Transcript
DEBORAH KRAT: Welcome, everybody, to the National Institute of Mental Health’s Practice-Based Suicide Prevention Research Center’s overview meeting for today. Just a few housekeeping notes. Participants have entered in listen-only mode. Cameras off and mics are muted. Please submit your questions via the Q&A box at any time during the webinar. Questions will be answered, time permitting, after each center presentation. If you have technical difficulties hearing or viewing the webinar, please note these in the Q&A box, and our technicians will work to fix the problem. You can also send an email to me, Deborah Krat, at dkrat@mn-e.com. And with that, I will now pass over to Joel Sherrill. Joel?
JOEL SHERRILL: Thank you, Deborah. Welcome, everyone, and thank you for joining us today. I’m Joel Sherrill. I’m the Deputy Director of the NIMH Division of Services and Intervention Research, and it’s my pleasure to welcome you all for this overview of the NIMH-supported Practice-based Suicide Prevention Research Centers. In a moment, we’ll start today with welcome messages from Dr. Joshua Gordon, the NIMH Director, and Shelli Avenevoli, the NIMH Deputy Director. And then I’ll provide a brief description of the goals of our research centers program. Following that, we’re very pleased to have the directors of these practice-based research centers join us to provide an overview of their centers and the work that they’re pursuing. But first, we have the messages from Dr. Gordon and then Dr. Avenevoli.
JOSHUA GORDON: Hello, and welcome to this National Institute of Mental Health webinar on the NIMH Practice-based Suicide Prevention Research Centers. I’m Dr. Joshua Gordon, the Director of NIMH. As all of you know, suicide is a significant public health concern. It’s also been a priority of mine since arriving here at NIMH in 2016. I saw advancing suicide research as an important immediate need that had to be prioritized. Thanks to the efforts of you and many others, the suicide prevention research field is robust. In fact, NIMH investments in suicide research have tripled, rising from $41 million in 2016 to $125 million in 2022.
This research portfolio has resulted in a number of effective approaches to suicide prevention. Unfortunately, however, the use of these practices lags behind the evidence. We could do so much better if more of these evidence-based practices were implemented in more places. Accordingly, in 2020, NIMH launched the Practice-based Suicide Prevention Research Centers in order to speed the translation of research into practice. These centers are funded to achieve outcomes that can’t be achieved through our usual grant mechanisms. You’ll hear today how these centers are working to get life-saving knowledge into communities, many of whom experience disproportionate suicide risk. I’d like to thank the NIMH staff that guided the development of the center goals, organized this meeting, and helped with the center’s progress and oversight. Thank you also to the many center directors and their teams who are doing this critical and often challenging work. And thank you to everyone who’s joined the webinar today to learn more about these important efforts.
SHELLI AVENEVOLI: So good afternoon, everyone. Oh, sorry, Joel, did I interrupt?
JOEL SHERRILL: I was just going to invite you to welcome us. Thanks, Shelli. Sorry.
SHELLI AVENEVOLI: Yeah, I was trying to be streamlined there. So good afternoon, everyone. It’s really my pleasure to be here, extend another warm welcome to you, and thank you so much for spending your time with us today. It’s really great to see some familiar faces and names in the crowd. So, this year has truly been a year of reflection for NIMH. As you know, we’re celebrating our 75th anniversary, which really gives us the opportunity to celebrate achievements in mental health research with all of you and the broader NIMH community. But it also gives us an opportunity to look forward and think about what we might accomplish in the future.
As many of you also may know, we’re preparing for the transition of Dr. Gordon. And he mentioned in his opening remarks just a moment ago that suicide prevention has been one of his top priorities and that the practice-based suicide prevention centers has really been one of our early vanguard projects in this space. But as we prepare for his departure, Dr. Gordon’s departure, I really wanted to be here today just to let you know that suicide prevention will remain a high priority for the National Institute of Mental Health. And also wanted to take advantage of this opportunity to highlight that the time is right to really consider how we might move forward. And that to me means this is perfect timing for this meeting. And we really look forward to hearing your insights over the next two days. We want to learn from you about how we might better translate findings into practice, as Josh mentioned. We want to make sure we can ensure a health equity lens in our research activities, enhance our diversity of perspective in this work as well. We want to hear from you also about the challenges you face.
And then, most importantly, about the opportunities we might embrace moving forward. So, I don’t want to take more of your time. We have a busy agenda. It looks wonderful. I, too, want to extend a big thank you to our staff for organizing this event.And also another thank you to all of you for devoting your time to this activity. Thank you.
JOEL SHERRILL: Thank you, Dr. Avenevoli. On behalf of our NIMH colleagues, we really want to thank you and Dr. Gordon for your ongoing commitment and support of suicide prevention research here at NIMH. Okay. So, as I mentioned, next I’ll provide some background about the Center Research Program. In 2020, NIMH launched the Practice-based Suicide Prevention Research Centers, which were modeled after the Advanced Laboratories for Accelerating the Reach and Impact of Treatments for Youth and Adults with Mental Illness, otherwise known as the ALACRITY Research Centers Program. These Practice-based Suicide Prevention Research Centers are integrated transdisciplinary research programs that are aimed at developing, refining, and testing effective and scalable approaches for reducing suicide rates in the United States. Next slide, please.
The overall goal of the Center’s program is to support high-impact practice-based research with near-term potential to both significantly advance NIMH’s suicide prevention research priorities, and importantly, to ultimately achieve the National Action Alliance for Suicide Prevention goals of reducing the rate of suicide in the United States. These centers are intended to represent incubators for practice-relevant research and to support rapid refinement and testing of effective and scalable approaches for intervening at key intercepts in the chain of care, for identifying high-risk youth, for promoting continuity across key care transitions, and for intervening, including prevention, treatment, and for addressing crisis and acute risk. Next slide, please.
As you listen to today’s presentation, you’ll hear how the centers incorporate features to enhance transition into practice. They emphasize practice-based research that leverages clinical practice settings as incubators for new research ideas and as laboratories for nimbly refining and testing interventions and service delivery strategies. They use deployment-focused approaches that consider the perspectives of key end users, service users, family members, providers, administrators, payers, and other policymakers to yield interventions and service strategies that are relevant and can be more rapidly integrated in the practice. And they incorporate transdisciplinary research teams that broaden traditional clinical and mental health services collaborations to include new insights and new technologies, new research methods from experts in areas such as data science, digital health, decision science, and health systems engineering to advance both clinical research and ultimately clinical practice. Next slide, please.
In their presentations today, the directors will be describing their center’s focus, their practice partners and settings, their interdisciplinary research teams, and their research projects. Like other NIMH centers, these centers are essentially comprised of cores and research projects. The Administrative Core provides oversight to facilitate innovative and nimble transdisciplinary research to establish practice partnerships and to engage end users, to encourage and support additional research and pilot studies consistent with the center mission, to support research opportunities for trainees and for new collaborating scholars, and to lead dissemination efforts. The Methods Core assists by serving as an incubator function to facilitate the development and provide support to the center projects. But beyond that, the Methods Core serves to incorporate additional transdisciplinary expertise and to generate new methods and new directions. And today, you’ll also hear center directors describe their larger scale signature research projects, their exploratory research projects, and additional seed projects that are consistent with their center’s focus. Next slide, please.
NIMH supports seven centers that are listed here on this slide, and we’ll be hearing from each of these centers. Today, you’ll hear how collectively the centers address suicide prevention across ages and across a range of groups who are at risk for suicide. And you’ll hear that across these centers, the research takes place in a partnership with a variety of settings where suicide risk might be identified and addressed, such as in specialty mental health care and mental health clinics, in primary care practices and networks, in emergency departments, and in other parts of health systems, and beyond that, in other systems and settings that serve individuals at risk for suicide in the community and, for example, in criminal justice settings. At this point, I’ll turn things over to my colleague, Dr. Jane Pearson, who’s special advisor to the Director on Suicide Research at NIMH, and she’ll introduce the centers and moderate the center overview presentations. With that, Jane.
JANE PEARSON: Thanks, Joel. It’s my pleasure to be a part of this webinar today, so many of us can hear about what all these great centers are up to. So, we will ask the directors of the centers to present a brief overview, like about 15 minutes, to give us a sense of what their centers are covering. And if there’s five minutes after each of those 15 minutes, we’ll take some questions, mostly from the other center directors, to see if we can address some outstanding questions there. It’d be great if the center directors could use their raise hand function, and then I can track some of those questions then. I’d encourage the viewing audience, if you have questions, to go ahead and put it in the Q&A, and I’ll try to track those. And if we’ve got time at the end, which I think we will, we’re off to a good start today, we can also try to address those at the end. You’ll notice a 10-minute break here planned after the fourth center presentation, after the Michigan State one, and we’ll plan on doing that, and then we’ll come back and hear the last three centers. So, to get things started, let’s get rolling with the University of Pennsylvania, and I believe it’s going to be Gregory K. Brown presenting. So, introduce yourself, Greg, and your center, take it away.
GREGORY K. BROWN: Thank you, Jane, and I just want to thank the NIMH leadership and organizers for putting this together. It was quite a feat. So, I will share my slides. Okay, let’s see. There we go. So, can everybody see my slides? So, hi, everyone. I’m Gregory K. Brown. I am the PI.
JANE PEARSON: Can you put them in presentation mode, Greg? Sorry.
GREGORY K. BROWN: Oh, they’re not? Okay. Not yet. It is not letting me do that. Interesting.
BAILEY/WEBINAR SUPPORT: And Greg, do you perhaps have two monitors?
GREGORY K. BROWN: Yeah, I do.
BAILEY/WEBINAR SUPPORT: Are you able to, when you go into your screen share, just make sure that you select whichever monitor, so the opposite one of the one that you have up now?
GREGORY K. BROWN: I’ll try again. Sorry. That’s all I can do.
BAILEY/WEBINAR SUPPORT: Okay. Would you like for me to share for you? I do have a copy of your slides.
GREGORY K. BROWN: No, because I upgraded them, so.
BAILEY/WEBINAR SUPPORT: No worries.
GREGORY K. BROWN: Okay. So, I will just share what I have here then. Sorry.
So, I’m Gregory K. Brown. I’m the PI and co-director of the Penn Innovation and Suicide Prevention Implementation Research Center, along with Dr. Maria Oquendo, who is also the PI of the center and co-director. Our mission is to … there are two major goals: 1) to develop, test, and scale up suicide prevention practices for racial, ethnic minorities, sexual minorities, and socioeconomic disadvantaged groups, and 2) develop and test implementation strategies to bring evidence-based suicide prevention practices to scale efficiently and with high fidelity.
The INSPIRE Center focuses on intervention and assessment practices that can be deployed rapidly in a range of practice settings and in collaboration with community partners. The center consists of an Administrative Core that includes myself, Dr. Oquendo, and Dr. David Mandel. We are advised by a data and safety monitoring board, as well as an external advisory committee that includes several people who are on this webinar.The Administrative Core directs a steering committee that includes all of our study investigators, as well as a stakeholders committee that is external to the project that includes community stakeholders and patient advocacy groups and people who work with higher risk individuals. The steering committee Administrative Core supports a signature project and three exploratory projects that I’ll talk about in a minute as well as a suicide prevention scholars program and a pilot projects program. We currently have four ongoing pilot projects. The Methods Core that is directed by Dr. David Mandel and Dr. Danielle Mowery supports all the research that is being conducted.
To get into a little more detail with each of the projects, for exploratory project one, it’s entitled Supporting Transitions to Adulthood and Reducing Suicide, or STARS. It’s directed by Dr. Lily Brown and Dr. Jose Bauermeister. It includes postdoctoral fellows, Marin Kautz House and Jenny Tran, as well as a support team and youth advisory group. The aim of this project is to conduct a systematic suicide prevention adaptation of iReach mobile app. The name of the intervention is STARS. It incorporates safety planning content and targets coping, social support, and positive affect using adapted framework for emergency adult and sexual gender minority individuals. The intervention includes six peer mentoring sessions to support app use. The second aim is to examine preliminary efficacy on suicidal ideation behavior and mechanisms of action of the STARS intervention relative to the control condition, which includes safety planning alone, using a prospective randomized control trial design. And the third aim is using re-aim metrics to examine whether STARS has preliminary evidence for impacting intervention implementation outcomes, along with emerging adult and sexual gender minority individuals compared to the control arm. This protocol is described in an article that was recently published in JMIR Research Protocol in 2023. You can access that there.
Our next project is an exposure-based implementation strategy to decrease clinician anxiety about delivering suicide prevention evidence-based practices, otherwise known as Project Calmer. It is directed by Dr. Emily Becker-Haimes and Dr. Maria Oquendo, and Dr. Shari Jager-Hyman is also a co-investigator, and it also includes Jesslyn Jamison, a postdoctoral fellow. The specific aims of this project is to develop an exposure-based implementation strategy, which we call EBIS, to reduce clinician anxiety and bolster self-efficacy to deliver suicide screening assessment intervention by leveraging exposure-based therapies to manage core clinician fears and anxiety related to suicide in high-risk patient encounters. The second aim is to iteratively refine the EBIS through pilot field testing and using rapid cycling prototyping with up to 20 clinicians. And the third aim is to test acceptability, feasibility, and preliminary effectiveness of the EBIS compared with implementation as usual to target suicide prevention practices and its implementation and outcomes in a pilot randomized trial with 40 clinicians. This protocol has been described in an article in Implementation Science Communications that is noted here.
