Could a Mental Health Coordinated Care Plan Be Applied to Plaque Psoriasis?

Supplements and Featured Publications, Profiles in Care: Importance of a Coordinated Health Care Plan in Plaque Psoriasis and Psoriatic Arthritis Management,

A Q&A With William B. Leasure, MD, and Mark Williams, MD, of Mayo Clinic

Mayo Clinic is a nonprofit health care organization with major campuses in Rochester, Minnesota; Phoenix, Arizona; and Jacksonville, Florida. William Leasure, MD, a psychiatrist and division chair of Integrated Behavioral Health at Mayo Clinic, discusses the collaborative care model used in mental health at Mayo Clinic that ensures care coordination between primary care and psychiatry. The model employs an evidence-based, patient-centered team care approach that leverages nurse care coordinators to achieve patient goals and improve treatment outcomes. This program has increased patient and provider satisfaction, improved depression symptoms, and increased the speed of improvement in depression for those participating.

AJMC®: What do you think makes for a good case management program or a good collaborative care model?

Leasure: There are a few core principles that are defined that have to be present, to be consistent with the collaborative care model. I can run through those as a starting point in terms of what’s defined and then we can expand from there if needed. Dr Mark Williams has been working with the collaborative care model much longer than I have, and is an expert, both nationally and internationally. So I have invited him to share his expertise as well.

Williams: I’d like to define a few of the words that we use. Case management, care management, collaborative care are all words that you’ll hear us using. It’s important to know that they come from a source or a definition or characteristics. Some people use case management broadly, and you have to look at what they’re using it for. Sometimes that means for your life you have this person who connects with you or until you move out of the state or whatever. You often hear the word care management linked to an episode of care. Until you get stabilized for your diabetes or until you get stabilized for your whatever condition, you have this link, then you graduate from it with an outcome. Then you’ll hear this collaborative care concept which is really coming out of how should primary care and mental health collaborate, but it can be more broadly used to think about how should any specialty and primary care collaborate. If you’re going to go to the literature, it’s important to think about what are you looking for and what’s the scientific basis around that term because people use these terms sometimes interchangeably, but they come from a different heritage.

Leasure: That’s great. That’s one of the things I was going to mention: the collaborative care model does highlight this partnership between primary care and mental health. There are other disease management models out there, but this one is focused on that connection between the primary care teams and the mental health teams, and joining them together deliberately in ways that have not always traditionally occurred. I’ll run through the core principles of the collaborative care model quickly, and we can jump in with any questions if needed. Those principles are patient-centered team care, so really keeping the patient’s goals, and what they’re looking at wanting to accomplish, at the center, but also working deliberately as a team, and centering processes on ways to facilitate that. Population-based care: using a registry to track patients over time so that patients aren’t falling through the cracks. You’re looking at an entire population, not just whoever shows up in front of you in the clinic that day. Measurement-based treatment to target: in mental health, for decades, there weren’t necessarily clear measures to target, but that’s really changed in the past couple of decades and one of the core principles is really measuring the care. You can imagine in diabetes or hypertension there are clear things to measure with lab values or blood pressure. Sometimes it’s a little bit harder to define in mental health but using some type of measure to assess your outcomes and then treating to a target. For depression, in our case, really trying to treat to the point of remission of depression, not just a general level of improvement. Evidence-based care, which speaks for itself, but reminds us to really stick to the evidence. Accountable care: having some type of accountability built into the model. That can vary depending on the state or the payer, but there are growing payers that are reimbursing for quality of care and clinical outcomes, not just the number of patients that you’re churning through or, in a fee for service model, the number of patients that you’re seeing in a clinic.

AJMC®: At your organization, how do you ensure that care is coordinated between these multiple stakeholders? And where are case managers involved in that?

