The Impact of Depression and Its Treatment on Cardiovascular Disease

[00:00] Okay, thank you. So it's my great pleasure to introduce my friend and colleague, Dr. Bruce Rolman to Cardiology Grand Rounds. Dr. Rolman is the University of Pittsburgh Medical Center Endowed Professor of General Internal Medicine and Professor of Psychiatry, Health Policy and Management.

[00:20] and translational science and biomedical informatics in the division of general internal medicine. He's the founding director of the Center for Behavioral Health and Smart Technology and the Center for Behavioral Health Media and Technology, the goals of which are to link faculty from across Pittsburgh's

[00:40] schools of health sciences and the medical center, Carnegie Mellon University, and the local entrepreneurial community to promote new collaborations and to support the next generation of innovators who will create, test, and disseminate scalable solutions for some of the most common, challenging, and costly behavioral health problems.

[01:00] This also provides an academic home for talented, new, and junior faculty. He received his bachelor's degree in biology from the University of Pennsylvania and went on to receive his degree in medicine from Jefferson Medical College. He did Masters of Public Health and Epidemiology in Johns Hopkins University and

[01:20] at University of Maryland, Baltimore County, did his internship in residency with a fellowship in general internal medicine at Johns Hopkins. He's a fellow of the Academy of Behavioral Medicine Research and the American College of Physicians, and has served as the president of the American Psychosomatic Society, among many other honors. He has a

[01:40] over 100 peer-reviewed publications. Among them reports that the several NIH-funded randomized clinical trials testing the collaborative care model for addressing psychosocial and psychosocial multimorbidity in cardiac and related populations. His current research also includes a focus on the use of mobile technologies to reach

[02:00] serve rural medical populations. I welcome Dr. Rolven to Cardiology Grand Rounds. Thank you, Dr. Berg and Dr. Velazquez, for inviting me to Yale to present our research here.

[02:20] from the University of Pittsburgh. And also, when putting my talk together, I was impressed by how many Yale people, Yale faculty are involved and I'll be presenting some of their work as well in my presentation this morning. I have no disclosures. I'm here for hire. So I have no disclosures over here, no Speaker, Bureau, Zonorary,

[02:40] et cetera, all the work that I'll show you has been NIH funded. The goals of our presentation are really to raise awareness of the prevalence of impact of depression on patients with cardiovascular disease, describe the collaborative care model or chronic care model for treating depression that we've been using in Pittsburgh and in other places.

[03:00] and also to present key findings from trials to treat depression in patients with cardiovascular disease.

[03:20] When I give this talk each year, I try to update my slides from the heart and stroke statistical update. These costs are only rising. Probably the only thing dropping are the number of MIs and bypass surgeries, but the other heart failure, deaths, all these other numbers unfortunately continue to trend up.

[03:40] disease, however, it is very common in the two. Depression is often comorbid with patients with cardiovascular disease, and depending on how you measure the depression, whether it's with a formal interview or using a screening questionnaire like the PHQ, up to half of patients with cardiovascular disease.

[04:00] have comorbid depression or elevated mood symptoms. And when it's present, as I'll show you in data, it roughly doubles the risk of mortality, reduces health-related quality of life. And unfortunately, although they're available, a variety of safe and effective treatments often goes unrecognized and untreated by cardiologists.

[04:20] primary care physicians, and other clinicians. So there was a landmark article that came out, wow, it's been like 14 years ago, advising routine depression screening for patients who are post-MI, post-bypass surgery, and those with angina.

[04:40] I don't know how many people in this room are familiar with this that came out. Most of the time when I present this, people say that they had never heard of this report before. But we'll be going back to this. But this is something some cardiologists and researchers, Matt, had mentioned American Psychosomatic Society. There are a number of people

[05:00] who are co-authors of this paper, were very involved with the society. But my role here coming to Yale is kind of to introduce a topic to people who may be less familiar. This is one of my favorite quotes by Louis Pasteur about chance favoring the prepared mind.

[05:20] By the time I came to the University of Pittsburgh, Johns Hopkins had my fellowship on a practice interview. And at the time, my research interest had been on adherence with healthcare recommendations and medications in particular. But it was really through mentorship at the university and specifically

[05:40] in people in our department of psychiatry that I became interested in depression and its impact on reducing patients' adherence with medications and other recommended care. And between 1995 and 1998, I started getting involved in the field of depression in primary care.

[06:00] care. And one day I saw an article in the Pittsburgh Post-Gazette that women did not do as well after bypass surgery as men. And authors had speculated maybe it was because women had smaller blood vessels, etc., etc. And there was this article that came out from the Society of Thrasyptics

[06:20] surgeons, looking at a large registry over 340,000 cases, and they were looking at things like cross-clamp time and ejection fraction and all sorts of other characteristics. But there was nothing in there about depression. And I had a conversation with one of my mentors at Pitt, and I said to him,

[06:40] that's followed. I wonder if it was because of depression that women did worse after bypass surgery and by this time the chance favors are prepared by and by this time I was aware that women suffered depression and other affective disorders, anxiety and so on, and roughly twice the rate as men. But the authors had not done anything.

[07:00] if there was nothing in the dataset, in the SDS was not collecting information on depression. And so one of my mentors, Dr. Chip Reynolds in psychiatry, suggested I look into the issue, and then I became obsessed with it. And I found all sorts of really interesting papers that got me turned onto the field. This is a classic paper.