Our third exploratory project is entitled Navigating Mental Health Treatment for Black Youth with Suicide Risk, otherwise known as Pathways to Care. It is directed by Dr. Rhonda Boyd and Tami Benton and includes co-investigators Jason Jones and Jeremy Esposito, conducted primarily at CHOP. The aims of this intervention include conducting a systematic adaptation of suicidal teens accessing treatment in the Emergency Department, or STAT-ED, for suicidal Black youth presenting in the ED. And this intervention addresses cultural stigma, problem identification, negative perceptions of treatment, and logistic barriers to engaging in follow-up mental health care. The second aim is to examine implementation outcomes of the patient navigation intervention for Black youth and caregivers compared to controls. And the third aim is to examine a preliminary efficacy of the intervention compared to usual care on primary outcomes, including mental health treatment initiation and number of visits, and secondary outcomes, including suicidal ideation.
Our signature project is entitled A Safety Planning Intervention and Follow-up Telehealth Service Model for Suicidal Individuals in Emergency Department Settings. It’s directed by myself and Dr. Courtney Benjamin Wolk and includes co-investigators Kelly Green, Shari Jager-Hyman, Maria Oquendo, and Megan Reilly, as well as many Methods Corps co-investigators. The specific aims of this project is to develop an external suicide prevention consultation center to provide telehealth delivery of safety planning for at-risk patients prior to ED discharge, plus follow-up telephone contact after ED discharge. There are three main aims. The first aim is to compare the effectiveness of the suicide prevention consultation center and enhanced usual care using a step wedge cluster randomized design for reducing suicide attempts and deaths and for increasing engagement in mental health outpatient care and reducing return suicide-related ED visits and hospitalizations over a six-month period. This design involves eight EDs, all within the PennHealth system. Our second aim is to examine preliminary implementation outcomes, including the number of patients receiving safety plans and follow-up calls, the fidelity to the safety planning intervention, and whether the consultation center is acceptable to and feasible for ED clinicians, health system leaders, and patients. The third aim, importantly, is to conduct a cost and cost offset analysis of the suicide prevention consultation center and usual care treatment. This intervention was recently published in the protocol in Contemporary Clinical Trials, noted here. And then our Methods Core, which is directed by Dr. David Mandel and Danielle Mowry, includes co-investigators, myself, Dr. Lily Brown, Dr. Freya Nezir, Dr. Bob Gallop, Dr. Amber Calloway, and Fran Barg, as well as a very talented research support team, individuals listed here, and our postdoctoral fellows, Hayoung Donnelly and Eric Crosby.
The Methods Core includes four aims. The first aim is to be a scientific incubator, to refine and develop new methods to answer three major challenges in the field of suicide prevention research. There are three projects within this aim. This includes optimizing intervention adaptation and implementation strategies or participatory design using pre-mortems, rapid prototyping and testing, as well as time and motion studies. The second project uses natural language processing, a type of machine learning, to develop cost-effective and accurate methods to assess fidelity for the safety planning intervention. And this is important because all of our data in the signature project is coming from the electronic medical record. And finally, to advance the science and measurement of implementation costs for suicide prevention interventions and adapt and refine a pragmatic approach to costing using time-directed activity-based costing methods and develop an online cost evaluation tool. The second aim is to provide a centralized state-of-the-art assessment, data management, statistical, and human subject protections to support the INSPIRE projects and subsequent research applications. The core will maximize the potential for data pooling and sharing through common use of measures and, importantly, data harmonization so we function as a center. The aim three includes dissemination of INSPIRE scientific resources. We have a website that is listed here. We do provide education in suicide prevention informatics through regular webinars and, of course, providing mentorship to our postdoctoral fellows and our junior faculty. And then aim four is to ensure rigorous evaluation of the research productivity for the center and impact by providing strategic methodological and technical support to the Administrative Core.
One of the things I want to highlight in our center, just as an example of the technical sophistication that this project supports, includes what we call a simple EMR viewer or electronic medical record viewer. So, in our signature project, our data for our effectiveness aims all come from the electronic medical record. And we need to be able to extract that data from the medical record, but at the same time de-identify it.So, we start with the, you know, the doctor writes a note in the patient’s chart. That data is then scrubbed using a program called Filter to take out all of the identified information that is then uploaded into another database, which houses the Filter ED viewer. And that center diagram is an illustration of the viewer that includes key information that we need to evaluate the suicide prevention consultation center. And you can see it includes the after-visit summary, the safety plan, the Columbia suicide disparity rating scale, and risk assessment. That viewer is run by a research coordinator who culls the data, uploads the data, the de-identified data, into REDCap for analysis. And how am I doing on time? Doing fine?
JANE PEARSON: You’re doing great.
GREGORY K. BROWN: Oh, thank you.
JANE PEARSON: Yeah, a minute or two ahead. So, we’ll just check and see if there’s any other folks from the centers who want to have, who have any questions.
JANE PEARSON: I actually have one question that’s very specific. In the suicide area, we’re used to talking about post-mortems, but what’s a pre-mortem? You mentioned. That’s a great question.
GREGORY K. BROWN: Is David Mandel on the line?
JANE PEARSON: There he is.
DAVID MANDELL: Yeah. Hi. A pre-mortem, which is sometimes referred to as brainwriting, a phrase I never quite understood, is a process of trying to anticipate all the things that can go wrong and developing proactively strategies to address. It involves listing in a very systematic and facilitated way, everybody’s worst fears. You start with the premise, it’s catastrophic thinking. This project has failed. Why has it failed? And you have people generate on their own quickly all the reasons it possibly could fail. You organize them along continual probability and how catastrophic it would truly be. And then for those things that are in the quadrant, that is high probability, high catastrophe, you then invest a lot of energy in thinking about what potential solutions and fail-safes are that you could put in place to prevent that from happening.
JANE PEARSON: Great explanation. I think that’s wonderful. And it sounds like good advice for anybody doing grant writing, too. Like, what if our plan doesn’t work in terms of how are we going to explain it? And thus, building in ways to check to see if it’s working the way you think it should be along the way too. So that’s my simple interpretation. But that sounds really interesting. Looking forward to that. Let’s see. We’ve got one quick question. “Is the filter ED viewer available to people outside of your center yet?”
GREGORY K. BROWN: Not yet. I don’t know, Danielle, if you’d like to comment on that?
DANIELLE MOWERY: Yeah, absolutely. I want to clarify. I think there’s kind of two main pieces here. I think there’s Filter, which was a de-identification tool for text that was originally developed at UCSF by Dr. Butes’ group. We’ve adapted this for our own electronic medical record data and enhanced it to be able to integrate both structured and unstructured data to do the de-identification. So that’s one piece. The second piece is the viewer itself. I think, as Greg alluded to, we’re going to get a lot of rich data coming from electronic health record. How do we maintain the gestalt of reading a case that’s been de-identified and helping the abstractors find the information that they need, reducing information foraging and to be able to abstract out those outcomes quickly so they can do the analysis piece? And the short of it is, our electronic health record is Epic, and we don’t have a second copy where we could just load this in.
That really meant that we needed to start from scratch and say, what’s the minimum viable product for being able to render this data back to the analyst and to work with the analyst to make sure that we’re going to provide it back to them in a way that’s easy to interpret? And we’ve gotten a lot of great feedback from some of our research coordinators and faculty on the grant to help us to demonstrate that. That framework itself will be open sourced and will be readily available, and we’ll put it on a GitHub at the completion and work with other sites that might want to use that framework to facilitate abstraction of sensitive and maybe even non-sensitive populations for their own studies. Absolutely.
JANE PEARSON: That’s terrific. That’s great. It’s a good illustration of how these centers are national resources and can share things they’ve learned along the way. Terrific. Let’s see. Do we have any other questions right now? I think that answered it for Anna. All right. So, the way we’ll roll here is if there’s not time for questions and I can’t help transition to the next person, if we’re going tight, the next person could just say who’s the next center presenting? So, thank you both. The next center we’re going to hear from is from the University of Pittsburgh with David Brent presenting.
DAVID BRENT: Hello, everyone. Let’s see. I got to share here. How is that working?
JANE PEARSON: We can see it, but it’s not in the slide. There you go. Perfect. Okay.
DAVID BRENT: All right. Anyway, thanks, Greg. I really enjoyed your presentation. I would say I would describe this presentation as a pre-mortem. You’ll see how closely that corresponds to your definition. Anyway, our A2 Center is based in pediatric primary care.& And I want to acknowledge Joel and Jane and Stephen O’Connor’s work. I remember the dark days when there wasn’t much in the way of intervention and services research, and they really, I think, kept this dream alive. And I don’t think we’d be here today if it wasn’t for them. So, I’m quite grateful.
I would say the genesis of our center, at least for me, was in seeing the progress that we had made in the adolescent suicide rate evaporate. I started working in suicide in about 1980. And over time, you know, we acquired more information, and we thought, you know, for about 15 years, the suicide rate was declining. And then in about 2007, it began to go up. And this graph stops at 2015, but it’s continued. And so clearly, whatever it is that we had been doing was no longer working. And so, you know, this generated for me at least three questions, which is: Where should we be looking? How should we be assessing them? And who are the kids that are at highest risk? And one of the things that collectively we considered was to pivot to primary care, because that’s where the kids are. They’re being screened often. It’s the last clinical contact before suicide death.
At this point, there’s, in some systems, there’s universal screening with a PHQ-9. And there’s evidence from adult studies that if you improve depression management, you can actually reduce the regional suicide rate. And I think, in light of the fact that there’s been the rate of suicide and suicidal behavior has grown most quickly in youth of color, that there’s evidence that treatment delivered in primary care actually can bridge health disparities. And this is a study that I participated in with Robin Weersing, where we delivered a treatment that she developed for depression and anxiety. And what you can see is that the effect sizes are actually much greater for the Hispanic patients than the non-Hispanic patients. And there are other studies like this. So that, I think, is one of the promises that this is a place where we can address some of the health equity issues that we’re now confronting. And this center had a prequel.
We had an ALACRITY Center beginning in 2018.And so, we tried, and that was just at University of Pittsburgh. Now, in this iteration, we’re collaborating with Children’s Hospital of Philadelphia, and we’ll get to that. We talked to pediatricians, and they identified three problems. One is, now that we’re screening for depression and somebody screens positive for suicidality, what do we do? The second one is that kids and families are referred for treatment, and they don’t go. And then, you know, they’re told that they have to develop a safety plan, but they don’t know how to do it.
We developed three digital tools to try to address that. And Oliver Lindhiem and Anna Radovic worked on the first project. We developed something called Screening Wizard, which assesses comorbidity and suicide risk, treatment preferences, and barriers, and gave some talking points to the clinician to try and come up with a treatment that was a better fit for the family. And in a randomized clinical trial, the rate of referrals increased substantially. In the second intervention, it was developed by Tina Goldstein and Brian Suffoletto, and it was a text, like a nudging intervention using motivational interviewing principles.
And in an open trial, we found that we were able to increase adherence to treatment. And then, in the third issue, Stephanie Stepp and Candace Biernesser worked on this, we developed something called Guide2BRITE, which was an app. We had developed an app called BRITE, which is a safety planning app. And the Guide2BRITE actually leads the clinician step by step through the development of a safety plan. And in the clinical trial that we conducted in the ALACRITY Center, those that were assigned to Guide2BRITE did have lower depression and ideation scores, although a lot of kids actually didn’t use the app, so there was definitely more work to do. And so, we then made a transition to suicide, you know, the practice-based suicide prevention centers. And so, some of the team who were added, so Bruce Rollman is the Methods Corps director, Jami Young is the site PI and co-director at CHOP. And I’m not going to go through everybody’s names here just in the interest of time, but basically, one of the things that we did was to develop an intervention where we put together those three digital tools in something that we called iChart. And so, our signature R01 is a randomized clinical trial of iChart for youth that screen positive for suicidality on the PHQ-9, with the goal to see if we can improve treatment engagement and also decrease the rate of suicide events.
Because of the challenge of trying to do a better job of identifying who’s at risk, we worked on a machine learning algorithm of the healthcare record that we then wanted to test on a cohort that we’re recruiting through the signature R01. But one of the issues with adolescent suicide is it’s quite impulsive. And so, we also wanted to be able to identify when somebody was at risk. And so, a subgroup of 500 kids we’re following, getting data from mobile sensing and social media to see if we can identify temporal inflections and ultimately be able to deliver interventions in time. And another overarching issue in the center is that we wanted to address some of the health disparities in youth suicide. The way that we are doing that is our qualitative work is focused on issues of health equity. And we’re trying to recruit a sample that’s at least 35% Black to ensure that any of the products that we deliver will be acceptable and effective for Black youth and families.
I think I have a little bit of time to introduce some of the other people. So, right next to the iChart is Kalab Abebe, who’s our project statistician, co-director of the Methods Corps.In the first line, next to improvement of suicidal youth is Nadine Mellum and Stephanie Stepp, Karen Schwartz, and Molly Davis, who lead the signature R01. The next line is Rich Tsui and Neil Ryan and Ron Barzilai, who are involved in the development of the machine learning algorithm. Mun Mun Dechahury, Randy Auerbach, and Nick Allen are involved in the social media and mobile sensing, respectively. And Molly shows up again because she’s the best trained in implementation science in our group, and so, was helping to lead the qualitative work. Rhonda Boyd, who also works on the Penn Center, and Cesar Escobar-Vieira and Patricia Dock, have helped in a work group on health equity that’s, you know, collectively, our goal is to make sure that whatever we develop will reduce rather than increase disparities.