Leasure: I will talk through how the model works. I think that will answer the question. Our model is focused on major depressive disorder. First, we screen for depression in the primary care setting, so patients are identified as screening positive for depression and then further assessment occurs to help determine if they do have a diagnosis of depression. If they are identified as having moderate or greater severity of depression, then they’re eligible for our depression care coordination program. We use the PHQ [Patient Health Questionnaire]-9, which is a very commonly used measure of depression and can be used to track improvement over time. The primary care providers can refer patients that they identify, but we also have an eligibility list where we identify patients who are primary care patients who have a PHQ-9 score in the moderate or severe range and also have a diagnosis of depression. We can proactively identify those and then approach the primary care provider and the patient to see if they’re interested in the program as well. Once they are referred, then the care coordinator reaches out to the patient to gather information similar to what we would get if we were seeing the patient in a psychiatry appointment in our office. Then we meet with the care coordinator on a weekly basis to review those new patients that they would have enrolled and their whole panel of patients to make recommendations on treatment, review where things are with the treatment, look at those measures like the PHQ-9, talk with the care coordinator about how things are progressing or not progressing in terms of goals, and work through any barriers that might be there to help move forward to reach those goals and help get the depression in remission. We use a registry to ensure we are reviewing those patients on a regular basis and really continuing to make recommendations to try to move the treatment forward. That’s one of the features of this model—that it really helps accelerate treatment changes and make more treatment changes over time to overcome the inertia that can happen in clinical practice. In a practice without collaborative care, we might see someone and then 3 months later see them for a follow-up. Maybe they stopped the medication after 2 weeks because they were having adverse effects, but we did not realize that because we had not had contact with them for quite some time. Maybe they remained at a low dose because it had been a while since we saw them or their primary care provider saw them. The idea with this model is that we’re preventing them from falling through the cracks, but also helping to make treatment adjustments to get them to those treatment goals.

Williams: The only thing I’d mention is at Mayo, William and I are currently involved in two programs, one for adolescents with depression and one for adults with depression. We’ve talked about adding other mental health conditions, but there is a care coordination program for complex adults at Mayo as well run by some other folks, and we have, in the past, tried to mix diabetes, cardiovascular disease, and depression as a model for care coordination. However, our current programs are mostly about … depression in adults and kids. The principles should be largely the same, but I don’t want to speak about things we’re not doing.

AJMC®: And what’s the background of the care coordinators? Are they nurses? Do they have a health care background? Is it more of an administrative role?

Williams: There are about 90 randomized control trials about this using people with different licensures, so it’s pretty hard to answer. For us, it’s RNs [registered nurses] that have mental health background or medical background, and whichever background they have, we try to supplement it with the other. If their background is in primary care, that’s great because they know how to work in primary care, but they might not know what to do with a suicidal patient. For those that have a mental health background, they don’t know how to operate in primary care, so it’s kind of a learning process there. This model is supported by Medicare billing codes. They don’t require it to be an RN in that role. In some places, the role is a social worker, and it has been tested with psychologists with non-RN level nurses in that role, but Mayo has chosen to use RNs.

AJMC®: So then after they go to the care coordinator, then it’s kind of a handoff to the psychiatrists, and then they handle all the treatment? How’s primary care?

Leasure: A key feature is that it’s not a handoff to the psychiatrist, but instead an opportunity for the psychiatrist to take a broad look at what is going on and then make recommendations for treatment. However, the final decisions and the treatment are driven by the primary care provider and the patient. They make the final call. The psychiatrist is serving in sort of an advising or consulting role. We don’t meet with the patient directly, at least as a core part of the model. There are times when there’s a level of complexity or a patient may not be getting better over time and we will recommend an appointment with a psychiatrist to help clarify or see if there are things that come out of that appointment that might not have come out in working with the care coordinator. But in most cases, the care coordinator is working directly with the patient, and the psychiatrist is meeting weekly with the care coordinator to review the progress and what’s going on, reviewing measures like the PHQ-9 and others, and then making recommendations about medication, psychotherapy, [and] other, maybe social things, that might be helpful— behavioral activation goals, other things, to try to help support the improvement. Particularly for medication recommendations—those would go back to the primary care provider for them to make the final decision on if they were comfortable with that, if that was something that they felt was a good direction to go based on their experience and relationship with the patient. In most cases, they do proceed with that [treatment plan], but there are times when that may not happen.

AJMC®: What methods do you use to coordinate care between the psychiatrist and other specialties?

The recommendations are entered as a clinical note in the EMR, which is used for all clinical care throughout our health system. The care coordinator will discuss recommendations with the patient and usually sends them electronically through the EMR to the primary care provider (PCP). If there are questions, the PCP may reach back out to the care coordinator electronically or by phone, or do the same with the psychiatrist. We try to foster back and forth communication to build the relationship between the primary care providers and the care managers and psychiatrists.