[07:20] by Dr. Nancy Fraser-Smith. It's probably been cited about 3,500 times in the literature, and I call it the ER paper. You are the ER paper of depression and heart disease. Dr. Fraser-Smith ended. She did something really very simple. Basically, they surveyed people

[07:40] people who survived acute MI at the Montreal Heart Institute with formal at the time DSM-3 criteria and to see who had formally had major depression versus who was not. And what they found, 16% of the group overall had met criteria for major depression, but in six

[08:00] months after survival of the acute MI, 17% of those patients were dead compared to just 3% of the people who were not depressed. And Dr. Fraser Smith and colleagues adjusted for a whole variety of characteristics that are kind of routine now in cardiovascular medicine, like smoking status, ejection, fraction.

[08:20] gender, etc., etc. And there are only three things that still held out in the adjusted multivariate, and that was still major depression. That was still roughly, you're four times more likely, more than four times more likely to die if you had comorbid major depression, calling an MI. But you look at this different

[08:40] difference over here is 17 and 3%, 14%. So as a young now assistant professor, I was thinking about this, wow, if we can cut that 14% in half to seven, that would be really big. And I thought about, you know, maybe if we could just tackle one disease, and this was it.

[09:00] was amazing. These curves in the 1990s reminded me of cancer curves, the HIV epidemic before antiretrovirals, and so on. It was huge in my mind. And it wasn't just post-MI. I mean, here's other work too by Francois L'Espreance and-

[09:20] Dr. Fraser-Smith in Montreal, looking at BEC depression inventory scores and mortality following MI. These curves reminded me of what you see like in cholesterol and survival. The higher the cholesterol level or HDL cholesterol, the lower the survival that you see, or the higher the blood pressure. This was a dose response.

[09:40] even at subclinical levels of depression. Here we see this in Ingrid Connery's work after bypass surgery, same kind of thing. If you're depressed, you're more likely to die, even after adjustment. Here's work from Wei Zhang at Duke, same idea with in

[10:00] failure. So across a variety of different, we can go on and on with papers like this, but across a variety of cardiac conditions, the more the presence of depression and the worse the depression, the lower the survival. This is work that we did here at the University of Pittsburgh Medical Center or

[10:20] or UPMC, where we screen people, I'll show you after with the PHQ2, it's just a two-item questionnaire that's easy. If you can remember the cage questionnaire is screened for alcoholism, it's easy to remember, this is even easier, it's only two items. And basically, if a patient said yes to one of the two items, these are patients with systolic

[10:40] heart failure or H ref. And if you screen positive 20% of those people were dead in a year compared to just 8%. So this wasn't just happening in Montreal and at Duke and other places. This was happening here in Pittsburgh. Even after multivariate, as I'll show you too here.

[11:00] it was still a significantly elevated risk of mortality from just those two simple questionnaires.

[11:20] patent the sun and Dr. Kronke is kind of a similar mindset. So this is basically, we'll talk about this in a moment, things that you can do here in New Haven in terms of the screening. So there are a couple of YAL investigators who were involved in this work too. This was a review article that recently came out.

[11:40] at in circulation and looking at depression and the effect estimates come from meta-analyses, as citations over here. But again, using more modern techniques, multivariate techniques, and so on, the association between depression and incident, MI, CHD, stroke, obesity, etc., is all there.

[12:00] So it's extremely, extremely important. Unfortunately, I can only scratch the surface of the field today, but I hope that there'll be a few people after the talk who might want to research the topic on their own and consider treatments, maybe new ones that have been developed, or ones that are available, like cardiac rehab, as a result.

[12:20] way of tackling effective disorders and improving quality of life and survival. Alison Gaffi had the pleasure of meeting her for dinner last night here at Yale, also has been working in this area too and showing the same thing with psychological distress, increasing cardiovascular disease.

[12:40] muscular morbidity and mortality. So I guess if we the bumper sticker, as we'd say, is depression kills. If you had to put this on the bumper sticker of your car or on your door or a button, depression kills. How it kills still yet.

[13:00] is up for debate. This is a review article that was in JAK a few years ago. There are a variety of both biologic mechanisms and also behavioral mediators by which depression might affect mortality. So like I mentioned before, we know that people who are depressed are less likely take their medications or less likely to quit.

[13:20] smoking or make other needed behavioral changes. There's been a variety of work also to show that people who are depressed have stickier platelets, they have autonomic dysfunction, higher blood pressure on average, so more inflammatory cytokines, and so on. So there may be a variety of mechanisms by which depression affects

[13:40] morbidity and mortality, probably no single one, which is probably, we'll talk about this later, why we haven't yet seen like depression treatment reducing mortality. But maybe some people here in this room or in their offices here in New Haven are working on these problems right now. I'd be very interested.

[14:00] and but certainly an active area of investigation. So in terms of things to take away from and things that you could do here on the wards today or tomorrow, begin by asking about depression. So this is from the circulation article I showed earlier.

[14:20] They had this kind of complicated scheme over here. We'll get to this in a moment, but they settled on using the pHQ2 followed by the pHQ9. So this is a pHQ2. So basically in the past two weeks, have you had little interest or pleasure doing things or felt down to pressed or hopeless? Some people will score this on a

[14:40] on a zero to three scale, we just did a yes, no. If you answered yes to either question, then we considered it a positive. So here's the sensitivity and specificity of this. So it's a very sensitive screen for depression, but the specificity is relatively low.

[15:00] And so that guideline paper from 2008 recommends a two-stage screening process following up with the two-item PHQ2 with a PHQ9. And this is what we've done also in our research studies to confirm that we have people who truly are depressed. So here's a sensitivity and specificity of the

[15:20] the pH Q nine. So in cardiac patients, you probably don't want to do the pH Q nine first, because you're going to miss a lot a lot of the patients sensitivity is only about half. But, but if it is positive, if it is positive then because of the high specificity, then the person has elevated mood symptoms.