And in terms of the R34s, we chose three R34s that addressed risk factors for suicidal behavior that are also worse in Black kids. And so, we thought not only is it something that would be useful in the treatment of suicidal kids, but it also might help to address some of the drivers of these healthcare disparities. And so, the first R34 that is led by Tina Goldstein and Ariel Williamson, who’s now moved to University of Oregon, Peter Franzen and Adrian Cerner, is an adaptation of a transdiagnostic treatment for sleep problems, and we’re adapting that to use in primary care for depressed kids who have sleep issues. The second intervention focuses on low physical activity and anhedonia, and is led by Jamie Zelazny, Charles Jonasson, and Rhonda Boyd. And it uses an app, Vira, which is developed by Nick Allen, to use mobile sensing to assess and deliver behavior activation. And then, the third intervention focuses on online victimization, which unfortunately is much more common in ethnic, racial, and gender minorities. And this is led by Cesar Escobar Viera, Candace Berneser, and Jason Jones. And the goal is to develop a chatbot that engages youth who are being cyber-victimized to basically help them to either disengage or use social media in a way that decreases victimization.
We have one overarching goal that it’s not a research project per se, but we want to train pediatricians in mental health skills. And this is a program at CHOP that’s led by Wendy Wallace. And one of our goals for the Methods Corps is to evaluate the impact of this program. And finally, we want to train a diverse cohort of early career scientists, and we’ve appointed three étude scholars to that end. This just gives you a big picture of what the projects look like.
The signature R01 is the gateway to all the other studies. And in it, we’re recruiting 1,200 adolescents. And then of those, 900 go into iChart. Those are ones that screen positive on suicide. And then about 300 go into the R34s. The social media and mobile sensing studies will take about 500 out of that 1,200. They may be in, you know, more than one study. And the predictive algorithms are based on all of the medical records, but then we’re going to see how that algorithm works in the people that we have more detailed follow-up.
And some things are going pretty well. We’ve engaged a number of practices. We’re close to our milestones at this point. We have a pretty high rate of baseline, accepted baselines after consent, and we have a very good retention for follow-up once they get into the study. One of the things that’s happened is as a result of the center, we’ve seeded and supported seven NIMH grants, two small business grants, a K Award, and three foundation grants, So I think that part is going really well.
You’ll hear about this maybe tomorrow, but we’ve developed one cross-site collaboration with Northwestern. And as I said, we’ve appointed three scholars and funded three pilots grants. But we have some challenges. One has to do with recruitment and primary care. And that is that we’re enrolling only about 18% of those that are potentially eligible. And right now, due to our capacity, it takes us about a month to get somebody into the iChart intervention. The second thing is when we looked at the engagement in iChart, only about a third of those who were randomized to iChart were using at least one of the skills, so, low use. And there was some evidence from another study that we did that those who are at higher risk were more likely to use the app. And then the third issue is that, and this is tentative, but at our first look at the data, the sample that we’re recruiting we thought was at high risk, but we only have one suicide attempt in the first 64 kids who we followed up. We’re trying to figure out what we should do about that. One issue was we’re working on increasing our capacity so that we can get people in. But I think that one of the issues that will address all of these is to enrich our sample to get a higher risk subgroup who may be more motivated for the intervention in the first place.
And, in terms of the engagement, we’d previously had a longer engagement process where we oriented people to the iChart app, and we spent a lot of time going over the elements that we had used before. This was led by Karen Schwartz and Molly Davis. We developed an onboarding manual that we believe is going to make a difference. One of the things that I was doing to try and prepare for this meeting was I went to this paper, which is really excellent. I think Brian Ahmedani may be the senior author, but I think the paper is led by Emily Braciszewski. And it looked at the presentation of youth who died by suicide age 10 to 24, where they showed up, what was their last clinical contact compared to a matched control group. And the metric on the y-axis is the odds ratio. They looked at anybody and then whoever had a mental health and whoever had a substance diagnosis. And what you can see is that if you just get anybody in primary care, you have about a 1.8 odds ratio. Whereas if you take somebody who’s been hospitalized, and these are probably mostly psychiatric hospitalizations, you have an odds ratio of 90.
I began to think like we had made this case, at least to ourselves, that what we were doing in primary care is the best way to find these kids. And unless we incorporate some of these other indicators, I think we’re going to end up with a group that’s too low risk to make a meaningful difference. Some of the things we’re thinking about is to try and, in addition to having suicidal ideation, to have a screen on substance abuse, or a visit to an ED or an inpatient unit for behavioral health, to maybe take kids who are depressed without a suicidal indicator, but some of these other issues. We’re considering that now, because I think that otherwise, I don’t think we’re going to make the kind of difference that we thought we were going to make. To highlight this, there’s a study that just came out in JAMA Network Open, that I thought was fascinating, that looked at suicide profiles over the last 20 years. And one of the things you can see here is that there’s one type of suicide that’s decreasing, and a couple that are increasing. And the one that’s decreasing is people that have pure mental health issues. The ones that are increasing are either polysubstance abuse or mental health plus substance abuse. And I think that that may explain why the suicide rate declined in the 90s, because our mental health treatments had improved.
The issue is that we’re facing a different problem now, that is probably related to more comorbid substance abuse. And I also believe that the rise in the diseases of despair in parents is driving the suicide rate in kids now as well. And this comes from some work I did with Robert Gibbons and Kwan Hurr, it’s ongoing, but it’s from a large claims database, 160 million people, and on the left hand side, you can see there’s really like a tripling of the rate of people who have parents who have diseases of despair: suicide attempt, substance abuse, alcohol-related disease. And if you look at the kids, not surprisingly, they have about double the odds ratio of a suicide attempt. If the diseases of despair have doubled over these last 10 years, it could explain as much as 50% of the increase in teen suicide rate. I just think that these, we need to align what we’re doing more related to those issues.
Our conclusion about, you know, what we’re doing and what we should be doing is that I think pediatric primary care, it’s a good place to find lots of kids who are being screened for depression and suicidal risk. And I think it’s a good place for preventive interventions, including promotion of gun safety, which I think is really important. But I think without enriching our sample with higher risk kids, it’s unlikely that we’re going to, that what we’re doing is going to have an impact on the suicide rate. And that if my thesis is correct, that the children of parents with diseases of despair and youth with comorbid mental health and substance abuse are a significant source of the increase in the teen suicide rate, then our screening and treatments need to be aligned with this if we’re going to reverse this increase in adolescent suicide. And that concludes my remarks.
JANE PEARSON: Thanks, David. Yes, thinking through those settings is always interesting because over the population, you might have a higher volume in primary care, but then you can think of lower volume, higher risk groups. So, it is a trade-off, I think. We have one question. “How are you finding the primary care docs? Are they receptive in picking up this role?” And let’s see, from those in residency training to seasoned doctors, “Are you seeing any difference between types of providers and their length of training, I guess, or recency to training?”
DAVID BRENT: So, I think that I’ll speak to our experience in Pittsburgh, and I think Jami Young is on the call and she can speak to, if she wants to, to speak about CHOP. What I would say is that in theory, they’re very interested. In reality, they’re just too busy to do much of this. And so basically, our research staff are doing the recruitment. They can make the referral to us, but then we do the consenting, and we are doing the interventions. And perhaps that’s realistic. I just don’t think they can do one more thing. Although, at an earlier point in time, I had a fantasy that they would be.
JANE PEARSON: Jami?
JAMI YOUNG: I would just echo that. I mean, I think we have a number of training programs here that have started at the residency level, and now we have a group of training programs that are focused on embedded providers. So, I think we’re increasing the number of people who are interested in mental health and can deal with some of the mental health challenges, but we’re hearing very strongly, and it’s one of the reasons why we did it this way, that they will need help by a social worker or somebody else embedded in primary care in order to do this work, especially as we’re asking them to screen for more and more conditions in their well visits.
JANE PEARSON: Thank you. I know from the American Academy of Pediatrics, the partnership that they had with AFSP and Lisa Horowitz from NIMH helped with their blueprint on, I think that’s what it is, on suicide prevention. Like step number one, develop a partner in the community who’s a behavioral health specialist. And I think you’ve been through some of those experiences as well. So, it’s like trying to balance just handing the kids off versus how much can they do in the staff. Do you see some variation there, or do you hear differences like between Pittsburgh or CHOP in Philadelphia?
DAVID BRENT: I don’t know how much it’s, I mean, Jami can speak to this. I think there’s more practitioner variation than site variation. You know, we had one person who had the, for health reasons isn’t working with us now, but who was very, very actively involved in dealing with mental health. And I think Wendy Wallace, who I think we’ll hear from a little later at CHOP, you know, was super involved. And then some of the people she’s trained, but I think that, you know, there’s a lot of variation within practitioners.
JAMI YOUNG: And just to add to that, I think that’s absolutely true. But I also think we have differences in terms of how many embedded integrated behavioral health clinicians we have. And even within sites, we have some sites at CHOP that have that and some sites that don’t. Some sites that have an embedded, what we call medical social worker who’s there for linkages of care. And so, I think part of our work from an implementation standpoint, we’ll be thinking about, you know, who are the best ultimate implementers of these interventions if they’re effective, and how do we try to ensure that different people with different backgrounds and training can do it. But I think that it will be the rare pediatrician who both has the passion and the time within their setting who can do the onboarding, let’s say, for iChart, as David described.
JANE PEARSON: And then did you experiment at all with telehealth in terms of consultation? Has that come up?
DAVID BRENT: Well, everything we’re doing is remote.
JANE PEARSON: Okay.
DAVID BRENT:And I mean, I think that the pediatricians are really happy about that because they don’t want us taking up any space.
JANE PEARSON:Okay. I wasn’t sure exactly if that was the complete…
DAVID BRENT: Yeah, I think we had been in office, and then with the pandemic, we switched, and we said, hey, look, it works just as well, and we’re out of their hair. So that’s pretty much what we were doing.
JANE PEARSON: Okay. Thanks for clarifying that. Okay, great. We have one question was, there were two conclusions you had, David. One was about the diseases of despair for the parents translating to kids’ higher risk, and then he says, what was the other conclusion?
DAVID BRENT: I would just say that I think that unless we enrich our sample for higher risk kids, that what we’re doing, it may do good for kids, but I don’t think it’s going to make a dent in the adolescent suicide rate.
JANE PEARSON: Or it might do good in the longer run versus an immediate high-risk acute, and that would take a long time to sort out probably.
DAVID BRENT: Well, you know, our center started, you know, it’s now five years, so I guess we have 12 years to disseminate before we’re average.
JANE PEARSON: Okay. Good way of looking at it. All right. Thank you both, and thanks for the questions folks are sending in. And we will transition now from David to Children’s, Nationwide Children’s Hospital, and it looks like it’s Jeff Bridge presenting.
JEFF BRIDGE: Hello, everybody.
JANE PEARSON: Okay. Take it away, Jeff.
JEFF BRIDGE: Okay. Thank you very much. Great to follow Dr. Brown and Dr. Brent. For those in the audience who don’t know this, Dr. Brent is my mentor. So, it seems like I’m always following David Brent, which is an honor. What I’m going to do is I’m going to pick up the discussion and I’m hoping that my slides, can everybody see this? All right. Let’s try this.
JANE PEARSON: Now it’s starting. Okay.
JEFF BRIDGE: Everyone see that now? Yes. Okay. Terrific. All right. So, I’m representing our team. RP50 is titled ASPIRES, the Center for Accelerating Suicide Prevention in Real-World Settings. I am the PI and co-program director along with my longtime colleague, Dr. Cynthia Fontanella, who is also on the call. This is just, this represents our building. This is to personalize this a little bit. So I am, we’re located here on the second floor of the Big Lots Behavioral Health Pavilion at Nationwide Children’s. And at Nationwide Children’s, it brings together the departments of pediatrics at The Ohio State University and the Division of Child and Adolescent Psychiatry at OSU in the Department of Psychiatry. And what I wanted to do first is just, so RP50 is housed within our Center for Suicide Prevention Research. And just to give you a little bit of a flavor for what our center does, as the name suggests, there is a research and prevention imperative that we follow. As part of that, we have about 40 plus prevention and research team members who do a variety of work.
And one of the things in prevention, our signature project for suicide prevention has been since we were established in 2015, is doing school-based suicide prevention work. And we have implemented the Signs of Suicide program in over 200 schools throughout central southeastern Ohio. So more than 100,000 students have benefited by receiving that prevention program. And through that work and other work that I’m not going to talk about, we’ve really developed a mechanism for implementation. And we, kind of, we had done it organically, more through a quality improvement lens. But I’m hoping that at the end of this, our discussion is that the audience will get a sense for how that prevention work really influenced our P50 application and the work that we proposed and are now doing. So, our ASPIRES Center, our overarching goal is to accelerate the delivery of youth suicide prevention interventions in clinical and community healthcare settings to reduce youth suicide. And to accomplish our goal, how we’ve envisioned this is that we want to apply a public health approach to rapidly develop, test, and refine these innovative interventions that really, when the research dollars go away, these interventions will have stickiness. They will stick around and continue to work within the practices.
