to that tool to see their own data. But as the Mass League, we have a view into all the health center's electronic health record data that's feeding into that.

So in terms of ownership, what we view that as is, it's the health center's data and their patients' data. We at the Mass League do not own that data. We simply are stewards. And we tend to use the word stewards of the data to use that data internally for advocacy on the health center's part, to use it to determine where they might be struggling with quality metrics, and where can we provide training and technical assistance?

But along with that comes outside researchers that request because community health center data is rich, and it's valuable, and it's really what we need for insight into health equity. So we do get a lot of requests for research. And my role in that is to really do the data governance around providing that data on behalf of the health centers to those researchers.

Karen Emmons: Fantastic. Well, we're going to dig a little bit more into the question of how research is conducted in the context of this data in just a second. So it sounds like there is this really vast amount of data, and the Mass League is in a role of stewarding that data on behalf of a number of organizations to do some of your own learning and feed that back to the health centers. Cheryl, how do you think about data ownership in relation to community-level data?

Cheryl Clark: Thank you so much for having me and for the question. And in my role at the League, I oversee the work to use data for research, generating new knowledge, and for evaluation purposes, trying to understand where we are and what programs are doing and how to shape what the future looks like. And at the League, the work that we do is in service of community health centers, both federally qualified health centers and community health centers across Massachusetts.

And the work that we do centers this idea called emancipatory research. And the idea of emancipatory research is really, in part, that we invest, and we ask the questions, and that the data to really follow the priorities of health centers, and that we're also oriented toward action and solving problems with those data. And that orientation is important as we think about what it means to be good stewards of data and what it means to be sharing and also prioritizing health center ownership of their own data.

It, in part, means that as we go through the processes of what data collection looks like-- for example, some of the work that we're engaged in is qualitative in nature, where we're trying to understand stories and trying to get a better sense of what's happening in community settings-- that when data are collected, it's done in a way that if there are survey instruments or if there are qualitative strategies that are put in place, that happens collaboratively. And when the data are quantitative-- and I really appreciate Susan's leadership in governance in making sure that this is done ethically and in partnership, that there is a strategy in place so that the data are interpreted in community, and stored, and that we all understand what they mean and how to disseminate as needed. So for me, the way that I think about this is really in that sort of community-led, community-owned, and community investment that is a part of what an emancipatory research perspective would bring.

Karen Emmons: That's fantastic. And as you know, so many of us are in the background of cheering the Institute on for its focus on emancipatory research, for its really trying to center all things community health center and health center community in the work that you're doing. So we're very excited about that and very interested to see how that unfolds in terms of the kinds of questions that are asked in research.

And particularly, I think the innovative ways of addressing some of the really thorny problems we have in trying to improve community health these days. So it's very exciting. Susan, could you say a little bit more about how you think about stewardship and governance of data, what the governance structures are that you put in place to protect the data, to make sure that the ownership stays where it should be, and then a little bit about how the research requests intersect with that?

Susan Adams: So we can see at the League the PHI-level data. So we can see patient-level identifiable data. There are only a few people at the Mass League who have that access. Most anyone else at the league would see a more aggregated higher level de-identified, more dashboards, and reports, and measures. But we do need to drill down to the patient level often for research requests and other needs and get to the deeper level of data.

When a researcher comes forward and says, I would love to have some of this data, I can't just hand over the entire EHR. So there is a very structured process we go through, a governance process that we try to follow. It's very layered. It's very time-consuming.

But I do this from a security standpoint and making sure that all approvals are in place. The health center's aware their data is being shared. They agree their data could be shared.

So what we would normally do is, we'd start with a meeting, and let's talk about what that data specification is going to look like. What data elements do you specifically need to do your work? Can I provide those to you? Do I even have access to those? Does it make sense?

And there's a lot of tweaking. We could go down a path of data standardization and where that is and disparate documentation and EHRs. So we really have to dig deep into how much of that data we have, how accurate that data is, and can we provide it? So we'll go through a lengthy process.

