Including People with Suicidal Ideation/Behavior in CNS Trials

In this episode of Mind the Methods (the ISCTM podcast), Dr. Mark Opler speaks with Dr. Elizabeth Ballard, Associate Scientist at the NIMH Intramural Research Program and the 2025 recipient of the ISCTM Lewis Alan Opler Prize, about the evolving science of suicidal ideation and behavior. Dr. Elizabeth Ballard explains why many commonly used tools were built for risk documentation and legal protection, not for detecting treatment-related changes in clinical trials, and how repurposing them can obscure true signals, particularly for rapidly acting interventions. Drawing on trial examples, Dr. Elizabeth Ballard highlights how different measures (full scales vs. single items/subscales) can yield different conclusions about treatment effects, underscoring that measurement choice can be the difference between “it didn’t work” and “we didn’t measure it correctly.” The conversation also addresses the field’s “catch-22”: trials often exclude participants with suicidal ideation/behavior, limiting what we can learn about how treatments affect people with lived experience. Dr. Ballard discusses practical steps to safely include higher-risk participants, IRB and DSMB education, clear monitoring and response pathways, adverse event planning, and shared resources across research networks. Looking forward, Dr. Elizabeth Ballard describes a “both/and” future that blends clinician interviews with self-report, ecological momentary assessment, and emerging objective/implicit measures, arguing that suicide risk is dynamic and variable, and our tools must evolve to capture that reality.

Transcript

SPEAKER_00 0:00

Welcome to the Mind and Methods Podcast, brought to you by the International Society for CNS Clinical Trials and Methodology, your gateway to the latest insights, innovations, and conversations at the forefront of central nervous system clinical research. This series dives into the challenges, progress, and people shaping the future of CNS therapeutics development and trial methodology. In this first season, we'll explore innovations in CNS trial design. From digital endpoints and adaptive protocols to patient-centric models and generative AI. Here, we bring together researchers, clinicians, regulators, and industry leaders to discuss the challenges and breakthroughs shaping the future of CNS therapeutics development. From advancing trial methodology to improving outcomes for patients, our conversations dive deep into the science and strategy behind CNS research. Thanks for joining us. Let's get started.

SPEAKER_01 0:52

Mind the Methods, brought to you by the International Society for CNS Trials and Methodology.

SPEAKER_02 1:00

Hello, everybody. I'm Dr. Mark Opler speaking to you today, and I'm talking with Dr. Elizabeth Ballard, associate scientist at the NIH intramural research program. In addition to her many other qualifications and notable accomplishments, Dr. Ballard is also the 2025 awardee for the ISCTM Lewis Allen Oakler Prize. And it's my great pleasure to be speaking with her today about her work and her perspectives on suicidal ideation behavior, research, assessment, and the future of the field. Elizabeth, thank you so much for taking the time to speak with me today.

SPEAKER_03 1:39

So excited to be here, Mark.

SPEAKER_02 1:42

So I want to start off with a broad question about the work that you do in suicidal ideation behavior, if I may. Why don't we start with the current state? Could you say a few words about what you see as the limitations on current measurement approaches to suicidal behavior research? And just what are the challenges and and barriers you feel that are most important for us to try to overcome?

SPEAKER_03 2:10

Yeah, very excited to talk about this. This is near and dear to my heart. Overall, what we know is suicide is arguably the leading cause of psychiatric-related death. The rates continue to rise even with a dip over the pandemic. And really, it's been real a struggle over the last 20 years for us in the United States to really figure out what's going to impact that rate. And in terms of the assessment, I think there's a lot of things that sort of get combined. Suicide also is potentially illegal liability. And that's something that clinicians often are worried about. You know, is my is my client, is my patient going to go on and kill themselves? And I think from an assessment perspective, is that we've sort of gotten that all twisted up in the sense that we have these assessments that are sometimes used for risk stratification, risk assessment, sort of ascertain medical legal liability, making sure you're documenting correctly. And oftentimes these assessments sort of throw the kitchen sick, you know, sort of all the risk factors at the same time to really sort of show that you've done your due diligence. However, that is quite different than something that you're going to do over the course of a clinical trial to ascertain whether somebody is improving in response to particular intervention. And so I think that's why I think it's particularly, you know, a fraught area in terms of risk assessments or assessments of the course of clinical trials, is because oftentimes we are repurposing assessments that were developed for sort of long-term risk for clinical trials management. And then especially for, you know, so the work that I do with rapid response, and we've talked a lot about this, Mark, before, you know, then you have sort of these new interventions that might have rapid responses. And that's even sort of further sort of a difficulty for people to figure this out. So again, a lot of these measures that we used were not designed for the purpose that we need them for, especially sort of for these clinical trials as related to individuals with suicidal thoughts and behaviors.