And so here is just a model of what this looks like. So, in the light blue, this represents our current state. We’ll call this current pre-P50 state. Our expertise in our center, it really focuses on epidemiology, health services research, all related toward youth suicide prevention. We looked at the P50 as an opportunity to expand.We’re partnering right now with colleagues on this panel on interventions. But we’ve always been sort of secondary. We’ve been sites.And we saw this as an opportunity for us to really expand what we’re doing and be primary, primary throughout Ohio. With that, the green represents more of this widespread adoption of evidence-based and evidence-informed practices. As others have described, we’ve put together a transdisciplinary team to solve these issues. And I’m not going into our training program so much on this presentation, in part because I’m happy to follow David for a variety of reasons. One is that I’m going to pick up on some of the challenges of doing this work in pediatric primary care. And so, we don’t have too many slides about our mentoring and training. What I will say is that we’ve launched our pilot program. Our pilot program has really focused on training emerging scholars as well as advanced scholars who might be in other related fields or disciplines that have an interest in doing youth suicide prevention work.
Our center was launched in August of 2022.We had our first round of pilot grant submissions. We awarded three pilot grants. Our due date for the second cycle is April 15. I think we have four applications so far and looking to continue with that. So just briefly, and I wish I had more time and I wish I had pictures of everybody. I thought that was a very nice touch. We have such a wonderful group. Our Methods Core is led by three multi-PIs, Guy Brock, Ann McAlearney, and Chris Bartlett. And as you can see, really there’s a breadth of expertise here. In biostatistics, health services, and implementation science research, and informatics. And our working groups really were developed based on our areas of our challenge areas and areas where we have gaps and we really need to work to inform the field. So not surprising implementation science and qualitative analyses, statistical modeling and data, digital health and data management, AI and predictive analytics, and population and social determinants of health.
Let me orient you to this slide. So, the way we were thinking of ASPIRE, how would our center, how would the projects be linked together in an informative way? And we really picked up from the National Action Alliance for Suicide Prevention and the focus on a continuum of care. And how do we prevent patients and families from falling through the cracks? Our four projects, our signature project and three exploratory projects, our signature project, which I’ll describe in detail in a minute, comes out of primary care. And you’ll hear more about that in a second. Our exploratory research project is involved in specialty care. So, this is really focused on improving family, familial risk factors that are linked with youth suicidal behavior. When we move through specialty care to more acute care, our second exploratory project, I’m going to describe the intensive crisis intervention that was developed here at Nationwide Children’s, and then from there, transitional care. And that’s a project to reduce access to lethal means during high-risk transitions.
This slide really, again, when we were thinking about ASPIRES and how could attaining a P50 from NIMH focused on rapidly reducing the youth suicide rate, how could that really just launch the work that we’ve been doing within our center? And the way we thought of this is that we can leverage the resources that we have within our existing center and the best practices that we’ve developed to create this innovative and expanded base. So, we’d go from individual projects to having integrated programs, from epidemiological research to more translational research. Again, not losing our core base, but adding to it. Really having this focus on continuous quality improvement and sustainability and doing both primary and secondary prevention work.
So, I am going to go through our three exploratory projects first, and I’m going to do this fairly quickly because our talk today is really focused mainly on the challenges that we have encountered, especially with our signature project. Our first project is led, it’s titled Acceptability, Feasibility, and Appropriateness of a System of Suicide Prevention for Youth with a Parental History of Suicide Attempt. This project builds off of the expertise of Doctors Arielel Sheftall and Ted Beauchaine.Dr. Sheftall left Nationwide Children’s, she’s now at the University of Rochester, and Dr. Beauchaine left Ohio State for Notre Dame, but we still obviously continue to partner on these projects. And how this work developed: We originally intended to do more of an intervention, but in talking with program and thinking through where the critical gaps were, we realized that this study really should be more pre-implementation. What Dr. Sheftall and Beauchaine and the team are doing is they’ve set out to characterize the children, the offspring of parents with a history of suicide attempt, with the goal here of being to identify interventions, existing interventions, and gaps to engage proximal targets to reduce current and future suicide risk in these youth. But also examine the acceptability, feasibility, and appropriateness of the PACE intervention mapping that would be developed by doing both quantitative and qualitative work.
The quantitative work, as its starting point, is that at the hospital, we document patient visits where the parent has indicated that they’ve had a history of suicidal behavior. When we look over our records over the last several years, we’ve identified thousands of young people with a parental history of suicide attempt. This creates an opportunity for our Methods Core to do some work in terms of data mining, using predictive analytics to build a study design where we can do a follow-back, but also a cohort study where we can follow these youth forward from the time of the index presentation where it was documented that the child has a parent with a history of suicidal behavior. And we can mine our databases to examine risk factors within a large clinical database. We also can get feedback qualitatively from key stakeholders, so administrators, clinicians, parents. And all of this is done with the goal of creating this clinical intervention mapping guide to help inform best practice. So that is our initial project. It was started in August of the first year of the grant. Excuse me, for taking a sip of water.
The second project is called intensive crisis intervention. And the study PIs here are Doctors Mary Fristad, David Axelson, and Jenny Hughes. This is in the acute care setting. Our hospital system over a decade ago developed ICI, which focuses on reducing suicidal behavior and high acuity at risk youth. And why we developed this is that our local ADAM board put out a request for proposals for psychiatric inpatient alternative treatments that could assist with the growing demand for acute mental health care. And the RFP specifically indicated that they wanted a short-term intensive stay treatment that could stabilize acute crises, offer medication management, and linkage to services upon discharge.
The team set off to create an intervention that could fit the description. And sort of the core tenets of ICI is it’s intensive, it’s short-term, there’s a cognitive behavioral approach, and a family-centered approach to the treatment. The model emphasizes that learned maladaptive responses to stressors contributing to crises can be changed quickly with intensive therapy for both the child and the family. And we’ve done some pilot work where we’ve shown in an open case, a case series study where we showed positive associations on reducing suicide and improving functioning. And also in a comparison to youth who were receiving inpatient unit, inpatient treatment, better outcomes, and reduced length of stay. So, this was very promising. And so that is going to be tested in a randomized controlled trial. There’s a lot of work to be done there, but that’s the ultimate goal there.
The third exploratory project is called Safer Still. Dr. Jack Stevens and myself are the study PIs. This focuses on transitional care. We know that the one-month period following discharge from an inpatient psychiatric setting following suicide risk is a very high-risk time for engaging in suicidal behavior. Dr. Stevens and I have been working on this proposal for many years now, and we’re delighted that it was well-received as part of this P50 application. Dr. Stevens’ expertise is in behavioral economics. This web-based application will test a series of nudges during that one-month post-hospital discharge period that promotes safe firearm storage. There’s also safe medication storage and other lethal means restriction, but really the emphasis is on safe firearm storage because our sample will be parents of teens who are at high risk because during the inpatient stay, they indicated that they have firearms in the home and that they don’t store them safely. And so there will be a randomized controlled trial to test the effectiveness, and we have implementation science aims as well.
Our signature project, so again, picking up from where Dr. Brent left off, our signature project is in pediatric primary care practices. And our study is going to utilize a hybrid one-stepped wedge effectiveness implementation design. If you are not familiar with this design, I will walk us through it in a second to determine the effectiveness of implementing a quality improvement intervention that delivers a clinical pathway within pediatric primary care for youth who are detected being at risk for suicide. I’m going to go through this quickly. Our patients, our study eligibility, we’re focused on youth age 12 to 17 who are presenting at one of 14 pediatric primary care centers/practices across our state. They are not affiliated with Nationwide Children’s main campus. We’re recruiting without regard to current or past mental health problems or comorbid physical illness. And we do require a parent or guardian of an eligible teen. The exclusion criteria, medically or cognitively unable to participate, age less than 12 or greater than 18, without access to permanent residence or phone, and non-English speaking, or in foster care.
I hope everybody can see this. I’ll just walk us through this. The yellow star represents where we are here in Columbus, Ohio. One of our practices, community practice, is in Hilliard, Ohio, which is a suburb. We’ve also partnered with Dayton Children’s. So, we have two Dayton Children’s pediatric primary care practices and another practice just south of their cornerstone. We’re up in Toledo. We have three practices within Toledo, three practices up in the Cleveland area, actually northeast of Cleveland, more in the Ashtabula area. One practice in Akron. I should say that these four practices are all federally qualified health centers. And then we have three practices down here in southeastern Ohio. So, this is close to the West Virginia border. This is in the Appalachian region of Ohio.
Our clinical pathway, very quickly. All of the practices that are participating screen currently for depression using some version of the PHQ-9. With screening for depression comes screening for suicide risk because the ninth item of the PHQ-9 is an item that is included to detect suicide risk. We are, as part of our pathway, layering on to that screening protocol, adding the ASQ (the ASQ). And the ASQ was developed at the National Institute of Mental Health by Dr. Lisa Horowitz and colleagues, Nationwide Children’s, myself, the team, we’re partners in developing the ask. Why are we including the ask on top of the PHQ-9? Work that we’ve done and others have done has shown that using a suicide-specific risk screener on top of depression screening can identify more youth who may be at risk for engaging in suicidal behavior. Second part of our clinical pathway is to train providers in how to do a brief suicide safety assessment. We’ve based this off of the model used at NIMH that was developed there, and we’ve adapted it in consultation with experts at NIMH and Chiati, Brombat, and others to make it relevant for the pediatric primary care setting.
The next step of this is to, at the conclusion of the BSSA, that will determine if a full mental health evaluation—but actually the first step would be whether or not an Emergency Department referral would be indicated or a mental health evaluation or outpatient referral or no further action required. The intervention includes brief safety planning within the practice. There’s a care and contact piece to this, so we’re able to implement text messages via a care and contact approach, and our team at Nationwide Children’s has developed a way to do this, and we’ve been implementing it in a variety of settings across our region. There’s care navigation. Nationwide Children’s runs an affordable care organization called Partners for Kids. As part of that, there’s care navigation for patients that meet certain eligibility criteria. Our patients who would be in that more intermediate risk for suicide meet that criteria, and so they’re eligible for the care navigation. Why I’m taking the time to describe this is that these are all elements that then can be sustainable, so we’re really concerned of adding elements that, again, when the grant dollars go away, that won’t last, so that, in our state would last, and then as part of this, we’re doing simulation training as well. How am I doing on time?
JANE PEARSON: You’re running out of it. Okay, so a couple more, yep.
JEFF BRIDGE: Okay, so our design, the yellow, represents our treatment as usual phase. Treatment as usual is followed by the intervention. I’ll use our first site, Toledo, as an example. All of these sites were randomized, so this is our random allocation here. All sites get at least six months of treatment as usual. Our first cluster goes live in month seven. That follows the intervention phase followed by sustainability. That is our design. Currently, we anticipated about an 11% screen positive rate using the PHQ-9 in our practices. What we have found is that, let me orient you to this slide, so annually, we had almost 13,000 patients predicted, which would lead us, if we do the math, annually, there would be 1,400 who screen positive, a monthly rate of about 61. Why it’s 61 is we would identify 121 youth, but we anticipated about 50% would participate.
This next slide is showing all of the actual data of the first four months of our study. We had a total visit numbers of over 4,000. We identified 221 based on screening, 57 based on our inclusion criteria. We recruited 11 participants. That’s a huge loss. Why did we do that? Why was it as bad as it was?I’m not going to focus on potential solutions. I’ll talk about challenges. There was a low rate of engagement with our potential eligible participants in some sites. Some sites just failed to give out the iPad despite weekly reminders. There’s a low callback rate. There was an increased no-show rate at several large primary care centers. That no-show rate is just patients not showing up to scheduled appointments. We encountered the holiday seasons, which impacted our ability to recruit. We found out that the PHQ-9, even though some practices were saying that it was being administered to all eligible patients, it was not being administered. We found out that there’s a large Spanish-speaking population at two of our practices that we were unaware of at the beginning. Then there’s stigma, lack of interest, and privacy concerns.
Our current recruitment strategy has been that we use iPads that are programmed in REDCap that have just a very simple asking parents if they are interested in participating or hearing more about a research project that their providers are partnering in to reduce the risk for suicide. The iPads are … cut it? Okay. All right.I think in the interest of time, I’m going to say that we’ve struggled to recruit. Now, we have alternative strategies that we’re working on to alleviate that concern. I’ll leave it at that.
JANE PEARSON: Okay. Thanks, Jeff. There’s so much to cover. I apologize. I think I also did a little late start there for you, so I didn’t know much. I think we can add that to the pre-morbid possibilities of things going wrong. I like that approach.
JEFF BRIDGE: I’m sorry.
JANE PEARSON: Thank you so much. No problem. Jennifer, you’re on for Michigan State.
JENNIFER JOHNSON: Fantastic. I will screen share. Okay. I am going to present about the National Center for Health and Justice Integration for Suicide Prevention, which is co-led by me at Michigan State, Brian Ahmedani at Henry Ford, and Lauren Weinstock at Brown University. As you all know, suicide is a top 10 cause of death for people under age 65 in the U.S., and we have effective interventions to prevent suicide. But the biggest problem in suicide prevention is this. It’s how do we get these effective suicide prevention interventions to people who need them, and especially those who aren’t well connected to healthcare. We do know some suicide risk factors. There’s been a lot of research on this, and this lets us know some about who is at risk, but they don’t let us know as much when they’re at risk.
So, let me tell you a story of someone we worked with. This is a man who had been trying to rebuild his life. One day, something went wrong, and he lost his job and his hope for the future. He started to get more anxious, and then he started drinking. And as he became more intoxicated, he started having suicidal thoughts. They got worse throughout the evening, and he decided to end his life by crashing his car on the highway, making it appear an accident. But as he drove, he got pulled over by law enforcement. The police didn’t realize they were encountering someone in an active suicide attempt, and instead of being taken to the hospital, he was arrested and detained in jail. The criminal legal or criminal justice system in the U.S. is vast and wide and complicated, and on any given day, most people who are criminal legal involved, even those who’ve previously been arrested, are actually in the community, not in custody.