And once we have that somewhat solidified, then we have to provide back to whoever's requesting that data what we can and can't provide. And if that meets all the needs-- how often we're going to pull that data, how often we're going to provide that data, who's the control sites, who's participating-- we would move into a cost approach because there is financial concerns with this. It's a lot of time and effort. There's data quality that we have to do to make sure that the data looks good, systems to pull the data. If we can't do it at the level of the Mass League staff, we have to outsource it to that vendor to pull that data, and that comes with a cost.

And once that's approved, you move into the legal framework, which is, if there's an IRB, if there is patient-level data, we need data use agreements in place. We have to contract with whoever's requesting the data. And then we have to get the health center's approval to give that data to you.

It's a lot of work, and it takes a lot of time. Normally, it does take a few months to get through and get those agreements in place and then actually execute the work. And usually, it's not a one-time data pull. It's going to be multiple data pulls because you want to see the progression and that the work that's being done at the health centers is helping improve that data and what the outcome is. So it is a lengthy process, and it does take time.

There is always hesitancy on the health center's part, and we have to keep that in mind. And so, again, as stewards of that data, we want to make sure that the work that's being done with that data benefits the health center. We would not want to just provide data and then have something published out there with the health center name and then be blindsided. And we want to make sure the data is interpreted properly. That's a whole other discussion about how there could be a perception of data missing or the health center's not doing good on a quality metric, but you might not know the reasons behind that.

So all those things have to be talked through, worked through. It's just not here's the data and go. And then you really have to think about, is it de-identified aggregated? Is there any identifiers? And what is the long-term storage of that data? So if you're done with your work, what happens to it? Does it go out for other people to use that might not understand what that data means? Or is it destroyed? So we cover all of that in our governance process. It's a lot.

Karen Emmons: It really helps to unpack that because I do think researchers think, the data is there. Just give it to me. And it's not like that at all. And it's hugely important data. We should be using it for research to answer important questions for the health centers and for their communities, but we have to do it in a proper way and in a proper cadence. Well, how is data usually handled after the work is done? Is it returned? Is it destroyed? Is there a typical or maybe there's not a typical?

Susan Adams: I think it varies. But typically, we will ask and have a form signed. What do you intend to do with that data afterwards? Where is it going to reside? An answer might be, it'll be sitting in our REDCap program, but it won't be used without going through permission to use it from us again.

So that's what we want to hear. Can I control it once it's left the walls? No, but the hope is it's used for that. If it's good, valuable data, I'd love to see it used again in the future, but with the understanding of limitations that might have been addressed during the use of it.

Karen Emmons: Yeah, and with knowledge of the health centers of what is being looked at. Yeah, that's really important. I know in our own work together with the Mass League and many community health centers, it's been really important to use that process to also develop the relationship and say, look, here are some questions we have that we'd like to use your data for. But do you have questions too?

Because while we as researchers are looking at the data, there may be things that the health center would want to know that we can look at with the advanced analytics and other things that researchers have access to that would really be helpful to health centers. So I think, as researchers, trying to think about your own needs, but also being open to using your process to help support the health centers can be really, really critical way to build those partnerships. And I think we should think of them as partnerships. So, Cheryl, that really comes back to as you're thinking about, centering the health center's questions, and resources, and needs, how this research is coming in plays, what kinds of frictions that creates.

And want to see if you would address a little bit this issue that Susan raised of context. I think, so often, researchers come in and think, I'm looking at this data. This data tells me everything I need to know. But in community research, there is so much going on that is not necessarily reflected in the data that you could really get wrong if you don't consider what the political and social context is in which that data is collected. So do you have any thoughts on that issue from that IHE perspective?

Cheryl Clark: Thank you. I appreciate that. And I have learned so much from Susan Adams with respect to what this looks like and just the years of experience that she's brought to the governance of this work. One specific example that came up is, as we were thinking about electronic health record data, in particular, and measures that may in some contexts sort of feed into data repositories from EHRs, we wanted to understand more about an important health issue-- Black maternal child health and also understanding more about birthing people and what that looks like in terms of getting people into prenatal care as a way of reducing mortality.

