Integrating behavioral health services in pediatric primary care has been identified as a core strategy for closing the worrying treatment gap faced by young people with depression, anxiety, ADHD, and other common conditions. Pediatricians and psychiatrists have stressed the importance of addressing behavioral health in primary care settings for more than a decade, but in the same breath, they've also acknowledged the many administrative and financial barriers that stand in the way of behavioral health integration.
Four years ago, the Smith Family Foundation and an interdisciplinary group of clinicians and researchers from Boston Medical Center (BMC) and Boston University (BU) took aim at these barriers with an initiative known as TEAM UP for Children. Pushing further than most integrated models, TEAM UP seeks to not only transform pediatric behavioral healthcare at community health centers (CHCs), but also to build the lasting capacity at CHCs necessary to sustain the transformation well into the future.
Just as innovative as the model itself is the way TEAM UP set about creating it. Within some broad parameters, every aspect of TEAM UP's implementation and evaluation was co-developed alongside the CHCs. Inspired by quality improvement frameworks and methods drawn from implementation science, a field of study focused on facilitating the adoption of evidence-based clinical practices, TEAM UP's collaborative, iterative approach was designed to avoid the potholes that have hobbled behavioral health integration in the past.
Within some broad parameters, every aspect of TEAM UP's implementation and evaluation was co-developed alongside community health centers.
TEAM UP is now starting to roll out its early results. A recent qualitative study surfaced key themes from in-depth interviews with dozens of CHC clinicians and staff, and a second study, published in Health Services Research, was an important first step in establishing TEAM UP's sustainability. The latter study — the first to report hard data from the initiative — compared health claims from the three TEAM UP sites and non-participating CHCs and found an increase in primary care engagement at the former without a corresponding increase in avoidable healthcare costs.
In the fall of 2019, with added support from The Klarman Family Foundation, TEAM UP is expanding to four more CHCs. As the second phase of the initiative kicks off, the co-leads of the evaluation team — pediatrician Megan Bair-Merritt, MD, MSCE, of BMC and the BU School of Medicine, and Chris Sheldrick, PhD, of the BU School of Public Health — sat down with HealthCity to discuss TEAM UP's novel approach to implementation and evaluation.
How does the TEAM UP model differ from other implementation approaches?
Chris Sheldrick (CS): There's a long history of scientists doing their work on the outside, trying to make discoveries and figure out the right way of doing things, and then going out to the community and saying, "Actually, the way you do it is wrong. You should do it this way." And then, naturally, there are conversations about barriers. "Why aren't you doing what I told you to do? There must be reasons." A fair amount of time and money has been spent advocating for evidence-based practices that just aren't practical and don't show real benefits on their own.
TEAM UP is a very different model. Behavioral healthcare needs to be integrated into pediatrics, but the details of how that happens get worked out with the actual practitioners. That leads to solutions that are consistent with the evidence but also practical on the ground.
Is the co-development model especially important or beneficial for CHCs?
Megan Bair-Merritt (MBM): I think any pediatric practice — whether it's at a CHC, or another site — knows their processes and culture and the populations they care for in a way that obviously we don't. The system has to work for them. It has to fit into their organizational culture, the priorities of their providers and staff, the needs of the patients they're seeing. As a general rule, it just doesn't work well to come in and say, "Here's the program." In the first cohort, for instance, we had three different sites using two different electronic medical records. And not only that, the way the individuals and systems interfaced with those electronic medical records was different.
CS: There hasn't been a full recognition of the complexity and sheer difficulty of doing very basic-level change in these practices. If a practice has never done evidence-based screening before, and they decide they want to hand out a new screener to each patient, changing the workflow to make that happen is actually unbelievably complicated and takes a long time and a lot of staff. Through the co-development process, TEAM UP offers a lot of support for the practices to help figure out all of those individual points and decisions along the way.
Boston Medical Center pediatrician Megan Bair-Merritt, MD, MSCE, and researcher Chris Sheldrick, PhD, of the Boston University School of Public Health, co-lead the evaluation of TEAM UP for Children.
Can you share an example of how those details get figured out?
MBM: In the first round, the evaluation and implementation team did something called a failure modes and effects analysis for two processes: behavioral health screening and warm handoffs. We brought together the entire team and brainstormed all of the ways a process could break down, and then assigned a score to each based on severity, how common it is, how likely you are to detect it. And then we problem-solved how to modify or improve the process.
There were some commonalities across sites, but each site had its own barriers. The analysis allowed us to identify the site-specific barriers as well as the larger barriers to greater implementation — and meta-barriers such as billing and sustainability.
How did the co-development model come into play with evaluation?
MBM: It was really important to us to think about the metrics as a piece of the co-development work. From the beginning, it was clear to us there were no comprehensive evaluations of pediatric behavioral health integration in a CHC environment that we could look to, either for the model or for the evaluation. So we actually co-developed unique indicators that we can now report out. It was a whole new conversation about what would be helpful to know. I've not been in a scenario before where the metrics were developed collaboratively with the folks who are doing the actual work.
After that deep thinking, we have moved into overtly saying to the health centers, "We won't interpret your data without asking you what you think it means." It was a simple, easy way to agree that it's a two-way street. While there are data, there is also experience — and both of those need to be understood in the context of the other.
Why is that two-way street so important?
MBM: Health centers, historically, are the recipients of metrics that have been developed by other organizations. They pick them up and have to report on them. And that data has often been used in a punitive way: "You're not meeting this benchmark, this provider's not meeting this quality measure, and so we have to put these things into place." Versus, how do you really use data to be curious and drive change in a positive way? We've been explicit about building a culture around data use where people are excited and interested.
CS: The data culture in TEAM UP is developing well. The staff and clinicians really value data. I'm a full-time researcher, and I spend a lot of my life analyzing data. With a lot of audiences, I'm told to keep it simple. Cut to the chase, tell a good story — that kind of thing. When I walk into these meetings, they want more complexity. They want to know, "Okay, well, what does this mean? How did you measure it? Why do we think it's different?" They're really interrogating the data and trying to understand what's going on. It's a really good sign when people engage in that way.
Clinicians and staff from the three community health centers were involved in every step of the implementation and evaluation process. (Courtesy TEAM UP for Children)
Do you feel the TEAM UP experience has broader lessons for implementation science?
CS: I think there should be a paradigm shift in the field. In medicine, the only example I can think of where scientists pretty much that know a treatment is good for everyone is a vaccine — and even then, not every vaccine. It's pretty rare to have that level of certainty.
We don't have anything close to that level of certainty in behavioral health. Yet behavioral health is what's tearing down our society and making kids and families miserable. We have to do something about it. We know a lot about what is likely to work, and I think the only way to make it happen is to have a tremendous amount of respect for the people on the ground and come together with them to jointly find solutions. Moving forward, I really think the co-development model used in this project is the way to go with implementation science. I think it could work in all kinds of different areas.
What does the next phase of evaluation hold?
CS: We have a slew of papers planned. And going into this next phase, based on all the co-development in the first phase that we just talked about, the chance for co-learning is tremendously more powerful even than what we've had so far. We have more data, and existing modeled relationships with sites to work through what those data really mean. There's just tremendous potential going forward.
MBM: We've learned a ton. How does data get put in? How does it get pulled out? Which metrics are reliable? Which aren't? And that's allowed us to change how we're getting data for the second round. As Chris said, it's going to get much more powerful. Our ultimate goal is to use that power not only to support the TEAM UP health centers in transforming how they care for children and families, but also to capture that collective experience and more broadly build the field of knowledge about integrated care in pediatrics.
This interview has been edited and condensed.