This table shows various parts of the criminal legal system. Importantly, jail is different from prison. People in jail are often in jail only for a few days. I think the median length of stay is seven days or so. They may not be convicted, and they may not even be charged. But part of what our research has found is that the fact of the criminal legal contact, whether it’s being pulled over by the police or spending a night in jail, marks suicide risk in the community.Because if you’re ever going to be arrested, it’s usually following a moment when things in your life are spinning out of control, like the story I just told you. And those interacting with criminal legal systems are especially likely to not be well connected to healthcare. And, you know, to our previous point that was made there, especially likely about two-thirds to three-quarters of substance use problems and about half have mental health disorders. And so there’s a lot of that mental health substance use comorbidity.
So just think in your mind, guess for a minute, what percent of adults in the U.S. who die by suicide do you think have spent at least one night in jail in the last year? Take a guess for yourself.Our research has shown that 20% or one in five of adults who die by suicide in the community in the U.S. each year have spent at least one night in jail in the last year. Therefore, you know, this research points in one way to one of the big problems in suicide prevention, which is how do you find people at risk when they’re at risk? And admittedly, you know, this isn’t in-the-moment risk, but it’s, you know, in the past month risk, in the past few months risk.And our research, which actually also involved Gregory K. Brown, who’s on this call, also identified the first evidence-based suicide prevention intervention for the 10 million releases from jail each year. You know, there’s recognition that suicide in jail is a problem, but actually odds/ratios for suicide after jail are much, much higher. And like I said, you know, one in five adults who die by suicide in the U.S. each year have spent a night in jail in the past year. So, that helps to answer one of the problems, you know, a big problem for suicide prevention, but unfortunately, as there always is, there’s a second big problem.
And this is a problem that’s plagued work with individuals interacting with jails for decades. The more than 3,000 jails in the U.S. are local. They take place at county and city levels. And the city and county, they’re, like I said, run by these local municipalities, and they’re independent of each other. And, you know, these are different, again, from prisons that are for sentenced people that are typically run by states or the national government. But these county and local jails are just never going to have the capacity to link people in the community back to the community services that they need. We’ve been trying to do that in various ways for 40 or 50 years around mental health, substance use, HIV, a thousand different things. It just isn’t feasible. There’s 10 million admissions or discharges from county and local jails every year.
People are only in for a few days. Releases are typically unscheduled. You might be in for two days. You might be in for eight days. You know, you might be in for 28 days.We don’t know, and you may not know until that day. And like I said, a majority of folks have mental health problems, 75% have addiction problems.And the jails just aren’t staffed to try to keep people alive in the jail and connect people back to community care. And then on the community side, the healthcare providers may not know that that person was in jail. They typically don’t. We’ve had cases where someone was in jail for a little bit, and the community provider, they missed their appointment, so they were dropped from care. So, you may think that it would be possible to link all these systems electronically, but it’s not. All the data are in different formats. And the jails tend not to have strong IT capacity anyway, and you’re typically talking about these small local institutions. It just isn’t feasible, and especially not across all of these different kinds of institutions at the scale we would need to do it.
As part of this work that we do, Fay Taxman and I, who’s another investigator in the center, had a conversation with CareSource Managed Care Organization in Ohio. And you may or may not know, it turns out arrest and release data out of jails are public, and health data, as we know, is private. And it turns out that technology has advanced to the point that it’s possible to scan all this publicly available jail booking and release data for matches to patients in a healthcare system. When I started doing this 20 years ago, the jails still had paper, electronic medical records, right? I mean, there’s all kinds of problems with this, but this is a very, very clever solution. CareSource had realized that they could, that this other data was publicly available, and that they could match this arrest and release data from the 108 Ohio jails against their own subscriber list each week. So, making the connection this way, like I said, is amazing. It solves a 40- or 50-year problem on the criminal justice side and care connection side, because it doesn’t require the release data to be in any particular format, doesn’t require direct coordination with the jails, which is complicated. It protects the privacy of the health data, and it means we have a scalable way to let the local health systems know that one of their patients was just released, so they can reach out and check with the person who may be in crisis.
NCHATS was the culmination of putting all these ideas in this past work together, this idea that contact with the criminal legal system marks a high-risk time for suicide in the community, that we have discovered effective suicide prevention approaches for these groups, and that we now have a way to let the health systems know to reach out. So, it’s solving, you know, a couple big problems simultaneously. As I said, this center is led by me, Brian Ahmedani, and Lauren Weinstock, and it’s a set of, like, as you know, eight related research projects that demonstrate and test suicide prevention approaches based on these ideas. I’m really excited. I think all of us are really excited that it’s pragmatic and scalable, we think, and maybe you guys can tell us that we’re wrong, but we believe that it involves the two largest randomized suicide prevention trials ever, or that we know of. One is 60,000 people, and one is 43,000 people, and the center is not only conducting clinical research saving lives, but we provide training and research for junior faculty, postdoc students, and training in suicide prevention best practices for health and the criminal legal systems.
We have projects across the country. We tried to have the project span many aspects of that model I put up of the criminal legal system earlier. Like I said, it goes everywhere from the 250 911 calls per year to the 53 million police contacts to the 10 million jail admissions to probation to all these different parts. Our four main projects have projects all across the country, and they happen after police contact, jails, courts. I’ll just, you know, go over these briefly. I’m happy to say more about them, but this, our signature project is our randomized trial of 60,000 people, and we’re identifying, takes place at the clinical sites, our Henry Ford Health and Health Partners in Minnesota, and it’s identifying, you know, using this idea to identify folks in these health systems who are getting released from jail and then doing some outreach from the health system side, and with the whole care pathway depending on, you know, level of risk and engagement and whatnot. Another of our pilot projects randomizes 43,000 people. This one we’re doing with CareSource in Ohio with their managed care subscribers who pass through jails, and they either get caring contact letters or care as usual.
Then there’s a whole other part that we do with trying to see if folks who have been seen in one of the large behavioral health providers, just to reach out: “Hey, how are you doing? Are you all right? Do you need anything? We’re here for you.” And re-engage in care if they want to. Project three looks at validating the MHRN screening tool for use in jail and in that period post-jail release, and then project four is led by Ben Cook and with his partner Jamie at the Cambridge Police Department looking at police and partnering with the Emergency Department to try to get people at risk for suicide into care. Because one of the challenges we have, the reason so many people go through jails in the U.S. is that the police don’t have a better place to take someone when the Emergency Department doesn’t want them if they’re too drunk, a little bit belligerent, they fall through the cracks. The jails have lots of people with developmental disabilities, with dementia, simply because the police didn’t have a better option, and that’s part of what Project 4 looks at. It’s a huge effort. Our center is truly national. It involves more than 15 university and health systems that we’re working with nationally, and we’re especially excited about our national group of partners. These are organizations of people with lived experience of suicide, lived experience of criminal legal contact, prosecutors, sheriffs, the American Foundation for Suicide Prevention, a fantastic group of partners that this center brings together to work across the silos.
What do we do with those partners? We really work to make NCHATS community engaged and based on community-based participatory research principles. One of the goals of the center—when system fragmentation is the problem, bringing key players together is a crucial step.Knowing each other and talking to each other that occurs as part of the center is actually the way we conceptualize that as one of the outputs of the center, is bringing all these constituencies together to talk about these things. We seek the partners’ input at every stage of the center and project activities, from conceiving to disseminating. Partnering organizations were part of the conception of many of the projects, like, I talked about Project 4. The MPI is from the police department. We have quarterly all-partner meetings, think tanks to bring all the partners together to think about each project, to launch it. We integrated their ideas into the calls for the RO3-type proposals, all projects, incorporate partners, et cetera.
Like I said, the goal is to create this group of folks that work across the health, suicide, and criminal legal systems to talk about that. One of the things that’s been interesting is that we started with a group of criminal legal partners who had all bought in to the fact that we’re addressing opioids as part of their jobs for judges and whatnot. I think that I just thought that if you had bought into behavioral health around opioids, you would just buy that suicide prevention is part of your job. There’s a lot of interesting stigma and misconceptions around suicide and what causes it, suicide risk and suicide prevention that we’re uncovering and addressing in the center, including in one of our four pilot projects.
I also want to say that our dissemination plan really relies on this idea of these folks finding opinion leaders among all of these partnering groups who then become stepping stones to dissemination and implementation among their constituencies. For example, if we have American Jail Association representation, we can leap into there. Really thinking about the social network that the center itself is building across all these constituencies and then a way of getting word out about the findings and getting things implemented.
We have cost-effectiveness analysis for all center projects. We have a set of implementation science activities across all the projects that will result in assessment of some of the implementation outcomes. I would say they’re all hybrid Type 1, but then the Methods Core has a set of activities that will result in a scale up focused implementation approaches for each intervention that can then be tested in future trials. I’ll just say that in conclusion, we have evidence-based suicide prevention interventions in the community. What this center does that we’re excited about is demonstrate and test some of the new ways to identify people most at risk, especially those who aren’t well connected to care, who are impulsive, whose life circumstances are chaotic, and try to connect them to those effective interventions to prevent suicide.
JANE PEARSON: Thank you, Jennifer. The scale of this is amazing. Really anxious to hear more about how it will be going. We’re about five minutes behind. Thank you all for being flexible enough and working through this. I’m going to turn it back to maybe Deb, who can tell us how we’re going to recalibrate for taking a 10-minute break, I think. Then when we get back, we’ll be hearing from the University of Massachusetts.
DEBORAH KRAT: Yes.
JANE PEARSON: There we’ve got it. Okay.
DEBORAH KRAT: Thank you very much, Jane, and thank you everybody so far who has presented. All these presentations have been great. With that, we’ll now move into our break time. Again, it’s a 10-minute break, so it’s now 1:45 PM. We will return back at 1:55 PM Eastern time.
JANE PEARSON: Thank you. Okay.
DEBORAH KRAT: Welcome back from the 10-minute break. I hope you all had an enjoying, relaxing break. We will now turn it over to our next set of presenters from the University of Massachusetts CAPE Center, Ed Boudreaux and Catarina Kiefe. It’s turned over to you.
JANE PEARSON: Thanks, Deb. We can see your slides.
CATARINA KIEFE: Thank you very much, Jane, and thank everybody at NIMH for this really interesting meeting. I have a tough act to follow. The previous presentations have been really amazing, but I’ll do my best. By way of introduction, many people in this space probably know one another, and I don’t. That’s because I am not a suicidologist. I am a methodologist. And I am multi-PI with Ed Boudreaux on our center. Ed Boudreaux pulled me into suicide research about 10 years ago when we got an R01 together, and the rest is history. But I still have not acquired the knowledge about suicide research that he has. Everything that’s content-related in the center is really something he created, and what’s methods-related is something in which I have input. That having been said, the CAPES, the P50, is part of a larger center at UMass, also called CAPES, with a mission of which is to prevent suicide by accelerating delivery of evidence-based compassionate suicide care. And within that, the P50 really specializes in suicide care technologies to support the primary mission of this center. We try to accelerate the delivery and provide evidence-based and compassionate care. Those are our main drivers. What motivates us in our true north, as Ed likes to call it, is that the overarching CAPES center needs to prioritize the efforts that are directed towards suicide prevention, and at least one of the four domains depicted here on this PowerPoint, namely implementation science, digital health innovation, predictive analytics, or patient-centered care. And of course, the ideal projects will incorporate more than one of these four domains. The P50 effort must advance suicide prevention, patient-centered care, implementation science, and digital health predictive analytics as a center. Not every one of our projects does that.
In the next slide, we talk a bit about how we’re going to do this. The target populations, our target populations in the P50 are a broad range of healthcare settings, allowing for evaluating common implementation themes and setting specific themes. We really pay basic attention, fundamental attention to zero suicide key performance elements, specifically identifying risks, engaging in care pathways, treating suicide risks, and ensuring transitions across care settings. The technologies we study are already mature and can either directly perform the performance elements of zero suicide or enable clinicians to do so. Finally, we’re using the common implementation science framework, the EPIS framework, to guide every project, including qualitative interviews in the garnering of data for every project. Every project is focused on development in real-world settings. We use common measures and data elements, and we plan to synergize across all projects.
This is an overview of our organizational structure. It’s a little bit busy, and I apologize for that. But what I want to really highlight is that we are, we try to, and I think we’ve accomplished being transdisciplinary, meaning multiple disciplines work together, not just to answer questions, but to generate new questions which wouldn’t be asked if it weren’t for the transdisciplinary nature of the center. So, like all the other centers, we have two directors, Ed Boudreaux and myself. We have an executive committee, which is a little bit larger, and basically, and we have the same basic fundamental structure of an Administrative Core, a Methods Core, and then our projects.