And we had to understand more about both the ways, and the clinical, and information technology workflows that happen at each health center, but also the broader context in which the data are collected. I guess it's important to understand that health centers and particularly federally qualified health centers have been funded through the federal government-- so the Health Resources Services Administration, which authorizes their funding and also helps to align health centers on quality metrics. And that is a quality metric, getting people into prenatal care early.

But there are also different contexts and ways that data are collected and reported. And those methods often are specific to the needs of the federal government, but, in this instance, did not always require demographic data to be collected. And so when we looked at the electronic health record data that we had, it didn't align with the data that was reported for federal reporting purposes. And so without having the expertise of health centers and of the data stewards, we would have gotten the wrong answer as we just tried to explore these data. So the key word, Karen, that I think you mentioned, partnership-- that partnership is a really critical piece of using data well.

Karen Emmons: Yeah, that is such really important point, Cheryl. And I think that partnership, as all things, can be engaged research, but especially when you're talking about data, the ability to interpret data without context is just terrible. And so we've learned from our own experiences where, in our implementation science partnership with the Mass League, we require all of our authors, all of our researchers on their papers to have a community partner on their papers and the role with a very important role, which is helping interpret the data and helping write the discussion section in a way that contextualizes the data.

And it just makes all the difference. I mean, we will come out as a research team with one view. And then we go back to our practice partners and say, OK, help us understand how you think about this.

It takes something that's in a box, and it just explodes it into often such deeper, richer meaning in such a positive way. It's really, really wonderful. So I encourage all of my research colleagues to really be thinking about, whether you're working with community partners or other kind of partners, it really matters to have that partnership.

And talking about papers and how we think about that, Susan, during COVID-- and COVID, I think, in many ways, changed a lot around data sharing and really made a major push around thinking about data sharing. And we had the good fortune of writing a paper together that highlighted a community perspective on the NIH data-sharing policy. We wanted to lift up another view beyond how people who are mostly focused on a research perspective would think about data sharing. What are some of your reflections on data sharing from a community perspective with community-level data? And what are some of the things you might want to caution researchers about in this space?

Susan Adams: I think Cheryl hit, with her example, some of my concerns. We definitely saw during the pandemic a lack of data collection. Things went to telehealth, so it might just have been a phone call. They weren't doing screeners and regular screenings that they should be doing. And when we started to share that data, the agencies that were looking at that data were saying, what is all this missingness?

So we had to dig deep and talk to the health centers and their patients, and why are they not answering this question? So some of it was the pandemic because it might have been a telephone call, so they did do the screenings. But other things, it's really cultural. They don't want to answer.

They're afraid to answer some of those questions because of the repercussions that they think might happen. And we see that a lot with the undocumented population, the immigrants coming in. There's even a front desk staff that is afraid to ask certain questions.

When you want to ask, what's your race/ethnicity, what's your income level, and the patient doesn't necessarily want to answer that. I would often see the front desk person just choose what they think is the appropriate answer or not answer it at all. And then that leads to the data integrity. And that's what we're finding, that there's a lot of work to do back education-wise at the health centers to make that staff comfortable asking and to make the patient comfortable answering it.

I worked as a former CIO. I implemented electronic measures, which actually helps the comfort level in capturing the data. But then you get into the digital literacy issue with some are not comfortable using technology. So it weighs pluses and minuses each way. I don't know what the best solution is, but it's a continuing education at the community health center level to capture that data.

But on our end, like Cheryl said, she could go in and look at a particular measure. It doesn't look so good. So why not? And we have to go back and dig into, why doesn't it? And it's not that they're not capturing it. It's not that they're doing bad on a quality measure. There's so many other factors that impact it that someone looking at that data might not realize.