[15:40] So then the next question becomes, will treating depression improve cardiovascular outcomes? And if so, which ones? This is also, we had dinner last night with Kim Smoldren here from Yale. And this was a paper that she recently published.

[16:00] looking at the triumph registry. And here's a suggestion that if you have depression and don't treat it, then you have significantly higher incidence of mortality after acute MI than if you do treat. Although this is not randomized trial data and there is also concern

[16:20] that maybe the patients are different. Those who are treated might be at higher quality hospitals, they might be higher socioeconomic status, and a variety of other reasons. So it's hard, it's hard. The randomized, this is not a randomized trial to say for sure, but it's a clue that perhaps treating depression might make a difference on recall.

[16:40] reducing mortality to that of people without depression. So these are a couple of the studies and some of them, Dr. Berg has been intimately involved with like the Enriched Trial and COPES study. And so these are some of the brand new studies.

[17:00] name studies, randomized trials looking at treating depression. I didn't even include our work on there, but there's probably maybe another eight or nine studies like this. But I'll go over a couple of the major studies. So I mentioned Dr. Fraser Smith first, who had the initial highly cited

[17:20] Well, she was funded in the 1990s in Canada to do a randomized trial of treating patients who survived acute MI in Montreal, and they were randomized to either a nurse-led intervention for psychological distress. They would screen the patients each month after MI and if they were psychological.

[17:40] psychologically distressed, they would make house visits and follow-up calls and so on. And here, this is very interesting. They found that there was no difference on mortality between the intervention or usual care, but women did worse. Women who were randomized to the intervention

[18:00] had higher mortality than those who are not. So that was really interesting. And why would that be? I don't know. I mean, maybe one thought is that since I've talked to Nancy about this and she can't exactly explain why that happened, maybe it's women don't like taking advice from other women. I'll show you later on also gender differences in this.

[18:20] in treatment effects, but it's unclear, but it probably points to the need if you're doing research to power these trials on a two-sided P value rather than one. This is the EnRICH trial, which at the time is and probably still is the largest randomized trial

[18:40] done for treating depression in cardiac patients. This was stopped for futility. This was an NHLBI funded trial. It was over $40 million done at multiple sites around the country. And you can see over here there was absolutely no difference in mortality between the intervention and usual care.

[19:00] intention was treating for psychological distress. If you were very highly symptomatic, then people were getting sertraline or Zoloft is a brand name. But there was absolutely no difference. People have published a number of studies looking at subset analyses, but in the intent to treat. There were also some issues, and we could talk about this later.

[19:20] about the delivering intervention. There was relatively, many people dropped out of the intervention that didn't follow through, and that's certainly a concern with doing these large randomized trials. At the time the EnRICH trial was going on, Pfizer had funded the SADHAR trial, which was using serous RNA.

[19:40] surgery, post-MI. And you can see over here, this is, was considered kind of a qualified success because of the composite endpoint had a trend reduction in mortality. But there were other issues with this. As we know with antidepressants, no single antidepressant works for everybody. And you know, any, any

[20:00] SSRI, for example, might only be about 60 to 70% effective at treating depression. So often involves a combination of medications. Other issues here was the study was underpowered, had a goal of about 600 people. And I think they got finished up with 379 or something. So this is considered a qualified

[20:20] Sandy Glassman followed up the survival of these patients for seven years and found no difference in mortality. So that's also interesting to see. Would the treatment make a difference? Nothing.

[20:40] and this was using medications, also sertraline, in people with heart failure. No difference not only in mood symptoms, which I'm showing here, but in the main outcomes paper, there was also no difference in survival from giving heart failure patients sertraline or placebo.

[21:00] did another trial of heart failure patients using acetylchopram, also known as Lexapro. Again, no difference in survival, just giving people a medication. So it's very different than like if you had done a beta blocker versus placebo or statin or something else, ACE inhibitor, there's absolutely no difference here. So

[21:20] We know that treating depression is going to take more than just giving people a pill. So what are some of the limitations of what we'll call these conventional trials? As I mentioned before, there was a single antidepressant used in some of them. Patients' preferences for treatment were not incorporated, so they may have dropped out if they were randomized.

[21:40] medication. Say for example, we didn't want to take medication, didn't want to participate in therapy. The screening and the intervention itself may not have been generalizable. How do people advertise for patients? Was it systematic screening or were people responding to ads in a bus or something? Some of them had brief

[22:00] follow-up. So as I showed like with the SADHART trial, the initial follow-up period is only six months, few clinical events, low study power, and just generally lack general sensibility typical practice settings. So around this time in the late 1990s, by this time, I

[22:20] was funded on an R01 to use a collaborative care child to treat anxiety and primary care. So it was familiar with this kind of model and the primary care model was developed in Seattle in the early-mid-90s by Ed Wagner and others and this was, we know in the United States that we're

[22:40] good at treating acute issues like infectious diseases or surgical conditions and so on, but we don't do as physicians and healthcare assistants don't do as good a job of treating chronic medical illness and this could be heart failure, hypertension, COPD, as well as mental health concerns. So Dr. Wagner and others have

[23:00] and developed this chronic care model that was linked to primary care, it uses a team approach where you have care managers who follow evidence-based guidelines and use registries to proactively follow up with patients between visits. So you're not just waiting for a patient come into the office and so on, using these registries.

[23:20] might see somebody who's maybe non-adherent or has high blood pressure or high cholesterol, and then reaching out to those people and seeing how they're doing, and then also connecting them to community supports, promoting self-care, getting leadership buy-in at an organization, and so on. And by the late 90s, early

[23:40] There was sufficient evidence to show that this model worked in a variety of chronic conditions, like heart failure and COPD and diabetes care and people in Seattle and other places around the country, including Pittsburgh, were beginning to use this model to treat mental health conditions. But nobody yet had used this model to treat.