In the Administrative Core, we have the necessary operations unit, but we also have a Business Development Unit directed by Dr. Dunlap, an Education and Scientific Development Unit directed by Dr. Nunes, which is very active, but I, in the interest of time, I’m not going to belabor what’s being done there. And we have an Ethical, Legal, and Social Implications Unit directed by Dr. Nebeker from the University of California and to others. This is, as you can see, there are many different disciplines in the Administrative Core, and then within the Methods Core, it is directed by myself and Emmanuel Agu, who I believe is in the audience today, from the Worcester Polytechnic Institute. The whole center is a collaboration between UMass and WPI, the Worcester Polytechnic Institute, and Changwu Yang, a biostatistician and psychometrician. So, within the Methods Core, there are five units and another one that’s shared with the Administrative Core. The design data and analysis unit is sort of the workhorse of data and design and analysis, and it’s directed by Dr. Yang, a psychometrician, and Bruce Barton, who is a highly experienced clinical trialist.
We have an Economic Evaluation Unit, which is going to, which designed and is beginning to perform economic evaluations for each of our four main projects, signature and the other, and the three exploratory projects. And it has two faculty in it, it’s directed by two faculty, there’s also staff, who are health services researchers and health economists. Then we have a Machine Learning Unit directed by Emmanuel Agu, a Person-centered Design Unit, which is directed by Ben Gerber, who’s a general internist with a strong interest in informatics, and Ben Guizzo Tulu, who is also from WPI and is a researcher in patient-centered design. And finally, we have an Implementation Science Unit directed by Stephanie Lemon and Kate Pivovarova. Kate plays that role, but also she leads the evaluation unit, which is common to both the Administrative Core and the Methods Core, and basically does, directs all evaluations. I’m going to talk about the three, the projects next, so I will leave this slide now.
Our first project, which I’m just going to give a high-level overview now, is called JASPR. JASPR is an existing patient-facing, tablet-based application for patients with suicide risk. It administers evidence-based suicide risk assessment and interventions, and it has access to tools that the patient can take JASPR home on her smartphone to access JASPR after the ED visit. Going on, we’ll hear more about JASPR in just a couple of minutes. Then we have the project we fondly call CATS, for computerized adaptive testing, which essentially seeks to be transformative in identifying and tracking suicide risk in primary care. It’s directed by Rachel David Martin, psychologist who’s in the audience here, and Ben Guizzo Tulu from WPL. This is an animated slide that they produced, and it’s kind of fun because it plays on the CATS acronym, of course. And the Aim 1 of this CATS project, it has three, is to identify effective approaches to implementing the CAT-MH in primary care. The Aim 1, informant interviews and clinical administrators interview, queried, and patient, for example, patient takes the CAT-MH results, are presented to the clinician, and the next steps are based on CAT-MH. In Aim 2, they will conduct usability and acceptability testing of the integration of CAT-MH into the EHR. And in Aim 3, they will evaluate the feasibility of EHR integrated CAT-MH and the deployment at one primary care clinic. And I will leave you with that on the CATS project.
The next project is on automated data-driven adaptable and transferable learning for suicide risk prediction. The PIs are Fei-Fan Liu, an informatician expert in NLP and AI, and Tony Rothschild, a psychiatrist. And basically, it’s the background of motivation for this is that machine learning for suicide prediction has gained a lot of attention nationwide, and it’s important, of course, no need to convince this audience thereof, but there is a dearth of tools to assess and improve prediction models generalizability, and there are gaps between research results and clinical utility, which stymie adoption. Aim 1 of ADAPT is to assess generalizability and availability of the NIH-funded mHERN risk prediction model. Aim 2 develops the ADAPT pipeline for automatic adaptation. It preprocesses model adaptation, hyperparameter tuning, and interpretation. And Aim 3 explores deep learning for suicide risk prediction. Oh, I thought I had automated the slide, but I didn’t. That’s why I was telling you about the aims.
The third and final exploratory project is on leveraging early mental health and covering risk for suicide. LEMURS, we are a biologically oriented, zoologically oriented group. We have CATS, we have LEMURS, and we’ll come up with other similar names for the pilot projects that we intend to fund. It’s run by Dr. Dickson-Gordon from UMass Amherst, who’s here today, and Ellen Rundensteiner from WPI. The goal of LEMURS is to streamline risk screening and monitoring of college campuses, which, as you know, are plagued by a high rate of onset suicidal thoughts, behaviors. And like several other centers presented here today, we do not focus on a pediatric age group overall. Most projects are age agnostic, but this one focuses on college students. And what they will do is they, in Aim 1, they’ll do stakeholder interviews to understand needs of students and clinicians. In Aim 2, they’ll develop a LEMURS dashboard for monitoring and integration. And Aim 3, they will evaluate the feasibility of LEMURS by correlating with daily self-reported suicidal thoughts.
Finally, getting to the JASPR, which is the signature project. As I indicated, I gave you the highlights before. JASPR is a chatbot-administered version of CAMS, well-known CAMS, which of course is one of the most well-studied approaches for suicide risk management. JASPR performs personalized safety planning and lethal mean safety counseling. Its reports are automatically presented in the EHR for review and action by the clinical personnel. The workflow is being co-designed with clinicians to integrate technical and clinical workflow to reduce burden and friction. It’s a very complex problem as we’re experiencing many interesting challenges as we try to implement it. JASPR facilitates delivery of evidence-based care. It administers CBT and DBT-based coping strategies and distraction while the patient is essentially sitting in the ER in some kind of dead time waiting for intervention. It has a library of diverse shared stories with inspirational stories, which have all been co-designed with people with lived experience. Finally, I’m nearing the end of my slides.
This is the basic design of JASPR, which in itself is a methodological innovation which was created by one of our lead statisticians, Dr. Cheng-Wu Yu. The design is a novel design. It gained the prize from the Society of Clinical Trials in that it consists of parallel efficacy and effectiveness studies. The efficacy part of the study, which we’re implementing right now, is a randomized controlled trial which is supposed to over 30 months enroll 670 people into JASPR intervention versus treatment as usual. The Part B is the effectiveness part of the study which will implement JASPR in Ohio and in Colorado in two systems that have EPIC as well as UMass. But of course, you’ve seen one EPIC, you’ve seen one EPIC. Every EPIC is different. There are multiple challenges which we’re facing, but we’re moving forward with it. This is the end of my talk. I’ll stop here.
JANE PEARSON: Thank you so much. I’m sure the cat fans were really appreciative. I don’t see any questions for you right now. That last design looks really interesting. I hope when we talk about methodology, we can get back to that. Jordan Smoller, it looks like you’re up and ready to roll with another center in Massachusetts. Go ahead.
JORDAN SMOLLER: Thank you, Jane. I am up and ready to roll. I want to just thank again you and NIMH for supporting our center, for supporting all of these centers, and for bringing everybody together for this meeting. It’s very hopeful and inspiring to see all of these people really trained on this problem in such innovative ways. We’re really looking forward to contributing to that as well. I will share my screen. I’ll just introduce you to our center, the Center for Suicide Research and Prevention, which I have the privilege of co-directing with Matt Nock, who will be sharing this presentation with me. Let me see if I can share my slides and see if they… Does that work?
JANE PEARSON: Yes. Thank you.
JORDAN SMOLLER: I’m hearing no objections. Okay, great. Our center is focused on implementing novel suicide prediction and prevention strategies in healthcare systems. We know, and I think it’s been mentioned here, that the majority of people who attempt to die by suicide are seen by healthcare providers in the weeks leading up to the event. We have this key venue for intervening and prevention. In particular, we focus on the two highest risk periods for suicide-related behavior in healthcare settings. That is, after patients leave an Emergency Department setting for psychiatric evaluation and after they are discharged from an inpatient stay or inpatient care.
The two major questions that we face are: How do we identify those at greatest risk? Then: What can we do to help those patients to prevent suicide and suicide-related behavior? We’ve designed our projects along a kind of two-by-two framework, which you see here. We have two projects focused on the ED setting and two focused on the inpatient setting. Then in the other axis, we have two that are more focused on implementing risk prediction strategies, and the other two more focused on preventive interventions. We’ll talk a little bit about each of these as we go. One of the things that the center has really enabled us to do, which has been just phenomenal, has been to tap into and bring together just an incredible depth and breadth of expertise, commitment, clinical work, and stakeholder groups across our community, across the Boston community, Harvard, Mass General Brigham, and other healthcare systems or hospitals that are affiliated with it. Many of these folks have been working either directly on suicide-related research or clinical activity, or they’re in transdisciplinary domains that we think are essential for allowing us to innovate and implement these practice-based strategies.
For example, at Harvard, where Matt is based as the chair of the Department of Psychology there, we have folks involved in the grant not only in psychology but in behavioral economics, computer science. Susan Murphy, who’s one of the architects of the clinical trial design that we are implementing, sociology, statistics. At Harvard Medical School, the Department of Biomedical Informatics is engaged with us. Tianxi Cai, who’s a professor there, is actually a co-lead on our Methods Core, and she has tremendous expertise in predictive analytics and learning healthcare system implementation and methods development. Department of Healthcare Policy, we have Ron Kessler, of course, who’s a world expert in psychiatric epidemiology, but also in studying the heterogeneity of treatment effects and precision kinds of approaches to suicide risk prediction and prevention.The School of Public Health, we have faculty who are engaged in social and behavioral health in methods and methods development.
We’re partnering with Cambridge Health Alliance, which is actually one of the sites that will also be implementing some of the work that we’re doing through the center. We have community partners, and I’ll mention a little bit more about this later, including a wonderful partnership with Samaritans of Boston that I’ll touch on later.And then in our healthcare system at Mass General Brigham, which serves a population of millions of patients, it’s the largest regional healthcare system in our area, an integrated health system, we partner very closely with our MGB behavioral and mental health leadership. The chair of biostatistics at Mass General, Andrea Folks, is also a co-lead of our Methods Core, has tremendous expertise in biostatistics, but also clinical trials and data management for those.
The Mongan Institute is an institute that’s devoted to implementation science in many different areas. And the director, Steve Bartels, is a part of our Administrative Core and has been essential to helping us think about how we implement and work in implementation science frameworks in the best way that we can. And of course, our hospital systems, McLean and Mass General, and Salem Hospital and others are where a lot of the frontline work is being done and will be involved in the work that we’re going to do in our projects. And then also two children’s hospitals represented with investigators who are contributing to the center as well at Franciscan Children’s and Boston Children’s Hospital.
We’ve been able to become sort of a hub of research that I think is really for the first time in our area able to do this. And as part of that, we have tremendous transdisciplinary expertise that we are calling on and integrating in this. And this was supposed to be kind of a cycle, maybe beginning in the top left there with some of the basic methodologic and computational technologies and biomedical informatics that are going to be very important, innovative trial design. Moving through the clinical trials, we’ll talk about community partnerships, the implementation science, and frontline care. It’s been wonderful to see this really gel and come together. And I’m going to turn it over to Matt, actually, to walk us through just a high-level overview of our projects.
MATTHEW NOCK: Great. Thanks, Jordan. And I want to thank, join Jordan and the other center directors in thanking our colleagues at NIMH for making this all possible. It’s really, really exciting to be part of this. Thanks to Jane and Joel and teams for having this come to fruition. We’re really grateful to be involved. As all other sites have done, I’ll talk a little bit about our four projects. As Jordan mentioned, they’re organized along two focused in the Emergency Department, two inpatient, two prediction, two intervention.
The first, the signature project is focused on implementing and testing a clinician decision support tool to prevent suicidal behavior. And the logic here is over the past decade, our team and a lot of other teams have been building prediction tools that pull from electronic health records and produce a risk score, largely, often you can think of it as a predictive probability of suicide attempt in an individual patient. And this has been done all over the world. And we know that we can make these predictions with a relatively arguably actionable degree of accuracy, but folks are just now starting to try and implement these risk tools and convert them into clinician decision support tools. And it turns out one of the things we’ve been learning is it’s much more difficult to do this in real time at the point of care than it is to do it on a static pre-existing database. But we are persisting and producing such a risk calculator or clinician decision support tool.
If you look at the figure on the bottom here, the first aim of our signature project is to test in a randomized control trial, a clinician decision support tool that collects data from the patient and about the patient, converts it into a risk score, predictive probability of suicide attempt, and gives it to their clinician while they’re in the Emergency Department for a psychiatric visit. So, patient comes into the ED and we’re doing this with 4,000 patients presenting to Mass General Hospital and MGB Salem Hospital, which is a sister hospital several miles away. Patient comes in, we, in real time, pull about 40 variables from their electronic health record, have them complete a brief 15-item iPad-based survey, combine that information to generate a risk score, high, medium, low risk for suicide. Then for a random half of patients, we give that risk score to their clinician. For the other random half, the risk score gets generated, but not given to the clinician. And we’re going to test the effect that that has on clinician decision-making and see whether clinicians are more likely to hospitalize high-risk patients, less likely to hospitalize lower risk patients. And we’ll see if that has a downstream impact on the rate of suicide attempts at one month and six months follow-up. We did an earlier study funded by the American Foundation for Suicide Prevention with 2,000 patients using a similar design minus the RCT to develop this risk model. So, this is really an implementation of work that has been going on for several years.
If we then go to the top of the slide, Aim 2 is to improve our machine learning prediction algorithm using natural language processing on clinician notes, using social determinants of health data, and using newer machine learning approaches that can be used to hopefully improve our predictive model. And the third aim is to develop a precision treatment rule to try and optimize treatment for high-risk patients. And this is work building on some really amazing work that Ron Kessler has done. He has a paper recently, him and his team in trauma psychiatry using VA data showing that overall hospitalization doesn’t seem to have a positive effect on patients at risk for suicide. More interestingly, about 25% of patients get better after hospitalization. Around 25% get worse. And for about 50%, there doesn’t seem to be any effect. And his analysis suggests that if this rule is implemented, we’d see about a 16% reduction in suicide attempts and about a 13% reduction in hospitalization. So, it would be good for the system overall. So as Aim 3, Ron and team are going to develop a similar precision treatment rule for this non-VA setting. We’re not testing it out here. Hopefully if this all works out, that would be the next study we would do. But this is the signature project in a nutshell.