AJMC®: The care coordinator is doing the reassessment and a lot of the communication with the patient. What’s the extent of the care coordinator’s role throughout the process?

Leasure: They begin by ensuring the patient is interested in the program and answering questions they might have about it. Then they proceed with what we call the enrollment appointment or interview where they meet with the patient in person or by phone to gather information and understand what the patient is dealing with. Like I mentioned, the information gathered is similar to the things that we would gather if we were seeing the patient for a consult, including past psychiatric history, previous episodes of depression, previous other diagnoses, psychotropic medications they have tried in the past, experience with psychotherapy, psychiatric hospitalizations, suicide attempts, self-harm, other treatments they might have had, social history, family history, and really getting a good handle on what’s going on with the patient at this point; … then come to the meeting with the psychiatrist, review those things, and then make recommendations. They continue to have regular contact with the patient ranging anywhere from about 1 to 4 weeks depending on how things are going, what kind of progress somebody’s making. That occurs over the next several months unless the patient gets into remission and in that case, they would be doing things like working on relapse prevention, reinforcing things that had helped get the patient into remission to help keep them there over time. Throughout that process, they are reassessing what’s going on in terms of the PHQ-9 and the [General Anxiety Disorder-7] which is an anxiety measure. They are working with the patient on developing behavioral activation goals, which are a component of cognitive behavioral therapy proven to help in depression. Developing those goals, checking in with them over time about how things are going with those goals, and, if they’re having trouble with them, then adjusting those. They also convey the recommendations that the psychiatrist or the primary care provider might have made, and then work through any other challenges that might have come up along the way.

AJMC®: And then would they evaluate for the need of escalation of therapy and then send over the recommendation for the primary care physician?

Leasure: It’s intended to be a partnership. We meet every week with the care coordinator to review what’s going on with the patients and they might themselves identify a need or recommendation and, honestly, that’s often what the case is, that they identify something the patient is struggling with and might say, “I think doing some cognitive behavioral therapy or trauma-focused psychotherapy might be helpful to work on this trauma-related issue. I think a referral to a social worker might be beneficial because there are social needs there that aren’t addressed.” There’s a whole range of other recommendations that might result, but care coordinators are able to offer those based on their experience in working with the patient in collaboration with the psychiatrist. Then those recommendations are summarized in a note in the medical record, which then goes on to the primary care provider.

AJMC®: Is all this documented in the electronic medical record?

Leasure: Exactly. Part of this model stems from there just not being enough psychiatrists to see everybody directly. This is really a way to extend the reach of the psychiatrist both to areas where they can’t physically be there, but also where they are not able to see 50 patients in a 2-hour period, which is what we usually do in terms of reviewing the patients that a care coordinator would have. They usually carry 40 to 50 patients, and so we will review their caseload on a weekly basis. Obviously, we can’t see that many patients in that period of time.

AJMC®: Then how are you coordinating with any providers they might be seeing outside of Mayo Clinic?

Leasure: That one can be a challenge at times. There are a number of patients that are seeing therapists in the community, and what we will try to do, one, is talk to the patient about things they might be working on in therapy, what their goals of therapy might be, sometimes make some suggestions based on themes that are coming up. We often try to get a release of information so the care coordinator can connect with the therapist. That’s sometimes able to happen. Sometimes it’s more difficult because of the therapist’s schedule. Also, sometimes there are barriers to getting the release of information. The ideal situation is that we can include outside providers so that we can make sure we’re continuing to function as a team rather than we’re one little team and then there’s somebody else or other people outside of that.

AJMC®: How do you determine which measures get escalated? Are they predetermined? Can you talk a little bit about that process? What determines which diagnosis codes get brought to your attention by case management?