[24:00] depression. So collaborative care in the depression realm, there have probably been over 80 trials that meta-analyses have looked at. The effect size improvement is considered small to medium at like 0.3. So not huge, but it was roughly, it was steady. Some trials have shown more.

[24:20] some had done less. But a couple key things is that you can leverage the help of mental health specialists because they were kind of like coaching care managers and providing specialty backup. And these so as a result, a psychiatrist could potentially consult on dozens of patients in an hour.

[24:40] rather than seeing them one at a time. And it can also speed access to mental health care. So in our model's collaborative care, we're doing these over the phone, but in some of our recent studies, we've also done this over the internet and using like the Epic Systems. They call it my UPMC. I don't know if they, what's your name?

[25:00] if they have a similar term here, myyel or something else maybe, to communicate with patients using chat and other secured messaging. More recently, collaborative care has become billable. So about two years ago, CMS had issued billing to

[25:20] codes for collaborative care. So this has always been the challenge of how do you pay for the care manager? And in some systems where you have an integrated healthcare delivery system like a VA or Kaiser or Intermountain, it may be easier to do than some places that are feed for service where how do you build for the care manager's time and the

[25:40] time and so on. Our team is conducted over the last 20 years, our team has conducted five R01 clinical trials of collaborative care. All of them were positive trials, two of them, Bypassing the Glues and Hopeful Heart, we'll talk about a little bit more in depth. So we had a lot of the experience of doing

[26:00] kind of studies. So the first was the bypassing the blues trial. So I showed you that work before. I was in the room when people at the American Psychostematic Society in 2002 in Barcelona, Sanjart and Enrich both presented the main outcomes of their of their of their

[26:20] trials, and you could actually like feel the air come out of the room. Dr. Berg was there, I think probably the only other person in this room, but you could kind of feel the air come out of the room. There was a lot of hope and excitement for these studies, but it didn't turn out as hope for. On a personal level, I had submitted an R01.

[26:40] to NHLBI after working about three years in the field collecting pilot data, getting the experiences, as well as doing the anxiety trial. I submitted the grant, I think it was in June of 2001 and in the fall of 2001, it came back unscored and of course it was like.

[27:00] tremendously disappointed. And I came to the March 2002 meeting to hear the main outcomes of Enriched in Sanhart because I just had to be in the room to hear what happened rather than to hear it secondhand. And I resubmitted our

[27:20] grant in June of 2002 and then it received a fundable score that fall. And I think part of the reason why is because the traditional models, the medication model and the other kind of models that were used, counseling models, it was a negative trial and I think NHLBI wanted to keep the field going.

[27:40] and looking for a new model, and that was bypassing the blues. So at the time when I started giving this talk, it used to be over 600,000 bypass surgeries were performed in the United States annually. Now the numbers are down I think because of stenting and statins and better care and so on. But anyway.

[28:00] Similarly, when you looked at CABG patients, whether it was pre-op or post-op, roughly a quarter of those patients met criteria for depression or elevated mood symptoms. And those patients tended to have delayed recovery from bypass, increased rate of readmission at cardiovascular events and mortality. So the specific aims of the

[28:20] bypassing the Blues trial was to see if collaborative care for depression could improve health-related quality of life that was our primary outcome, physical functioning, mood symptoms, healthcare re-utilization, and healthcare costs. So it was quite ambitious. This was the basic study design. So

[28:40] Again, this is from 2002. And before the work had come out with the PHQ, we had proposed doing the PHQ2 screen in the hospital because we needed to get consent also. And then we would call patients up two weeks later. We didn't want to consider people depressed in the hospital.

[29:00] after the bypass because who wouldn't have tired or little energy or troubled sleeping if you just had your chest cracked open. So we waited until people got home and we restrean them with a PHQ-9 two weeks later and then they were randomized to either the collaborative care for depression or to their doctor's usual care.

[29:20] And then the other innovative design that we had was a non-depressed comparison group, because we knew some of these patients would get readmitted, some would die, you know, they would generate healthcare costs. So I thought also if this was a negative trial, like some of the others I mentioned, that we would at least be able to compare the depressed and the non-depressed contemporary.

[29:40] contemporaneously recruited comparison cohort. We were up and running in seven Pittsburgh area hospitals. This was one of the posters that we used. It was also our first study to make a logo. We're quite proud of. We had care managers. We hired nurse care managers who did the recruitment.

[30:00] They also interacted with patients over the phone to assess treatment preferences, impart self-management skills, promote adherence and adjust pharmacotherapy in response to the PHQ-9 scores, and also communicate with the primary care physicians, you know, either by fax and telephone, you know, depending on the urgency.

[30:20] We didn't give out any free samples of any medications. If we recommended an antidepressant, patients had to pay for it on their own. And if we couldn't get people well, we referred people to a community mental health professional. Pharmacotherapy is SSRIs largely and then avoided.

[30:40] do harm avoiding tricyclics and trying to get people off of benzos where we could. We had a weekly case review that the nurse care managers attended in a room like this, and we would project on the wall case registry screens like this one.

[31:00] And we would have a case discussion around the room with the study psychiatrist, Dr. Chip Reynolds, project coordinator, and myself and our study nurses. And you can see over here like the collaborative care model talked about registries. This was our registry. So we could see the PHQ score over the time, medications, what the treatment recommendations were and so on.

[31:20] so on and make appropriate recommendations. We told the primary care doctors, so this was the other innovation too. The other studies I showed, none of them involved the primary care physician or were kind of tied in. So how would you make this generalizable? So these were effectiveness trials where we sent PCP patients.