Karen Emmons: Yeah. I think it's taking that inquisitiveness approach, like, this doesn't look like I expected why rather than saying, aha, I found something here, and now I'm going to run with that. It is really important to be, especially when your expectations are not met or something looks different than you expect, why is that is really, really important. Cheryl, you do a tremendous amount of work in the health equity space. And any thoughts on how health equity means data sharing and how we might think about this, especially as you're rolling out the procedures of the Institute's work?

Cheryl Clark: I am reminded that the Centers for Disease Control and Prevention has a framework around research and getting feedback on the work to really being community-led, where the work comes from communities to what we hope to add, which is this emancipatory part where you really are able to have investment in the work and where you hold yourself accountable for moving the needle. These frameworks are not just helpful for understanding the kind of work that you're doing.

They're also really important for structuring the way that you engage in research. And it really is a different process when you approach a community organization as a researcher, and you have your literature search and your ideas about how the project ought to go than when you build relationships, when you have the humility to make sure that in order for this work to happen, that you really demure and defer to community expertise, to places that actually on the ground what's happening.

And for me, that's important for data ownership because it's very hard after you have a research project, you have all the questions done, and you've collected the data to gain insights that you didn't have before the data were collected. And so part of that process is making sure that everyone's at the table and that power is shared as the data are formulated as you think about what it is that you're trying to do, as you figure out what sources that you want to use for collection, and then what happens after. Who gets the credit for the dissemination? Who gets to use the insights? All of those processes, I think, are pretty critical.

And we're learning. I'm really excited that at IHE, one of the first projects that had the honor of working on with the health center located in Springfield-- it's actually a completely qualitative project, but has some mixed methods in that we're trying to understand what some of the facilitators are to improve access to treatment for substance use disorder and how to build even stronger relationships with community-based organizations. And so a lot of the data that will be collected will be stories and qualitative information.

But we also will work collaboratively to try to understand more about how to collect visit data and to use EHR data to monitor progress. And so I think that example just illustrates a couple of things. One is just how much more rich the research questions are. When they really do come from the health center, you start to work on things that really can get implemented and can lead to change. And also, it illustrates how effective I think processes are when you design work from start to finish in partnership.

Karen Emmons: Yeah, that's such a great example. And that work is so, so important. Both of you, in some of your discussions today, have talked around the issue of returning results to community.

And I want to just put a really clear pin in this issue because when researchers come in and they utilize data from community groups-- and again, this ownership issue comes into play. They're thinking often, OK, I have the data. I have my questions. I'm going to answer it.

When you're doing engaged or emancipatory research, that gets shaken up a little bit. It's more of a partnership. We're answering questions together.

But, to me, the return-of-results piece is absolutely critical, regardless of how you approach your science. Going back to the people who created the data, who have ownership of the data, and letting them know what you found, I think, is really, really important. And I'm wondering if either of you have thoughts on that. What are best practices for that? How-- examples that you've seen that it's done really well or warning shots that you want to make for people not to do it, but any thoughts you have on that, I think, would be really helpful.

Cheryl Clark: And I'm happy to just mention a couple of projects. So I should also mention that I'm also a physician and a researcher at Brigham and Women's Hospital and Associate Professor at Harvard Medical School and, in that capacity, have had a chance to work on a couple of projects that I think are best practices. The All of Us Research Program is one of the National Institutes of Health's largest programs, and it's funded to advance this concept of precision medicine, that we really ought to tailor our understanding of science, and lifestyle, and biology to the individual.

And part of the ethos of All of Us is that you need to return data that you collect and make sure that people who are participating in research get the benefit of that research. All of Us collects data for genomic research and returns those gene findings back to participants with explanations so that people get a chance to understand a bit more about that science and that background as well as their own heritage, in some ways, or their own genetic heritage. And I think that's a very nice example of making good on that promise.

I've also had the privilege of working with the Jackson Heart Study in Jackson, Mississippi, which is the longest running longitudinal study of African-Americans and heart disease in the United States. And it is a requirement that researchers who work in the Jackson Heart Study cohort prepare lay summaries so that when the research is completed, the key take-home messages are communicated in a way that is easy for just the general public to understand.