Williams: The model that we’re describing comes out of a model called IMPACT, from the University of Washington, and it’s basically designed around that. What we tend to focus on at this stage is mostly clinical; we’re not as good at risk-avoidance [in] terms of cost. We’re just not that sophisticated. We don’t have utilization data, for example. What we focus on is things like suicidality and active substance use. We will notice if somebody’s got a worsening medical condition that we might want to [address], but we’re laser-focused on improving their depression. If somebody has a chronic other medical condition, arthritis for example, if it’s interfering with their sleep and their quality of life and enjoyment, we will focus on it, but if it’s not, we might not pay much attention to it. We will look at social issues: if they’re becoming homeless, if there’s abuse in the home. It’s all around depression and things that impact depression. You can think about how all kinds of things might impact depression: job loss, abusive spouse. All kinds of things that might impact depression might be the highlight of our focus because our goal is to get this person to where either they are not depressed or whether they can manage their level of depression sort of like chronic pain. There are people that have chronic pain [where] you can’t eliminate their pain, but can they keep working, functioning, manage, have access to resources? Those are our two outcomes.

AJMC®: Within your model, what do you think are still some challenges and unmet needs that you’re working on? And how are you addressing those?

Williams: I can think of a few, and I’m sure William can think of a few too. One of the challenges is it’s easier in a large health care system to implement something than it is to tweak it because of committee structures and permissions and cost. It gets a little bit frozen, and that’s a challenging thing. I feel that we could do better about fidelity measurement [the extent to which delivery of an intervention adheres to the protocol or program model originally developed] as a regular part of these programs. There is an evidence base around outcomes, and if you don’t do certain things, some of the things that Dr Leasure mentioned, then you may not get those outcomes. How do you measure in a large system with clinics spread all over that people are sticking to the critical parts of the model and they can innovate around the others depending on their site? Then how do you add other conditions? Primary care providers need help with depression, but they also need help with bipolar illness, schizophrenia, or substance use disorders. How do you add those in and make sure that your team is ready to manage those conditions when there’s not a lot of resources? One of the biggest challenges in rural clinics is, if you measure for something and you find it and there’s nowhere to send that person or nowhere to manage it, then you don’t always want to look for something. But, of course, it’s still there, and you’ve got to deal with it. And then these patients have other things other than mental health problems. They clearly have medical problems. They clearly have social determinant issues. It’s a little tricky to know how to make a more consistent link with the medical teams about an arthritis and depression thing. You would need access to maybe a rheumatologist, for example. Or social determinants—it’s ideal if you’re linked to social supports in the community so that you can help a person who just can’t get in anywhere. Can you find them a place to get the help that they need? So often, we’re struggling to get someone linked anywhere in the health system for whatever their issues are. If we had better community links wherever our sites are, that would help. But all that takes a lot of time to set up.

Leasure: Another challenge is ongoing training and education for the team members. There’s the initial part to get everyone on a level when say you’re starting up a new program or when you bring someone in, but then there’s the ongoing part, both with clinical education, but also improving the program, the tweaks that Dr Williams mentioned. Then, even if we have a wide range of experience of our care coordinators, so someone might have been a psychiatric nurse for 20 years and now been doing this for 10 years; they have a huge amount of experience. Then we may bring in a new RN who has never worked specifically in psychiatry before but maybe has some primary care experience or other medical experience. How do we get their skills up to a minimally sufficient level but then create a curriculum or ongoing training that meets their needs but also isn’t too low for the more experienced person or vice versa? Threading that needle is challenging. It …takes time to be able to do that for our care coordinators and our RNs, but also [for] the primary care providers. Another challenge is finding the time and the resources to get everyone connected on a regular basis rather than just the individual pairing of the care coordinator and the psychiatrist—trying to create a larger team so we can learn from each other. There are some models for that. You might be familiar with the ECHO [Extension for Community Healthcare Outcomes] model, out of the University of New Mexico, which is a way of disseminating information amongst larger groups of people. We’ve explored doing something like that with this and other things for spreading more specialized mental health expertise but have not gotten into fully developing that yet. The other thing would be just ongoing studying of our outcomes and using those to continue to make tweaks. Having a real time dashboard of looking at our outcomes and then being able to use those to make adjustments, study best practices, is ideal. Then we could ask questions like, “Is one particular site or care coordinator, or care coordinator, psychiatrist pairing, continually doing better or worse than another? What can we do to try to improve the outcomes of that group?”

AJMC®: Does the information go back from the patient? I understand drilling patient information down to the patient. What about going back the other way? Are there mechanisms for the patient whether with their portal, or their telehealth, or something where the patient can go back to the care coordinator, care management person? Is there any live interaction, either virtual or through the phone?