[31:40] our treatment recommendations, they knew which arm our study their patients were in. We made recommendations for pharmacotherapy, but ultimately they were the ones to prescribe. So long story short, our nurses, roughly over a three-year period, approached over 3,000 bypass patients.

[32:00] patients, about 80% agreed to sign consent and completed the PHQ, just over half screened positive as kind of expected. And then you can see over here about a third of patients, just under a third, scored about 10 or above when we were able to telephone them two weeks after hospital discharge, and we randomized 3.0.

[32:20] Our goal was 300 and we did just above, as well as our non-depressed comparison cohort. I want to point out here we oversampled non-whites and that's in the non-depressed group and that's why there's a difference between the depressed and non-depressed by race.

[32:40] it's largely the same. This is the main outcomes from our paper, our main outcomes paper we published a number of years ago. We saw an effect size improvement of 0.3. This is measured between the yellow intervention line and the usual care. So you can see even with treatment,

[33:00] we couldn't get people to the level of health-related quality of life as a non-depressed comparison group. That was a top line. So they certainly significantly improved beginning at two months. We saw an early and sustained improvement, but not to the level of the non-depressed, kind of pointing out that we may need better treatments for depression.

[33:20] Here's the other, looking at the mood symptoms. So you can see also significant improvements in mood. The non-depressed group is at the bottom. This is an impact by gender. Remember Dr. Fraser-Smith's M-HART trial that showed people, women did worse.

[33:40] I believe found the same thing where women tended to do not as well. And here we could not say in the JAMA paper that women didn't do as well because confidence, the intervals overlap. But you can see the effect size improvements were different between men and women. So it looked like men tended to do better.

[34:00] than women and by effects on the main measures over here. So this is part of our secondary analysis, looking at readmissions. So the very, you can see over here, the pink line is very similar to the blue line or the non-depressed.

[34:20] There was a trend that intervention men, where the effect size was stronger, were less likely to be hospitalized. These are the number of care manager contacts. So you can see at eight months, it was roughly a little bit more.

[34:40] than a phone call a month, a little bit more, but it varied depending on how much engagement people were interested in, but the numbers between men and women had similar numbers of calls. Pharmacotherapy, so you can see over here, roughly a quarter of the patients were on an antidepressant.

[35:00] at baseline and then it went up to about a little over 40% over time. But it wasn't like we were just throwing people on pills. You know, you can see over here, look at two months, it's 36% in the usual care and then kind of drop back. And that may be because of the screening and notification that we did and so people started the medication but didn't necessarily.

[35:20] necessarily stay on it if they're randomized, but I don't think that we can attribute our improvement to just being, you know, putting people on pills. So here was the other thing that we did. It took about five years to do this part of the study, and these were looking at the claims costs. We got Medicare claims, Blue Cross Blue Shield in Pittsburgh.

[35:40] called Highmark and UPMC Health Plan claims on roughly half of the patients' continuous claims. And you can see over here the cost differences between the two. It was not significant, but it was a median. The intervention patients had a median of about $2,000 less.

[36:00] When we do these bootstrapping things, our study statisticians and healthcare economists, Julie Donahoe did, you can see over here the incremental cost effectiveness ratio is about negative $10,000 per additional quality-adjusted life year. So this is highly unusual. Most things we do in healthcare.

[36:20] cost more to get better outcomes. I'll show you a second in some cardiology studies, but you can see also here repeating the study statistically about a thousand times, we show that it's most likely cost reducing and quality increasing where it falls in that bottom right quadrant.

[36:40] So, putting this into perspective with other studies, nearly negative 10,000 is really just incredible. I mean, there are only a few things in healthcare that we do have a negative quality adjusted life years, like prenatal care and a couple of vaccines. Not much else. And you can see at the top here, it's 10,000.

[37:00] half of this slide are the costs, maybe these costs should be higher these days, but it's probably still relatively similar. It's highly cost-effective if you use the $20,000 standard of highly cost-effective interventions, but $50,000 is more typical and

[37:20] Some people have talked about moving up to 75,000 because of healthcare inflation. But certainly, treating depression alone is highly cost-effective. But if we put this in perspective of things cardiologists and primary care doctors routinely do, like screening lipids in middle-aged men, treating

[37:40] type 2 diabetes, I mean, it looks like a bargain. This is some long time follow up. One of the medical residents working with us presented these analyses at the ACC a couple years ago and you can see over here there was a significant

[38:00] increased in the depressed versus non-depressed patients even after adjustment over time. So it's kind of similar to Ingrid Connery's work that I showed earlier, but we did not find any differences by treatment assignment. That is, the intervention people in our bypassing the BLUES trial did not have

[38:20] lower mortality than the usual care patients. Nor do we see anything different. This is in men where we had the strongest effect size, but we didn't see any differences in either the overall or in women. So the final things to kind of come out of bypassing the blues is that this telephone delivered

[38:40] collaborative care, could improve health-related quality of life, that was our primary outcome, improve physical function, reduce mood symptoms, it also reduced post-cabbage pain, which I there was in a secondary article by Natalia Morone, and also reduced health care costs.

[39:00] or trends that was highly cost effective. So I always feel like, you know, there's so many papers that come out every week in the literature. And I was looking at the email, Dr. Berg sent me last night of all the papers that just came out of Yale Cardiology this past week. It's a very impressive number.

[39:20] But I always feel that too many studies just sit on the shelf and people don't read them. So I really think about what can we do to kind of get our work out there and that people use it. I'm really delighted that our work was also cited in other scientific statements from American Heart Association.

[39:40] And also going back in 2014, we had a colleague in thoracic surgery at University of Pittsburgh, who since moved on to University of West Virginia, but he was involved with the SDS. And now depression is one of the things that are abstract in the SDS dataset.