Researchers and the findings are also highlighted in newsletters. So not only do you get a chance to learn a bit more about what the research is that's been done, but you get to learn a bit more about the researchers. And it sort of humanizes the people who are conducting the work. So I think those are really strong practices for folks to consider.

And I would say, there are also technical things to remember as you are putting together your human subjects' protocols. Often, it is necessary to ask for permission to recontact participants after the work is done-- and so remembering to include dissemination as a part of your protocol so that you already have the ability to either contact people during the process or after the process is finished so that you can return those results-- so important to think about it in the beginning so that you write that into your study process.

Karen Emmons: Those are such fabulous examples. And I think that last point about having that in your mind from the very beginning and then having it built into your processes is the way that it gets done as an afterthought. It usually doesn't get done or get done well. Susan, anything you want to add to that?

Susan Adams: No, I don't think so. I think because where I sit more on the technical data delivery governance side, I don't necessarily always get to see the outcome of that data.

Karen Emmons: Well, that's something researchers should remember because as you're the person that's stewarding that process, it might be good for you to see that good things have come from it. So that's just a lesson to the wise, yeah, for sure. Well, this has been such a fun conversation. I'm going to just ask if you have any last thoughts for our audience about data sharing, the way in which we should be thinking about it in terms of partnership, in terms of community context and health equity. Any last thoughts?

Susan Adams: I think I would just-- again, not to recycle back, but I think it's just understanding the challenges of the data, keeping that in mind that it is time and effort. I do sit on quite a few committees right now that are trying to create, whether it be a statewide or a national data warehouse or lake house and have access to a variety of health care information, not just community health center, but let's get the payer information in there. Let's get the hospital information in there and create these more robust data warehouses.

However, with that, it's going to come a tricky governance for us in making that secure, and accessible, and understandable. So I just think we have to continue to work towards data standardization, and there's a lot of initiatives going on with that. There are national data sets out there, and adopting those more across the EHR and EMR platforms is going to be beneficial because we'll have better interoperability, better data sharing.

It's getting there. It really is. Everybody's working towards it. It may be in a different variety of ways, but we all have similar goals in making that data available to everyone, including researchers.

Karen Emmons: Yeah, that's fantastic. Cheryl, any last thoughts?

Cheryl Clark: I guess the only other piece that I would add is that this concept of data sharing and data ownership is really-- it's a value more than anything. And part of that is making sure that as we build our research projects, that we value those partnerships and make sure that they're there and present and that those processes are there as we put research projects together, and that, quite frankly, we also make the investments that we need in community settings so that community-based scholars also have the standing and the resources to own their own data and to ask their questions and to do this work. So all of that is a part of the process of health equity and data ownership.

Karen Emmons: And I would be remiss if I didn't pull out one of the threads that came across our conversation today, and that is about partnership and particularly using community-engaged principles as you're thinking about using community-level data. If you go to the Harvard Catalyst website, you will find that Community Engagement Program portion of the website-- really fantastic resources that are available to help investigators think through every aspect of doing community-level work. But in particular, we have something called the Community Coalition for Equity and Research that is really designed to bring you researchers together with community members who can give you feedback and ideas on your data, including around data sharing and data access issues. So that's a great resource for anybody who's working in this space.

This has been such a delightful conversation. I'm so grateful to both Susan and Cheryl for spending the last bit of time with me and really thinking through these issues. I know they're going to be really important and helpful to our community. So thank you so much for your time.

Susan Adams: Thank you.

Cheryl Clark: Thank you.

[music playing]

Oby Ukadike-oyer: Thank you for listening. If you enjoyed this episode, please rate us on iTunes, and help us spread the word about the amazing research taking place across the Harvard community and beyond. We are always looking to connect and collaborate with the research community and would like to hear from you. Please feel free to email us at onlineeducation.catalyst.harvard.edu to inquire about being a guest on the podcast.