Leasure: Yes, the ongoing contacts are usually by phone. What they try to do is set up a time for the next contact to occur, let’s just say in 2 weeks, and that would be a check-in to repeat the measures, review progress during that time, talk about barriers or potential treatment challenges, and then talk about any recommendations that might have been made. That’s real time conversation. Sometimes that happens through the portal as well. You can imagine that sometimes communication through electronic communication [is] not quite as rich as a live conversation, but even aside from that, the patient can reach out at any time to the care coordinator, either by phone or by a portal message.

AJMC®: And then what outcomes have you been looking at, and what improvements have you seen since you implemented the program?

Williams: Well, the original outcome was remission at 6 months or response, and that’s defined with the PHQ-9 in mind. Our state has a way to measure that, and it’s required across all primary care sites. We originally were trying to work on that, and we still do look at that, but with Epic coming, we have not had an easy way to, through our data set, compare response and remission outcomes by site, so that we can see who’s doing the best and what can we learn from them. It’s on our wish list. We also can look at things like referral rates. We didn’t formally look at this, but prior to care coordination, many patients out of primary care were going into our specialty psychiatry clinic to be seen because primary care didn’t know where else to send them and their families. That really decreased dramatically when we started doing this kind of work. Now, we were doing other things as well, so it’s a little hard to know cause and effect purely, but that’s been shown outside of our practice that when you develop more of a net to manage a problem in primary care, it changes the escalation to specialty care to those that really need that level rather than those that might be managed in primary care. Otherwise, it’s a challenge when the primary care team just doesn’t have the background or the feeling of being backed up. We find that the primary care feeling like, “Oh, you’ll be around if I get into trouble with this new medicine. OK, then maybe I’ll go ahead and prescribe it. Before that, I [didn’t] prescribe any of those types of medicines,” kind of thing. We do need more specialty care or more targeted need for specialty care, might be a better way to put it. We have been trying to learn how this works with [the emergency department] and hospital utilization, but it’s so hard to do that in regular practice because the utilization data is all in the insurance company’s hands, where they have all the bills, unless you look at only utilization in your system. The clinical data is in the clinical hands, and it’s been hard to marry the two together in a way so that we can track that on a regular basis. We have a lot of patient and provider satisfaction every time we’ve measured it. Patients and providers have had really good satisfaction. When we measure clinical outcomes like remission rates, and we’ve probably done over 40 different retrospective projects for different things, “What about this? What about age group? What about people with that problem, and that?” Every time, those that are in care coordination have a better clinical outcome on depression symptoms than they do in regular practice. We also measured the speed of improvement. How quickly did people get into remission? We found that it took longer than you think, like 2 to 3 months to get into remission, which seems like a long time if you’re depressed, obviously. It was much longer outside of care coordination, more like 6 months or something, so those are some of the different outcomes we’ve looked at. Of course, from the billing perspective, now that there’s a CMS billing code and the private insurances are starting to pay for it, some folks are looking at efficiencies, and how much are your costs covered as well. And that’s a challenge because each insurance company might have a different discount rate with whichever clinic you go to, and then some patients aren’t insured, some patients are. That’s a little tricky, but the goal is to try to be as cost effective as you can.

AJMC®: How did you originally identify depression, or mental health [as an] area that you wanted to do a care coordination plan, or program?

Williams: Well, being psychiatrists, we’re a little bit bent that way, of course, but in the beginning, in Minnesota, [it was] through the Institute for Clinical Systems Improvement [ICSI]. It’s a [quality improvement] organization in the state of Minnesota. They had some funding from the major insurance companies in the state to look at depression. Everyone agreed that depression wasn’t well managed in primary care and that there was a model with lots of strong evidence that it did work, and so, why don’t we implement that? And they originally had some reimbursement. The 6 major insurance companies used some of their administrative funds to pay a reimbursement for any clinic that did the training. I was on that steering committee for the Mayo group, but it was all across the state. We focus on depression because it’s among the most common conditions in primary care, so it’s easy to find patients that have that and it’s very disabling. It’s on the top 3 disabling conditions or top 5, depending on how you look at it, for the world; it is a disability, and it’s very treatable. Now, we also focused on anxiety, and we’re thinking about expanding into bipolar illness, and we tried that briefly, and we’re thinking of going back to that and substance use. Each of those, each condition has its own, “yes, buts,” in terms of what you expect from a person when they’re better, and what resources do you need. Can someone be released from care? As I’ve talked about before, is it a chronic thing? Someone in a course of life-threatening cancer, how do you decide you’re done with their care because he goes into hospice, right? Somebody with a condition that could, based on studies, get into remission in a certain time, and you have an outcome that’s relevant; that’s how you think about what you choose. With arthritis, you’d have to think about those kinds of things. What would you measure as your outcome because you probably always will have some pain and some movement problems, but what would be a sign that you’re ready to graduate from your model? And you can do that with depression.