SPEAKER_02 4:15

That's a great point. Just to follow up on a comment that you you made that I think merits a a little bit more exploration. You know, the the note about you know trying to repurpose tools developed in for one for one objective or in one era in another. You know, could you say just a word or two maybe about your view on the current state of our measurement tools and the specific issues that you feel are most critical to overcome to drag them kicking and screaming into whatever year this happens to be, I can't keep track.

SPEAKER_03 4:52

Yeah, I mean, I'm gonna focus on suicide, but oftentimes I sort of give a comparison of depression because I'm sort of more used to that. In to a certain extent, you think about it, depression is a diagnosis. There's an episode that you are with a certain number of symptoms that you're trying to create, sort of this episode as part of the assessment. Suicidal thoughts and behaviors are just that. So a thought or a behavior. It's not necessarily a syndrome or a constellation of symptoms. And so oftentimes we have, like I said, these risk assessments that sort of have a lot of facets that are very important for predicting risk. But for the most part, in a trial, what you want to know is are there suicidal thoughts? What are the characteristics of the suicidal thoughts, or has there been a suicidal behavior and how to categorize the suicidal behavior? And so there are commonly used instruments out there. I think we all know sort of what those are in terms of CSS RS, the SSI, sort of other suicide-related measures that are out there. But like I said, they're usually sort of used in many different contexts. They're used in screening, they're used in risk prediction. And oftentimes the sort of tracking of symptoms over a clinical trial is sort of the second or third purpose.

SPEAKER_01 6:09

Yeah.

SPEAKER_03 6:09

And so what that means is that oftentimes, and we've shown this, depending on what measure you use, you will get different results for your clinical trial. And we've done, we did one of my first analyses on postdoc, was just looking at different sort of suicide-related measures across a ketamine trial. And if you use the full scale for suicide ideation for people who don't know, sort of have lots of different facets of suicidal thoughts. You don't bind an effect of ketamine on suicidal thoughts. But if you use single items or if you use sort of a brief subscale of the SSI, you do see this differentiation drug as compared to placebo. And I think this is really critical. And sort of one of my soapboxes for all of this is that we do need better treatments for individuals with suicidal thoughts and behaviors. And what I want to make sure is that we're not sort of misassessing using the wrong instruments for our trials and then sort of thinking, oh, well, that didn't work when really it was an assessment issue, not sort of a drug issue or intervention issue.

SPEAKER_02 7:13

So before we go on, you know, one thing I think we're sort of obliged to talk about here, you know, is the fact that most clinical trials of almost any condition, with rare exceptions, usually exclude individuals with suicidal ideation behavior. So it strikes me in listening to you talk about this area that we we have a real catch-22, a real conundrum. It's very difficult to make any meaningful statements about how suicidal ideations and behaviors are affected by our treatments and trials if we're continually excluding patients who show any evidence of those phenomena. What are your thoughts on this? I mean, what's the best way to proceed? Clearly, you know, we we don't want to do anything that might be, you know, risk. We we can't put individuals' lives at risk. On the other hand, we need to understand this phenomenon in a much deeper way. It's vitally important. How do we strike the balance?

SPEAKER_03 8:18

Yeah, I again, this is one of my soapboxes in that, you know, you wouldn't think of a cancer treatment and then sort of say, oh, no, but we can't have anybody who might die of cancer in our trial. Um, and I think that sometimes we do that in psychiatry, especially related to suicidal behavior. And there's a number of, there's a number of restrictions and there's a number of facets. There's first the level of ethical or IRB review, in that, you know, I've definitely encountered IRB. So they didn't even want you to ask the question of, you know, are there suicidal thoughts and behaviors? So there's sort of that issue. There's also the issue of um adverse events and sort of what happened, it's a of what happens during the trial if somebody does report the suicidal thoughts and behaviors. Nobody wants a suicide death in their trial. I I do understand that. And at the same time, if you're going to do high-risk trials that are really going to impact suicidal behavior, there might have to be some sitting in that discomfort to see that that happens. Um so I think in terms of what needs to happen as going forward is a lot of sort of communication across researchers. You know, once you've figured, once you've really educated your IRB about the necessities, once you've sort of figured out the systems that are required to do appropriate monitoring of your patients in trial, how to report adverse events, once you sort of educated your DSMB, then we should be sharing those resources because oftentimes the IRBs really just want to know like, have you thought about it? Is there research evidence? You know, are we, are you asking us to take on something really risky? And so pointing to the research, saying, no, that this is how this other group does it. This is the research in this area, really can go a long way. And so I think, like, to give the cancer model, you know, I think that there's sort of the model of childhood cancer and all of these sort of research groups working together. And I just think for suicide, we have to have something similar. We have to be able to share those resources because it is so tough to do.