[40:00] So maybe there's a young Bruce Ralman out there who's going to look at this someday and repeat those analyses and this time have depression as a co-variant. We were also mentioned and cited in other scientific statements too about secondary prevention after

[40:20] bypass surgery. So I mean, this is like, you really want to see your work get out there and kind of get promoted. It's a big deal. When we get like a random email from some some other researchers who want our data to include it into in a meta analysis or, or something, it really is important as a way of getting your work out there and hopefully making a different

[40:40] We also work teamed with UPMC Health Plan. We did a couple years back, we did a depression initiative with their care managers where we trained them on a checklist, ABCC, DEFF. And very simply, we trained the health plans care managers when they called patients up

[41:00] after MI and went home to screen for these kinds of things. I got the idea from Atul Kowande's checklist book. And over time, it wasn't a randomized trial, but the health plan found that in the first year, it saved nearly $4 million of costs.

[41:20] So going back to another question, would depression treatment improve cardiovascular outcomes? We never really answered it. Did I need the studies that we did, I did, or presented? So our work started to move in the mid-00s to heart failure. At the time, I was interested in bypass because you knew when someone had a bypass surgery.

[41:40] and you can pursue depression afterwards. But heart failure to me was always kind of a vague condition. There was no starting date. It might be at a hospital admission date, but there was no formal starting date. And also, were we going with preserved ejection, fraction? Were we going with reduced ejection, fraction? A lot of people, when I tend on the inside,

[42:00] patient service were labeled with heart failure, but they didn't actually have heart failure. Maybe they had diastolic dysfunction, maybe they had COPD, but they didn't have it. So my attention started turning to this in the mid, late 00s. And we did the pilot work that led to our hopeful heart trial.

[42:20] It was published last year. So similar to depression and other cardiac conditions, very prevalent in heart failure patients. And I was also really, really intrigued because heart failure was the, while it was growing in incidence, unlike bypass, which was coming down, survival, survival hadn't much

[42:40] chain for small changes in survival over the last 10, 15 years. But it was not dramatic like post-MI survival, post-cabbage survival, and so on. So there was still a relatively high rate of mortality that you could do one of these randomized trials that you didn't need thousands and thousands of people.

[43:00] showed mortality differences if they existed. So my thinking was what if we can apply collaborative care to heart failure care? So that led to hopeful heart, but hopeful heart also had some other innovations too. My late colleague,

[43:20] colleague, Wayne Caton at University of Washington, Seattle, had presented some really nice, or not presented, he led some really elegant collaborative care trials in the OOs. This one, the Pathways trial, was treating depression in patients with diabetes. And he found over here that it did not make a difference.

[43:40] It was a small improvement on mood symptoms in this trial. It was like 0.2, 0.21, but there was no difference on A1C. And he followed it up also with what was called the team care trial. And the team care was the first blended collaborative care trial. And in this case, the blend was depression.

[44:00] infection, and also diabetes care. And he found that he could train up a nurse to do this together. They had a roughly like a point drop in A1C. Systolic blood pressure and LDL cholesterol all significantly dropped, as well as the mood symptoms. Mood symptom improvements. It was like a

[44:20] 0.67, which considered a moderate to large effect size improvement. Really, really wonderful results. They were published in New England Journal about 12 years ago. They've been cited now over a thousand times. And this is another landmark paper. But the idea was, and one of the things I didn't like about our bypass study is that we

[44:40] we just focused on the depression, how could that be implemented? Or were we calling for a separate siloed, another team of care managers through the mental health care? Could we train up a single nurse care manager to do heart failure care as well as depression care? And that was really one of the other innovations in helpful heart.

[45:00] blended collaborative care for depression and heart failure. So we had similar aims as in bypassing the blues, a similar design, but this time it was another innovation. I learned from another colleague at APS, Dr. Ken Friedland had published an influential article on control groups in

[45:20] behavioral health trials. And Dr. Friedland had talked about having a need to have a potent attention control group. So in many of our other studies, these 80 trials of collaborative care, including my own, had always used a usual care comparison group, not a true

[45:40] tension control. So unlike a team care didn't have this either. So they didn't have a separate arm that was just diabetes care management, you know, it was the it was the blend versus usual care. hopeful heart had a third arm in the that people were randomized to and that was

[46:00] care for the heart failure alone. So we wanted to have a potent attention control group, as well as the blended, and we retained a usual care control group so that we can compare to other collaborative care arms. Sounds a little complicated. But I'll show you the results over here. So also over roughly a three year period, our nursing

[46:20] has approached nearly 12,000 patients. You can see here roughly half screen positive in the hospital where we got sign consent and roughly a third screen positive on the pH Q 9 about two weeks after they got home. Again, we didn't want to call people to press in the hospital because if you had heart failure, you might be having trouble.

[46:40] sleeping, low energy, and we didn't want to confuse that. Here are the associate demographics. Again, a slight difference in race because of oversampling. But the other thing I want to point out that's really key, look at the ejection fraction. These were all patients with who had an EF of 45% or lower.

[47:00] to be eligible for the trial, but the depressed and non-depressed, you can't attribute, other studies have reported this too, you can't attribute the low mood to a low ejection fraction. That's the point I'd like to make here and the rates of comorbidities, diabetes, hypertension, post-MI were similar between the depressed and non-depressed.

[47:20] As expected, mood, health-related quality, function, and function were lower, but these were self-reported. Look at New York Heart Association class III and class IV. So people self-reported their health is worse if they were depressed. And I think that's really important for cardiologists to recognize it wasn't the ejection.