This next project, it builds on this and is an intervention with patients at high risk coming through the Emergency Department. And the motivation, the logic here, is talking to our clinical colleagues, frontline clinicians, other stakeholders, people who have lost loved ones to suicide, or who have been suicidal themselves. We hear over and over again, people are at risk, they call a hotline or they get themselves into emergency care.Sometimes they go to an inpatient unit, they get intensive treatment, then they’re discharged and often get zero treatment or somewhat lost at sea. And we wait until they identify themselves and then come back in again. So, trying to break this cycle and do something different, what we’re doing in this project is identifying the people in the top 50% of risk scores from the signature project, and then randomly assigning them to an enhanced outreach intervention using a hybrid Type 1 trial design, which is focused on testing efficacy of this intervention while at the same time collecting data about implementation of this intervention. And we’re doing this in partnership with the Samaritans crisis line, our local and national crisis line that often have people calling them for help, and they’re referring them to emergency services or otherwise. In this project, the Samaritans team is going to be trained to do outreach, some segment of the Samaritans team to do outreach to people identified at high risk and intervened upon in this trial. So, we’re going to do this with 300 patients, 150 randomly assigned to this enhanced outreach, 150 will get care as usual. We’re going to collect data as we go and some feedback will be provided as I’ll describe to the clinical team. And we’ll do assessments at four and eight and 12 weeks after being seen in the Emergency Department.
This next slide provides a little bit more detail about what it is the clinicians are going to be doing in this trial. After a patient is seen in the Emergency Department and discharged within 24 hours, they’ll be contacted by a Samaritans clinician who will do a risk assessment, safety planning, review their treatment plan, identify potential barriers to care, and then we’ll be in touch with them. We’ll reach out to them at a minimum of weekly contacts for 12 weeks. And the team is going to track those contacts and collect data we’ve developed over the years. I’m doing a lot of intensive longitudinal monitoring via smartphones and wearables. We’ve been following people adverse for suicide during this post-discharge period, and we’ve developed a pretty robust risk protocol for what to do when we get information from someone in real time indicating that they’re at risk.
And a lot of this work was guided by a consensus statement that we developed in collaboration with many of the folks on this call, Jane Pearson, Galia Siegal, David Brent, and others, where we had a consensus meeting to figure out what do you do with someone when they’re at high risk in the moment. And so that information will be used to monitor and follow patients in this trial and provide information back to clinicians, to the Samaritans workers as we go. I won’t go through each of the components here, but you can see what it is that clinicians will be doing and how frequently they’ll be doing them. A lot of this is what other teams are doing. And it’s frankly, as we thought about this, we thought, well, what would we do if someone, a loved one we knew, was at risk for suicide? What are the things that we would do? How would we follow up? And it’s things like reaching out to the person in regular cadence, figuring out what’s happening with them. They’re not fine with their clinician because they don’t like the clinician. Great. Let’s help you find another clinician. They’re experiencing a problem with homelessness, with unstable housing. Great. How can we help them address that? Family issues and so on.So, it’s really just doing intensive case management during this really important transition period.
The next two studies are focused on the inpatient setting. And the first is focused on improving suicide risk prediction in racial, ethnic, and linguistic minority youth. This is focused on the inpatient period for youth ages 10 to 17 and trying to improve upon the EHR-based risk algorithms that I mentioned earlier. Here, though, we’re focusing on two different health systems, the Mass General Brigham System and the Cambridge Health Alliance System, which is much more diverse in its users. And we’re focused on these two different settings and samples of minoritized adolescents. And we’re going to test the improvement that we see by incorporating information about social determinants of health. We’re building this by applying machine learning and natural language processing to a number of years of data that are already in hand in these two different healthcare settings, focused on the post-discharge period. And we’re capitalizing on the fact that Cambridge Health Alliance, for several years now, has been, for all clinical visits for youth, administering the Kiddie computerized adaptive testing battery. We’re getting K-CAT indices of social determinants of health that have been collected and are now available for use. We can test out the extent to which this measure and these assessments improve our predictions.
And the last of our four studies is an intensive intervention in real time, also building on some work that we’ve been doing using smartphones and wearable biosensors to identify folks at risk.Different from CBT or DBT or medication trials where we’re intervening on a regular cadence, this is really focused on what we do in the moment when someone says they’re having intense thoughts of suicide or high suicide intent. And we’re testing this using a micro-randomized trial or MRT applied to the testing of a just-in-time adaptive intervention or JEDI. And so the idea here is that we’re at randomly timed intervals, surveying people on the inpatient unit and afterwards to ask about suicidal thoughts and suicide intent. And when they report high suicidal intent, we respond and we test out various kinds of interventions that are randomized within person. And this is a mockup to hopefully simply show what the design looks like. So, we can think of this as someone completing a rating for suicide intent, let’s say on a zero to 10 scale. And if they respond zero, we’ll call them, they’ll have no risk. One through three, we’ll call them low. Four through seven at medium risk, and eight through 10 at high risk. So, if the person at a certain assessment point indicates they’re having no suicidal intent, we do nothing. If they’re having low suicidal intent, then we’ll randomly assign them to either get nothing or some pop-up message to say, “Hey, it sounds like you’re having high suicidal thoughts or some suicidal thoughts. Here’s what we recommend you do,”—a standard intervention.
Why would we randomize to nothing? It might be the case that when we surface an intervention, even something as simple as a pop-up, people habituate to it and actually is counterproductive. So maybe some value at doing nothing at this lower level of risk. At medium risk, we randomly assign to do nothing, the pop-up or a more interactive chatbot-type intervention. And then at high risk, we’re no longer doing nothing. We’re doing the automated pop-up, interactive pop-up, a text message, or a phone call. And then we have some follow-up assessments on whether they use their safety plan or its components, which they’re encouraged to do as part of these interventions, and whether this led to a proximal change in suicidal thoughts and intent. And we’ll get some passive data on whether they’re actually using their safety plan and how quickly and so on.
And then we’ll look at downstream impact on suicide attempts and hospitalization. And we’re hoping that what we learn here is what are the most effective interventions in the moment for people at different levels of risk, and we’ll also capture some contextual data. It might be the case that the non-human intervention is most effective for people when they’re at high risk, or it might be that it’s most effective when a person’s with someone else because they don’t want to answer a phone call when they’re with someone. And this has come from doing work on intensive monitoring for a long time. We sometimes reach out, we currently reach out to people at high risk, and we often hear them say, “I’m a teacher, I’m in class, I can’t talk to you right now.” You know, the human intervention isn’t always the most preferred, and so this study will allow us to test this empirically.
JORDAN SMOLLER: We just wanted to mention a couple of other things along the way, and one of them is, of course, one of our major goals is to advance the workforce and training of individuals interested in or new to suicide-related research. We did launch our pilot and emerging or scholars program, which, as others have mentioned, includes both emerging scholars, early career, or advanced scholars. And we had a great response, a very, you know, a wonderful set of really competitive, innovative applications, and we were able to, with great difficulty, select two.
Our pilot grant went to Rene Lento, who’s going to be looking at implementation of safety planning in the EHR, and what its effect is on, what effect training in quality improvement has on utilization of safety planning. And Jasmine Brooks Stephens is working with Kate Bentley, who’s one of our investigators, who’s co-leading a couple of our exploratory projects, Kate is, and her work as an emerging scholar this year will be on understanding clinical decision-making for racial and ethnic minorities at risk for suicide. So that program is off to a great start. Matt?
MATHEW NOCK: I will jump in. We were asked to comment on some of the innovative strategies. I’ll be fairly brief because I’m noticing time. We’re really excited about a lot of the ways we’re trying to innovate here, and especially grateful to Joel Sherrill, who, after he had his informational session about this call for proposals a few years back, and we had a brief meeting, as one does with folks from NIMH, and we described what we’re wanting to do, and Joel said, no, you completely misunderstand the assignment. We do not want, you know, research projects as you’ve been doing. We want things that are innovative, and we want things that are very practice-focused and are going to try and really move the needle. We’ve taken that to heart and tried to really innovate wherever we can and have things be very much focused on getting input from clinicians and trying to change the way we do care and test improvements. I’ve described several of these, so I’ll be very brief.
I already talked about our real-time EHR-based risk score, which we’ll be running in real-time within patients’ medical record, and we’re designing it in a way that anybody using EPIC, or we have EPIC in our system, will be able to use this tool and plug it right in and test it out in their setting. We’re developing EHR-based analytic tools in our Methods Core, which we’re really excited about. Our Methods Core is going to develop and disseminate in an interactive web API that will allow researchers from other sites to visualize and search their EHR for features that are associated with suicidal outcomes, calculate suicide risk scores based on patient characteristics, and they’re building an interactive federated transfer learning algorithm that can train an algorithm at other sites based on the distribution of features and models trained across participating sites. So, people at other institutions outside of ours can plug in and build, and this is what this little mock-up on the bottom is intended to look like, plug in, train a model for their site, and benefit from other training that is done locally.As I mentioned with Ron Kessler and others, building precision treatment rules that will also be shared, Jordan mentioned, and we’ll say a little bit more, I think, about our community partnerships, and we’re working with Susan Murphy and Walter Dempsey and others to evaluate JEDIs via this MRT design. I see James, so we’re virtually out of time, I suspect.
JORDAN SMOLLER: All right, so this is our last slide. I just wanted to highlight that, you know, another thing that this has enabled us to do is to build wonderful partnerships, and this one we’re particularly excited about and thankful for, and this, around the time that the call for proposals came out, we were taking a serious look in our system about what can we do that’s evidence-based to reduce risk, and, you know, obviously, as people have talked about, there are a number of evidence-based interventions, including caring outreach and safety planning, when people leave, for example, acute care settings, but the reality is our system, many of the clinicians were not set up to do this. Their workflows, you know, after discharge were not in place, and we connected with Samaritans of Boston, which is essentially our local and has been around for decades suicide lifeline crisis center, now 988, has obviously deep expertise in community-based suicide prevention, a highly trained volunteer workforce that answers, you know, thousands of calls from folks when they are in crisis.
They have extensive training, but we were then able to think together and co-develop and co-design this enhanced outreach intervention that Matt mentioned we’re going to be testing in a randomized trial using a different model, so adapting what their usual model of befriending anonymous support to actually training and providing care, providing support from a coach who can help deliver some of these evidence-based interventions in collaboration with us in the healthcare system, so a tight collaboration there, and it has just really been a wonderful partnership, and we’re very hopeful about this program and if we can show that proof of concept about scaling it up, so I’m going to stop there in the interest of time and just thank everybody again.
JANE PEARSON: There’s a question, yeah, about community resources that maybe you can answer in the Q&A, and we’ll move on to Kate Comtois. Thank you both. Very exciting. Kate, are you ready? Here we go. Here we go.
KATE COMTOIS: We are having our moment of praying for slide success, so hold on. Come on. All right.How did we do?
JANE PEARSON: It’s your view with the notes, so I don’t know if you can flip your screen. That looks better.
KATE COMTOIS: Whoops. Wait a minute. Hold on. I’m going to try it this way. Whoops. Is that better?
JANE PEARSON: That’s good.
KATE COMTOIS: Okay. All right. Thank you so much. I’m going to present about our Suicide Care Research Center that got started about February of 2023. This is a picture of us from our big kickoff this past January with our advisory boards, and our goal and purpose is to improve suicide care for adolescent and young adult patients, so we’re looking at folks from age 13 to 30 kind of working through that developmental period, their families, and their clinicians, and that’s kind of that triangle of the patient, the family, and the clinicians is pretty central to our work. Our goal is to develop a better model for suicide care that can be easily used in full or just in parts by diverse primary care and other outpatient settings.
We’re focused in the outpatient medical settings because we’re seeing that oftentimes outpatient medical settings are relying on the Emergency Department to manage suicide risk, which is often unnecessary, can be overwhelming to the emergency room. Frequently, it’s very overwhelming to suicidal patients and their families, and if what ends up happening is there’s a referral back to the ED, then there probably wasn’t that much point to the referral in the first place. And we’re looking at a lot of programs where there is collaborative care or other integrated behavioral health models, which have spread out considerably, but the suicide care elements of these interventions are kind of lagging behind. We really wanted to embrace the sort of team science, transdisciplinary research team approach, so there are suicidologists among us, a number of folks with deep expertise in clinical and health services research, including models such as optimization, as well as kind of large-scale health services, and we work closely with our biomedical informatics team.
Our goal is to, by sort of integrating together, come up with new frameworks, concepts, and methodologies that will improve suicide care, again, for patients, their families, and their clinicians. We have a very much kind of co-designed model, so kind of core to the idea of this center is that if we develop interventions and then try to implement them, that we’re going to run into many, many problems, and that it is going to be a much better design to do the co-design work at the beginning. We focus all along what we think of as the suicide care pathway that you might be following in an outpatient medical setting, if your risk was detected, some kind of assessment and shared decision-making, there’s some sort of short-term risk management, safety planning, then there followed by long-term surveillance and follow-up, because in primary care, you’re, especially, you’re going to be there a long time, and so the end of treatment is not the end of care.
