Research Tracks Impact of Collaborative Care Model at Penn Medicine
Researchers at Penn Medicine have published the first community-based study to demonstrate improvements in suicidal ideation, depression, and anxiety among individuals with suicidal ideation receiving collaborative care services.
Collaborative care is an evidence-based approach to identifying and treating patients with behavioral health conditions such as anxiety and depression in primary care settings. The researchers examined data from Penn Integrated Care, a collaborative care model (CoCM) program including an intake and referral management center plus traditional CoCM services implemented in primary care clinics within Penn Medicine.
One of the co-authors, Gabriela Khazanov, Ph.D., a research psychologist in the Center of Excellence for Substance Addiction Treatment and Education (CESATE) at the Philadelphia VA and a research associate in the University of Pennsylvania Perelman School of Medicine, recently spoke with Healthcare Innovation about the research, which was published in BMC Primary Care.
Healthcare Innovation: Could you talk about some of the reasons that behavioral health integration in primary care is such an important topic right now?
Khazanov: Yes, of course. We know that there are many individuals in this country who want and need access to mental healthcare, but it's very difficult to get it. The idea of collaborative care is that it's a way to provide mental healthcare services within a primary care practice. That means mental healthcare is easier to access for individuals who are just showing up to their regular primary care appointments, so those individuals don't need to go through the process of finding care in the community on their own. It's a way of making sure that they have easy access to the services that they need.
HCI: One of the things health system execs tell us is that there's an overall shortage of behavioral health providers. How do the health systems setting up these collaborative care models find enough providers to include in these team-based care arrangements?
Khazanov: I think there are a few ways that it is more of a sustainable system. Using Penn as an example, when a primary care provider wants to refer somebody to behavioral healthcare, they refer someone to a resource center, and based on that person's presentations or their symptoms and the disorders that they have and how severe they are, they are sent to a number of different options. They could just be referred to self-help resources or they can see the behavioral health practitioner within primary care, or they're referred out to the community. The 30% or so of individuals who are referred back to primary care to get behavioral healthcare — those are short sessions that are time-limited. Typically, it's about a 30-minute-long session every other week or every month. So that's part of the way that their resources are spread across as many patients as possible.
HCI: Does that require a different sort of team setup to triage people to the correct service?
Khazanov: Yes, typically there are a few different people involved. Of course, there is the primary care practitioner who's doing the initial session and evaluating if there's any reason for referral. There's also the mental health practitioner, usually that's a master’s-level practitioner who does evidence-based therapy. Within primary care, there's also a consulting psychiatrist. The Penn program is unique because it also has an intake triage and referral center. It has bachelor’s-level intake coordinators who ask questions over the phone to the patient about their symptoms, the issues that they're struggling with, and makes different suggestions for referral.
HCI: Has Penn Medicine increased the number of primary care offices that have behavioral health integrated?
Khazanov: We started in 2018 with eight practices, and now it's in over 35 practices across Penn Medicine. One thing I should note is that in the beginning, the system was operating at a loss financially, but because of the ability to bill for collaborative care and because of the way the system is set up, they were able to make it a viable program that is able to be disseminated more broadly.
HCI: Before talking about this paper and your focus on studying the impact on depression, anxiety and suicidal ideation, has Penn studied whether it is having the desired impact of more patients gaining access to mental health treatment and getting it more quickly than previously?
Khazanov: Yes, with a couple of caveats. This paper focuses on a particular subset of individuals, but there have been other papers published finding that it's really increased access to care. So quite a lot of individuals are being provided care, either within the practice itself or by getting support for referrals in the community — more than you would expect without this system. The caveat is that we don't really have a control, right? So both in this study and in other studies, we know what's happening for these patients within the system, but we don't have a good sense for how much people are accessing care outside of the system.
HCI: But have there been other controlled trials about the effectiveness of the collaborative care model on mental health outcomes, or even medical outcomes or total cost of care?
Khazanov: There are lots of studies showing that it's very effective, and most of those studies have focused on patients with mild to moderate symptoms. For those patients, studies have shown conclusively that it increases access to care, that you can improve symptoms, even with relatively brief periods of care.
HCI: From the paper you co-authored, it sounds like a lot of times people with more serious psychopathologies are sent outside the system to seek specialty care and sometimes they have trouble obtaining that. So this was asking: what if those people are actually treated within the collaborative care model? And maybe this is the first time that's been looked at…
Khazanov: Yes, exactly. So this is one of the first studies of a naturalistic community setting where individuals, especially with suicidal ideation, were kept within the collaborative care model and not only referred out. The reason we were able to look at that was because of this intake and resource center, which assesses patients and then also bases decisions on their symptoms, but doesn't exclude for suicidal ideation. It doesn't automatically say that everybody with suicidal ideation has to be referred out. So we had patients with some level of suicidal ideation that was not acute. They weren’t in an acute crisis. We found that those patients were able to be treated successfully within the model as well.
HCI: Could you briefly describe what the findings were?
Khazanov: The main finding of the paper was that individuals within the collaborative care model had improvements in their thoughts of suicide and their symptoms of depression and anxiety over the course of collaborative care. We also found that symptoms improved more with longer periods of care, but that was only up to six months. So it seemed like collaborative care is really useful, and the more treatment the better, within the constraint of six months.
HCI: The paper also found significant differences in decline in depression or anxiety across race, ethnicity and age. Was that surprising at all, or kind of expected? Could that be the basis of more research to understand why that is?
Khazanov: It would definitely be interesting to understand why that is. It does fit with some previous findings showing that collaborative care can be especially effective for minoritized populations, and that might be because there are potentially more barriers for those individuals to access mental healthcare — the stigma associated with it and the resources required to access that care. So when it's offered within this convenient and easily accessible system, it seems to be particularly effective. So it does kind of fit with those findings, but I think we do need to figure out exactly when that's helpful, and also how to scale up those systems so that it's able to increase access more broadly.
HCI: Do you think one impact of this paper could be more places that already using the collaborative care model would see more of these patients in that setting, rather than having them seen outside of it?
Khazanov: Yes. That is exactly what we were hoping to show. And although this isn't the only paper describing it, the model where individuals are being evaluated and being triaged and referred is particularly helpful for suicidal ideation, because you have that assessment. If somebody is at acute risk, they can get those services that they need. But if they're not at acute risk, they can just be treated like anybody else would be treated.
HCI: You mentioned earlier that the use of the collaborative care billing codes makes this make sense financially for Penn to do. But are there still barriers to health systems across the country adopting collaborative care models? Are we seeing it being taken up broadly? Or are there still cultural or other issues that might make adoption slower than we might want it to be?
Khazanov: There’s still a shortage of mental health providers. There's a lot of difficulty accessing care, even within this system. Ideally, it would be in many more practices. There are still pretty significant barriers just in terms of finding resources. There’s a lot of management and supervision and hiring that has to happen, so it's not implemented as widely as it could be.
HCI: Could other value-based care models reward this and nudge more health systems in this direction?
Khazanov: I think that's totally right, that the payment model could have this as a goal. Just increasing reimbursement for these types of services, and for the behavioral health practitioners, so that they're incentivized to have these roles within these practices is one of the biggest things that would be really helpful for this program and for other programs.