[47:40] fraction. It wasn't the comorbidities, it was the comorbid depression. So similar 12-month intervention. Here we had the blended, similar case reviews, weekly case reviews. But I'll cut up because the hour is late to the main outcomes. This is what we reported the very

[48:00] top line over here is a non-depressed comparison group. And our primary outcome measure over here was the SF-12MCS, or mental health composite score. And there was, again, a significant improvement in the blended versus the usual care, but you can see the EUC

[48:20] or collaborative care for heart failure alone, and the blended were very similar. Those are the middle two lines. These are the point differences by gender. Again, we powered our child to look at within gender differences because of our

[48:40] earlier work and MHeart and Enriched and others. So we could see over here blended versus usual care. In this case, it looks like women did better in bypassing the blues or trends that men did better. These are the effect sides as far as plots are being

[49:00] exercise differences. Again, the main outcome measure was the SF12 MCS, but there were not any differences. You can see in the Kansas City cardiomyopathy questionnaire or the SF12 physical function.

[49:20] looked at mood symptoms, we found significant improvements in the blended over enhanced usual care, which is the other thing that was really important to show that collaborative care for depression is really more than just calling people on the phone and seeing how they're doing. You know, because the blended arm

[49:40] had a significantly improved effect size improvement of mood symptoms over the collaborative care for heart failure alone. So we knew our intervention was doing something. And again, this was something that no other collaborative care trial had had this kind of potent attention control. Readmissions.

[50:00] Today we didn't find any differences. We're still extending this follow-up as well as for mortality, but there were no differences either by baseline depression status. Roughly about 80% of people were hospitalized. It's extremely high rate. We also presented in the e-apendix data looking at the per person per month.

[50:20] re-hospitalization rates, because some people were hospitalized multiple times, and a capital admirer fails to capture that. Again, there were no differences, either by baseline, impression status, or treatment assignment. Same with mortality. So we are continuing to follow up with mortality.

[50:40] working with a fellow back at the University of Pittsburgh and we hope to present the five year mortality rates. The other thing that's very new, we just submitted this this week to present at a Society for General Internal Medicine this year. We got the insurance claims. This time we did it in three years.

[51:00] We got the Medicare, the UPMC health plan, and the Highmark Blue Cross Blue Shield, this roughly on 80% of the patients who were involved in the trial. You can see the mean and median costs for outpatient. Inpatient is really dominating the costs over here and you can see the totals.

[51:20] The trends are in the right direction, but we're still analyzing the data. These are typically you see claims costs being highly skewed and we had a couple, the top 1% of people were over $2.7 million into 12 months after index admission for heart failure.

[51:40] Now why are they so expensive? We found roughly about 12 people had had heart transplants, there were LVADs and there were other things. So we think that we need to look at the data a little more, cost data a little more carefully because treating depression probably isn't going to make a difference on whether or not you get on the transplant list and so on.

[52:00] Stay tuned. I'm also really pleased and I give a shout out to Matt who's involved, Matt Burdock is involved in these trials through two European replication trials going on right now and and working with colleagues over there. Kristen Hermann-Lingan also is an APS former president.

[52:20] He's leading the TEACH trial and the ESCAPE trial. He's involved with the team at ESCAPE, which is a multinational European trial, looking at blended collaborative care for people with multiple cardiac conditions and the other was for people with heart failure. So it'll be interesting.

[52:40] and you see also how what their results are in a different healthcare system and state this is I guess we're a few years from hearing that. So kind of summarized this morning's talk, depression is prevalent and predictive of poorer outcomes for patients with cardiovascular disease. Treatment

[53:00] improve quality of life, mood and pain, and it's highly cost effective. But it's unclear if treatment actually reduces cardiovascular morbidity and mortality. People have proposed really large studies like in the thousands, but these are like, you know, tens of millions of dollars to do these studies now.

[53:20] I don't know if they'll ever get done. Maybe it might just be meta-analyses. But that's no reason not to screen for and to treat depression in your cardiac patients. We do a lot of things in healthcare like back surgery and shoulder surgery, cataract surgery, you know, that probably doesn't do anything for a long life, but it's expensive and definitely

[53:40] improves quality of life. So I think we need to think about that and as a cardiologist too, if your patients are coming in with heart failure symptoms and they're and they're and you're providing the best evidence-based guidelines but they're still really symptomatic, you know, don't wait for depression to be considered as a diagnosis of exclusion after everything else is working.

[54:00] out, it's really common and maybe for those patients in particular, it's really important to screen for depression who are still symptomatic despite your best efforts and objective measures like ejection, fraction, echocardiograms, cardiac casts, and so on. So please do consider depression

[54:20] as part of the treatment. But still I would say that more effective treatments are needed. So in our, I showed you the checklist and part of the checklist was E for exercise or cardiac rehab. I think cardiac rehab in most places have done a really good job in recognizing the impact of depression.

[54:40] And also because of the exercise is another effective treatment that you could do. Not everyone needs a pill and for people with just mild symptoms, it might be just, you know, counseling and exercise makes a difference. I kind of wanted to summarize this. I showed you this slide earlier from the 2008.

[55:00] Lickstein paper. But basically if you could just remember this, screen your patients with a pH Q2 down to press or hopeless, tired of her little energy most days over the past two weeks. If they say yes to one of those, then consider the pH Q9. If you can't find the pH Q9.

[55:20] just go Google pHQ 9 PDF and there are a lot of beautifully formatted PDFs of the pHQ 9 that you can administer. If the score is less than a 10 you could just stop. Okay so they probably, I showed you the sensitivity and specificity so with this kind of two-step method if it's less than 10 you could just

[55:40] stop. If it's 10 to 14, then it could be watchful waiting or maybe consider treatment if the patient had a prior history of depression. So maybe they were depressed five years ago or something, but if they never had depression in their lives, then it might be just something to reevaluate or put in your consult note to the primary care physician or recommend.