AJMC®: The people that graduate and they have remission, are they sort of always flagged as potential entrance back, or they’re just taken out of the care coordination model?

Williams: Our registry, it’s like a little moving window. It’s only focusing on those who are in care coordination. We have another Epic tool that we can see people who become eligible for care coordination, and the trouble, like arthritis and depression, [it] is a chronic illness, and it’s very common for people to have another episode; I mean, a subset of people will have one episode of that with one episode of depression. About 50% of the time, you’ll have a second episode. I guess patients who graduate from our program, we hope that we’ve taught them relapse prevention. “How did you get depressed, and what was the sign that you were depressed?” We’ve tried these 2 things in therapy. “What really worked for you to get better? If you continue those and you get in trouble, what if you were to go back to that therapist that you really got along with,” or “Notice when you get your sleep out of control; you really got to watch your sleep.” We try our best to put ourselves out of business with having them not necessarily need[ing] to come back.

AJMC®: And then, how do you see care management changing for these patients in the future?

Williams: Well, I guess it’s a little bit of a crystal ball type deal. Care management has been studied in multiple ways. The payers have tried care management where you have your insurance payer calling you, “I’m your care manager from X,” or whatever. You have a care manager for each medical condition in some places. In one health system you might have 4 care managers calling you for your 4 conditions. Our national health care system is still largely tied into fee for service, which makes it harder to know what to do when you’re billing the government, or the payer for each type of care management system when it’s one person. I guess one question is will we go to value-based care, and if we do, does that mean that we can approach a person with all of their conditions rather than say, “I know you have that other stuff. I’m focusing on this because I’m your diabetes care manager. I don’t want to know about that other stuff unless it relates to diabetes.” How do you organize your health care system around that and still have outcomes and follow up? In mental health, the provider shortage is not going to get better. Some have estimated we need 4 times the current mental health people to even meet the demand in the US, and that’s not going to happen. I’m seeing more and more spread of care management for mental health around the country. I’m involved in some training in Michigan, for example. Because these primary care sites, they can’t wait for mental health to show up, they’re going to have to figure out how to manage their population, and mental health has such an impact on all of those other illnesses. My guess is we’ll see a lot more spread of this model into primary care settings across the country, and then of course it’ll be up to us as the mental health providers to up our game and start managing the things that are even harder for the primary care to manage (addiction, schizophrenia, and bipolar), and figure out how to do that. They’ve done that in Washington, only in these FQHC, Federal Qualified Health Care Centers, but it hasn’t been broadly done in many other places. There are publications on care management, for PTSD, for chronic pain, for bipolar disorder, for anxiety, for depression. The data is out there, but it’s just really hard to get the resources pulled together and experiment and try to get that to happen without a grant or something like that. I think we need to expand care coordination as part of our help in primary care, because we won’t just have enough mental health providers otherwise.

One of the pieces of this model is that you have to get your hands dirty a little bit, and so, sometimes as a specialty area, we have been approached, for example, by [gastroenterology] because a lot of the people with irritable bowel syndrome, for example, have mental health problems.
—Mark Williams, MD

AJMC®: We use the term “quantify risk” a lot of times in terms of stratifying where these potential high-risk patients are, and that we’re quantifying the risk of. When you think about cardiovascular disease, sort of easy. High blood sugar, diabetes, triglycerides are out of control, cholesterol is out of control, blood pressure. In a disease state like yours, it’s a lot more qualitative. Are you able to use tools the same way from a risk stratification perspective? There’s no way of quantifying severity in the plaque psoriasis world currently. How do you quantify risk in a depression or anxiety patient? How do you enter in severity?