We’re following a co-design using human-centered design as a core element of what we’re doing, using the DDBT, or Discover, Design, Build, Test model developed by Pat Arian and Sean Munson and Erin Lyon in their ALACRITY Center here at the university. We are working on a component model to integrate and optimize interventions and technologies using the MOST optimization framework, and so we do a lot of the co-design on the ground with the clinicians, as well as with young adult patients and families. That’s our goal.
We also want to kind of step out and think about scalability early, so three times a year, we have our researchers meeting with non-academics, particularly national leaders in suicide prevention from Zero Suicide, the Action Alliance, other, you know, kind of other folks thinking in collaborative care, and also primary care network leaders, so the kind of places where we would want to see our interventions end up. We really want to give an opportunity for them to give feedback on the research as it develops. And people with lived experience, having times, three times a year, when people can present their research to an independent group of people with lived experience to give feedback and input as the project goes along. In addition, some teams also have people with lived experience on their teams. We’re looking at four key mechanisms to achieve our suicide outcomes.
We have clinical mechanisms and implementation mechanisms. The clinical mechanisms are therapeutic alliance and self-efficacy, particularly self-efficacy around suicide risk management, that is, does the patient feel they have the self-efficacy to manage their own risk? Does the family have self-efficacy to manage their family member’s risk? And do clinicians have the self-efficacy to manage risk in the outpatient setting without needing to refer to the emergency room? Our implementation mechanisms, firstly, is usability and acceptability. If it isn’t usable and acceptable, we just figure it’s not going to get used, and so that’s a really core mechanism we believe has, we have to kind of achieve usability and acceptability to get to our suicide outcomes. And we also need it to get to reach, that is, we need these interventions to actually occur in the settings where they’re needed. Again, if that’s not happening, there won’t be as much suicide outcome benefit across the population.
Putting this all together, this is our overall model. We have the suicide care pathway. We have interventions at different elements along the pathway. I’ll show you that more in a minute. We have our two, we have acceptability and therapeutic alliance and self-efficacy, kind of as our initial mechanisms, all driving toward reach, as well as driving directly to suicide outcomes.
This is our organizational structure of our center. Sort of the Administrative Core is divided into the Coordinating Core, Policy Core, and Team Science Core. So, the Coordinating Core obviously coordinates. The Team Science Core is where our R03 and collaborating scholars and all of our training and mentoring is centered here, led by Doyanne Darnell and John Fortney. And we also have the Policy Core, which is kind of coordinating all of our activities with non-academic mentors and our advisory boards. And we’ll be doing, helping us do a national policy briefing of the interventions that we have developed in the center to relevant people who might be able to scale the interventions. We have our Methods Core led by Molly Adrian with Andrea Hartzler, Trevor Cohen, and Kevin Hallgren, developing our products and doing some scientific inquiry. And we have our projects, which I’ll tell you about in a minute.
We have three exploratory and one signature project. I’m going to sort of orient you to them as we move sort of along the suicide care pathway. All the way at the beginning of the pathway, we have, we’re looking at what is it going to take to get systems that are outside of the electronic health record, outside of healthcare, facilely used within the system. In this case, we’re looking at ecological momentary assessment, which can certainly beat out a PHQ-9 in terms of the ability to predict, but how on earth is that data going to be used by a primary care physician, by behavioral health staff in primary care, in any useful way?And how do they need that data presented? How do they need to manage risk, et cetera, to do that? This project’s led by Ian Bennett and his team.And you can see, we are right here up at the front of the suicide care pathway, as well as long-term surveillance, and they’re looking at all of the mechanisms.
Next along the way is the point when we get to clinical decision-making. This is a project using JASPR, which we’ve heard some about already, as a tool for clinical decision-making in outpatient medical settings in pediatrics. And again, we’re clustering here in the assessment decision-making short-term management, as well as some surveillance, with a particular focus on self-efficacy, as well as acceptability and usability, and acceptability, again, from pediatric patients, their families, and their clinicians. This is our signature project. This is Molly Adrian’s project, along with her team. And we are, I’ll tell you more about this project in a minute, but this is kind of as we’ve moved down here to short-term management but getting into treatment. This is where, this is actually based in specialty care, but the idea is the referral comes from primary care to this quick specialty care program that can pick them up right away as an alternative to ED referral, see the patient, stabilize them quickly, and be able to return them to primary care in a situation where the primary care team is able to kind of pick up the ball and keep going, plus any other outpatient referrals.
And lastly, the project that I’m working on with my team is looking at the treatment end. We’re very much focused on treatment and into follow-up, and we’re really focusing on the mechanism of therapeutic alliance. What can we do to improve how collaborative care staff or other integrated behavioral health staff interact with patients based on the Aeschi Principles, which are established principles in suicide prevention, focused on the clinical narrative and developing collaboration. And our hope is that for the patients who are being managed in the primary care environment, that we can help them not only manage their risk, but actually recover from their suicidality and not go on with a chronic risk pattern or have to go to the ER.
I’m just going to give a couple of words about things that we have accomplished so far in this year that we just got started. One has been our kind of actually actualizing our proposed methodology. It was really core to, as I mentioned before, our methodology of integrating human-centered design and multi-phase optimization strategies to kind of develop and prepare for optimization and result in interventions that can be moving right directly into hybrid implementation trials, hopefully in large primary care networks across the Intermountain West. The way that we do this—the center coming out of the larger-most model—is we need to really consider all of the components of an intervention, meaning the content, features of the program, things that are done to improve compliance by the patients, but also things that improve compliance by the clinicians. And what we want to do is develop those components using human-centered design or co-design with our partners. And we, like I said, have adapted the discover, design, build, test model of the Alacrity Center as a way of operationalizing human-centered design in our circumstances. And we’ve spent an enormous time this year really looking at how to kind of get all of us on the same page and be able to really integrate these terms and make it possible for our three applicants to be able to kind of fit themselves within this model, but also to kind of move into optimization.
Our SOARS trial is an optimization trial for our signature project. I did want to say a few words about that. So, the idea here is that when you’re doing optimization, that you have intervention components. And what we’re trying to do before we go to a hybrid implementation trial is figure out how do we make sure this is the best and most efficient model of the intervention we can do. We want to identify weak or ineffective components and get them out. We want to figure out the magnitude of the component’s effects and whether there’s kind of interaction effects, whether one component is improving or weakening another component. And in this case, we’re using a factorial design. There are other designs that you can use with the idea that when we go to that hybrid trial, we should have the optimal intervention to actually try to implement. In the SOARS study, we are using a model of examining and optimizing around three factors. We have a common therapeutic assessment based on CAMS plus care navigation. This happens for all participants. And then we have three elements… SOARS was developed in the field. It was developed within clinics at Seattle Children’s Hospital and was showing really good pilot success. It’s not clear, however, whether we need the entire package. These three core elements of the package, the driver focused skills training, which is skills training directly for the adolescent, skills training for the caregiver, and lethal sort of expanded versus simpler version of lethal means restriction. Do we need all of these components? How core are they? Or could we potentially do this with less?This is what we came into the grant proposal with. And then, of course, we had to actually operationalize this. So, we’re having fun with rainbows.Molly’s team has figured out all of the different factorial design here and has figured out how to pretty amazingly smoothly be able to run all the different options for the MOST trial.
The last thing I just wanted to highlight is another thing we’ve been doing over this initial year. One of our goals, if we’re going to be able to make these interventions work and actually be usable, is we have to be able to get data into the EHR, and we’ve got to be able to get data out of the EHR. And in our case, Epic is also our EHR that the majority of our systems are using. We have been trying to essentially do pilot work with … our interventions aren’t ready yet, but we don’t want it to be that we’re figuring this out once our interventions are ready. We’re trying to figure out a lot of this infrastructure now. We’ve been collaborating with a team in UW primary care. They’re already wanting to work with Epic to improve a SMART set for suicide care. We have joined on with them and the programmers at our ITHS is our kind of our Epic support staff, essentially. And we’ve also had a number of our members join the Brain Trust for Suicide Prevention that Epic National is hosting, and we’re really working on getting that data in. And then we’re simultaneously using a pilot trial that I developed to work on getting all the data out. We want to be able to have all the data that comes in from participants and families be matched with EHR data.
So, this project, which is led by myself, Trevor Cohen, Julie Richards, and Denise Chang, is like we’re working here initially with very, very clumsy elements that were developed sort of on the fly with Epic. And we’re going to use our human-centered design to improve the usability of the system and get it implemented in Epic. And we’re working with our research information technologies office to be able to extract the data and, like I said, match it with my pilot data so that, what our goal is, is an overall data warehouse. We’re working kind of in these first couple of years of the center to kind of get all the kinks out of this. As all of our other interventions are ready, we’ll be able to incorporate them.
And lastly, I just wanted to highlight, we were successful in funding two R03s this year. We’ll fund two—our R03s are two years—we’ll fund two per year for years one to four. We have also brought on two additional collaborating scholars. And I’ll stop there.
JANE PEARSON: Thanks, Kate. It’s amazing. And I’m hearing common themes. Why don’t you just stay on? You want to stop sharing?And I’ll invite all the center directors to turn on their cameras. And I’ll just have a few remarks. All these projects are quite different, of course, but there are some themes and some of those themes were asked on the RFA for the center, so we shouldn’t be completely surprised. You guys got scores and you’re funded. So, you responded. But some of the things were pretty striking to me about just trying to help clinicians do better. So, you’ve got a bunch of different tools rolling, how to be less anxious, decision-making, therapeutic alliance, a lot of those pieces, which says we feel like we’ve got some interventions that we think work, and how can we optimize them. And we know we have a lot of training to do, too, so there were already questions about resources in the community. And then of course, that training is going to vary by the setting. And then Kate was just mentioning even training parents. Lots of novel approaches to reach out. Oh, dear. So, Matt Nock can’t get on. We’ll see if the host can let you on, Matt. Okay.
Novel approaches for outreach, which are great, and novel partners, which, you know, as the centers meet again later today, and then in a couple days, we’ll hear a lot more details. But in terms of the technology that we’re trying to leverage, decision support, and actually getting the users to engage in what we think are very helpful things are all some common challenges. And then the biggest challenge is trying to shift like systems, groups of people all at once, whether it’s two systems, the justice system, the mental health system, if it’s within a care flow, trying to get primary care providers, other clinicians, everybody who’s already busy already dealing with everything they’re supposed to deal with, to do things somewhat differently, track things, follow up, and so on. None of this is new. It’s just letting our broader audience know you are all taking on quite a bit. And it looks really exciting and promising, and it’s really going to move things ahead. I’m just so thrilled that we were able to do centers, that we’re having the meeting today, and we’re going to be talking a lot more about it.
Are there any comments across center directors? Anybody wants to raise your hand, say anything? Maybe you’re all ready for a break at this point. We can do that too. I just didn’t know if anybody else had comments. Some of the folks didn’t have a lot of chances to get questions or answers. Go ahead, Jennifer.
JENNIFER JOHNSON: I’ll just say as an add-on to what you said, the systems we’re working across aren’t systems, right? When you talk about the health system, it’s a thousand pieces, the justice is a thousand pieces. And even people who are working across systems, like the whole hospital, that’s a lot of pieces. If these things were truly systems, this would be easier, but they’re fragmented parts. And I think that’s what we’re all dealing with.
JANE PEARSON: Yeah, that’s a really good point. And I would also say, I think Catarina said, “If you’ve seen one Epic, you’ve seen one Epic.” So even if there’s some commonalities, even with some infrastructure you’re trying to work with, those are our challenges. Matt?
MATTHEW NOCK: Just building on that, I’ll just say, I want to say thank you again, Jane, to you and everyone involved. This feels like one of those situations where it’s, you know, be careful what you wish for, and now we have these centers with a lot of work to do. And I heard from everyone, we’re all encountering lots of complications, but it’s really nice to be doing this at all, and certainly together with six other centers and with support from NIMH. So, thank you.
JANE PEARSON: Yep. We’re really hoping you feel comfortable reaching out to see if somebody’s, you know, solved some problems that you also have, and see if there’s another approach just to make some more synergy across all of this. That’s great. Any other comments? Okay. Well, I want to thank our audience who joined us today for listening in, and we’ll be wrapping up this part. And again, thank you all for your teams. I know you were the spokespersons for your teams, and I know you’ve got very large teams and a lot of work going on. We’ve had a lot of people here at NIMH who, you know, made sure the center whole thing worked in terms of getting everything written, in, reviewed, and so on. And I want to thank Brendan Weintraub, who was doing most of the traffic control here, keeping everybody on track. And of course, to our contractor, MEES, who’s keeping us all in line and rolling ahead. And also to Josh Gordon and Shelli Avenevoli for supporting this and trying to get this moving and keeping suicide prevention a priority.
I’ll dismiss the audience right now, if they want to leave, and then just have your center directors hold on for a few moments, letting you know how this will work. I just want to let you know that we will be using a different Zoom link to start the next session. So, I think, Brendan, it’s labeled closed session first day. Does that sound right? In the email.
BRENDAN WEINTRAUB: Yes, that’s correct.
JANE PEARSON: Okay. So, you’ll have a bit of a break, and we’re going to start again. Should we start at 3:25? I think we can do that if that gives people enough time to switch over. Okay, let’s do 3:25 on a new Zoom link, and then we’ll start our breakout sessions. So, good to see you all. Looking forward to more discussion—and thank you again. Brendan says to reach out to him if you’ve got any questions, but we’ll see you all in a bit. Thank you. Thank you, everybody.