[56:00] cardiac rehab or something else would stay physically active. But if it's over 15, that's considered a moderate to high score, then I think it would be key to be treatment. And you might want to talk about medications, chiefly SSRIs and SNRIs or buproprion, counseling, maybe even.

[56:20] if people have complicated psychosocial issues, maybe they should see a mental health specialist, and of course, exercise, which is generally always a good deal to recommend in cardiac patients. I just also wanted to say thank you. This was our research team when we were at the peak of doing Hopeful Heart as well as the online treatment trial for mood and anxiety.

[56:40] none of the artwork in Pittsburgh, none of this would have been possible without having a terrific team and in the middle is Dr. B. Belknap who's involved with the TEACH trial and that's going on right now in Europe and you know come to our website to see more. Thank you very much.

[57:00] Questions? So, Bruce, thank you for a great talk. I think my question is, you know, having been kind of following this field, there was some time and being involved with something to pull out. You know, some people you mentioned along the way of my time at Duke.

[57:20] I've always been struck by the fact that it is very hard to define a dose response of treatment per se. I wonder if you can elaborate on whether you think that relates to the fact that there's depression treatment in general.

[57:40] And you said that should be tailored is not necessarily going to be. Is that going to reduce pressure in all people in the equal amount? So I mean how do you integrate that information into further studies of pressure? You have to imagine that.

[58:00] Sure. Well, some of it is also like what patients are willing to do. So cardiac patients may be on seven to 10 medications or more, and they don't want to take a pill. And that's okay. I mean, like certainly my patients with cardiac disease, I try to twist their arm for a statin if they're opaque or intolerant or an ACE inhibitor or something.

[58:20] not everyone wants a medication or will benefit from medication, but chiefly like exercise, socialization, really kind of like just basic things. Counseling if they're willing to do that. I think there's also new options that have opened up with telehealth that they can get.

[58:40] access to a mental health counselor, but they have to really want it and I think also to see where people are in their condition. Eryx is some of the other things I do like if I have a diabetic patient with peripheral neuropathy then I might think about like duloxetine or something like that that has FDA approval for peripheral neuropathy and also five or

[59:00] myalgia. So I try to work with people based on what are the indications for a medication, but really just seeing where they are. Some people are difficult and you're not going to get them to respond and they don't want to see a mental health specialist. It's definitely a challenge.

[59:20] Dr. Jim B challenge.

[59:40] those. And then the more recent trials, the collaborative care model, the LEND collaborative care are really focusing on some of the factors that are associated with depression. So it's not just that we're here's this patient who has multiple morbidity and that includes depression, but they're having difficulty managing their multiple morbidities.

[01:00:00] And so part of what collaborative care does is it supports them in that management. And so that addresses some of the behavioral factors that are associated with depression, and it may be that focusing there really starts to be legal, and by relieving the patient of some of that burden can affect depression.

[01:00:20] The other thing is the single women in Montreal and in Rich, a follow-up analysis in the Montreal study revealed that among women, those who responded to the treatment within the first two weeks had better survive

[01:00:40] than any other group in the trial. And of course it's a secondary analysis, it's not a pre-planned analysis, but it speaks to if you're going to do a treatment, you better be targeting what is the operative variable. And if you're not, you can actually make that.

[01:01:00] worse. A fair amount of these express women and the nurse is showing up and the nurse is not being helpful. One can imagine that they are more distressed after the nurse leaves because they were counting on the nurse to help. And that then also reminds me of the work of the old

[01:01:20] younger, within the distribution of patients, both semi, the women are younger and are more likely to have social determinants of health, living in poverty, abusive relationships, and things of that kind. And it makes me wonder whether the depressed women who are coming into

[01:01:40] Yeah, are also part of what's driving that depression. Are these other factors that it's like, I've got this horrific situation at home and you're trying to treat my depression. Yeah, stay with that idea. That's that's that's actually and Chris Callahan wrote a really nice editorial to our may not our

[01:02:00] Hopeful Heart Main Outcomes paper and it's where also my work has been going also in recent years, thinking about these social determinants. So I see patients in our clinic and there's a lot of, you know, we hear about it, but there's a lot of income inequality. There's a lot of people who can't afford these medications or like especially like diabetes and some of the newer GLP.

[01:02:20] ones, SGLT2s and whatever that are on patent right now. And we have this also in our case discussions when we would talk to patients, you know, on our case reviews, not talk directly, but care managers would say they can't afford this, they can't afford the copays to go to or the transportation or take time off to go to cardiac rehab.

[01:02:40] So I mean, these are like big issues. Like we know cardiac rehab works and, and we're eating healthy. What's eating healthy for you or for me? So these are really, really big issues. I think to me, I think this is going to be, I see in the Biden administration is trying to address some of these issues. Last year they had the child support.

[01:03:00] payments and others. I see it also in value-based health designs where you might have copays wave for getting a statin or something. But I think this is going to be a big issue. I'm really curious in part to see how the, well, I'm curious about how the European trials turn out, but in Europe they tend to have less in common in a

[01:03:20] quality issues than we have in the United States. So it's going to be really interesting to see how those studies work out. I think we're somewhat limited because of people can afford their meds or couldn't take. We freely recommended cardiac rehab or other things, but they couldn't take time off from work. And these get into like, being

[01:03:40] Beyond our work as a clinician, we're in this box, but now these are bigger societal things where policy changes need to be done if we're to maximize the potential benefits from our effective treatments. Thank you. Thanks so much for joining us on that. I'm sure if people have questions, they can reach out to us.

[01:04:00] that moment and probably didn't have a better story. There's a lot of people online and I checked. So thanks everyone for joining. Thank you for inviting me.