Williams: Good question, and it’s in the eye of the beholder, right? The person who is interested in risk from the perspective of utilization might do it slightly differently than the person who’s interested in risk in the sense of suicide, or in severity of the illness. Right now, I’m doing a little machine learning project on depressed patients presenting into primary care to figure out who develops high utilization of the ED. Some of the things that stand out, of course, are things like looking in the rear-view mirror. Has this person needed to be hospitalized before? Is this person presenting with the highest severity of illness? What about comorbid alcohol use or drug use? Are they employed or are they struggling to find work? That’ll affect their insurance, and that’ll affect their access to health care and medications. Have they failed treatment? There’s a definition of treatment-resistant depression, and it’s often defined as failing previous antidepressants. Have they already failed two previous antidepressants? They’re harder to improve. Have they had previous suicide attempts? What you do when we are assessing someone as we look back; is this somebody who’s had a lot of mental health comorbidities? Are they currently using substances? Are they isolated? Have they been in the hospital? Have they had a suicide attempt? Do they have a strong family history? Those are all things that we think about as more challenging to treat depression, versus someone who’s had their first episode of depression after a loss of somebody, and they’ve never had it before and no comorbidities. Those are some of the things we use, but whether they can quantify them all is a little tricky.

Leasure: And then there are severity ranges for the PHQ-9 as well, that can help estimate severity.

AJMC®: Have there been other specialties reaching out to you as an example because they see improved outcomes with your program?

Williams: Yeah, there have been, and it’s interesting. One of the pieces of this model is that you have to get your hands dirty a little bit, and so, sometimes a specialty area, we have been approached, for example, by [gastroenterology] because a lot of the people with irritable bowel syndrome, for example, have mental health problems. What I talked about is care management for those folks for the mental health component of their problems, but this would mean that the psychiatrist would be advising you, and then you would need to actually write some antidepressant prescriptions or follow up on what’s happening. That was a little challenging because there was less of an interest in taking an active role in management. The primary care area has asked us in rural areas, “Could you take on some of these more complicated patients?” let’s say schizophrenic. The challenge with schizophrenia is it’s a very debilitating illness, and you need long term psychiatric care. If you take on the real serious illnesses, do you end up not being able to afford your care managers because you fill them up, and then you can’t get more care managers because you’re only allotted a few. It’s a balancing act between those that never discharge, and those that the primary care team needs help for. What we’re going to have to figure out, I think, is ways to step up and have better relationships with the effort of step care. You start with this level of care, and you step up those that need more help. The thing is, not all “stepped up” areas are ready for a stepped-up possibility. Average length of time to get into a psychiatric appointment can be 3 months or 2 to 4 months. That’s just a long, long time if you’re on the cusp of losing your job with bipolar illness, for example. You can’t wait that long. What we have to do is figure out how to have a stabilization process with a better option to step up the care for those folks in order to handle more things, which is a little challenging.

AJMC®: What would be your advice to another organization that wanted to do a similar program?

Williams: Well, if they were interested in the depression point to start with, I think it’s an ideal place to start because there’s so much written out there about how to do this already. You don’t have to make it up that much, and there’s a billing code now through CMS to bill for it. The billing system reinforces some of the main components of the model, in that you can’t get the money unless you do these components, which is helpful in that regard, and then there are some groups around that will help you get started depending on where you are. In Michigan, this group called MICSI, it’s like ICSI with an M, because Michigan, [and they are] doing that with Blue Cross. Blue Cross has decided this is so important, they’re actually funding training across their whole state to implement this model in primary care clinics. Then there’s a place called the AIMS Center through the University of Washington. They set up this nonprofit organization that can coach a primary care group to get started on something like this. Depression has a lot more pathways set up to do it. Doing the other types of conditions, there is some information out there, but it’s not as easy just to have a recipe. The thing that always is important as implementation science would say, “Even though you have a recipe, you got to know what’s really critical in that recipe,” because everybody’s going to say, “That sounds good.” Here we have this kind of patient or here we have these doctors, and they’ll want to tweak it having someone who knows about, “Now you can change that, but don’t change this if you want to maintain the outcomes.”