SPEAKER_02 10:26

Yeah. So it's a great point. You touch on an important issue, which I think is, you know, the sort of the institutional stigma around suicidal ideation behavior. I mean, the whole topic is is very fraught in certain circles. And I think that's also true. It pains me to say this a little bit, but I think that's also true broadly, writ in healthcare. We have a lot of folks in a lot of different facets of our healthcare system for whom suicidality is a topic they do not want to touch. Yeah, given that we're starting to see larger and larger studies in clinical trials in areas like obesity. And these are areas that you know typically are managed by folks who don't necessarily have a lot of training in mental health. They don't specialize in it. You know, their focus is other branches of medicine, not behavioral health, not mental health. You know, what are your thoughts, if I may, about you know the need for better suicide assessments in areas outside of CNS, you know, especially obesity, where if if memory serves, there is an increased risk relative to the general population. And it is a higher risk population than is average. How do we help you know investigators working outside of neuroscience who are dealing with high-risk populations get over their stigma and do better jobs, even with our sometimes crude tools? How do we get them to do a better job and not just you know walk through the checklist, but ask the questions that need to be asked in the way that they need to be asked?

SPEAKER_03 12:12

Yeah, I think that's a fantastic question. I think there's the issue of how do you track the ID, you know, potential ideation or risk, you know, probably with an obesity study. You could get away a bit more with a depression measure with one suicide question and sort of support them in that way. But then there's this larger issue of sort of just identification and screening for suicide risk overall, not maybe necessarily as an outcome, but just as a point of good quality care. And I'm going to show my bias, but as my graduate work was developed, an NIMH developed tool called the Ask, the Ask Suicide Questions, which was specifically meant to address the question that you're asking about sort of a brief suicide-related question items that could be implemented in a non-psychiatric setting. And so I think the even if you don't go out and research that, I do think it's a good tool. But I think a lot of the lessons learned from that effort has shown me people, especially in non-mental health settings, need to know exactly what to say and what to do. They need it spelled out, like left to their own devices, who knows what they're going to assess. And then we need to then precisely spell out what is it that they're going to do with that information. Because asking people to do this sort of identification and then just say, well, and then follow up with whatever your mental health resources are, that that's just not going to work. That's not going to fly. And so, at least in this screening sort of pathway, you know, you do you implement safety planning, you implement sort of psychiatric hospitalization if it's indicated, really sort of help them spell out like this probably is only going to happen, you know, 3% of the time that you're going to get this positive response. But if it does, like what are you going to do about it? And so I think that there are really strong researchers out who have sort of developed these algorithms and pathways and sort of bringing that in would go a long way. Again, it's going to be rare, but those those cases I think could make or break a screening program.

SPEAKER_02 14:21

Let me shift gears slightly if I could. You know, I know that there is interest and certainly use out there of tools that are more patient-reported rather than clinician reported. Let me ask you, you know, in your expert opinion, what's the appropriate role for patient-reported outcome measures in understanding suicidal ideation? And what's the best way to integrate these into clinical trial protocols?

SPEAKER_03 14:50

So obviously, I don't think anything's going to sort of overpass the clinician, a good clinician interview, sort of sitting down with people. But again, what makes suicide a bit more distinct from something like a manic episode where you can visually, you know, observe you're sitting with somebody in the room, can very much tell what's going on with them, is that it's it's it's it's internal. People are going to report what they want to report or not. And so I think there's you know, strong need for sort of the clinician to make a safe environment for people to know what's going to happen. You know, if I report these suicidal thoughts, am I going to be kicked out of the trial or am I going to be supported with that? Um, or am I going to be hospitalized immediately? But I do think sort of patient-reported outcomes has its role. And then a lot, I know we're not quite there yet. My work has been looking at more implicit measures of suicide and sort of what else can we ascertain about suicide risk by other sort of metrics, whether sort of it says implicit response to tasks. One thing I'm really interested in is sort of suicide ideation variability and reactivity. And I think those are really new, emerging, interesting areas that we can also go after.

SPEAKER_02 16:08

So to drill down just a little bit, you know, I know there are some studies now in the clinical trial space that are using self-report versions of, say, the CSSRS. Right. Or some of the previously well-established self-report measures for suicidality. You know, from what you're saying, you know, it is such a subjective phenomenon. So maybe there is some wisdom to that approach, um, but it won't supplant the need for you know a good investigator to establish clearly with the patient, even before they use that measure, the freedom to answer honestly.

SPEAKER_03 16:45

Right. And yeah, and even, you know, the new technologies like ecological momentary assessment or smartphone assessment, of course, that's all self-reported as well. They might not even be in your office when that's happening. And so what that's been showing us it is is possible, and you do get more reporting, but you also get more variable reporting. And that's what's really been, you know, sort of as I mentioned about variability is, you know, somebody reports suicidal thoughts in the morning, but then not in the evening. And sort of what how do you interpret that? Um, sort of you would have to have sort of guidelines and sort of a plan of sort of how you ascertain risk, especially if somebody's sort of saying they have it or they don't, because what we know is suicidal thoughts very much vary over time.

SPEAKER_02 17:33

It's interesting you should say that. I mean, it also suggests possibly that our conceptualization of suicidality and suicidal ideation behavior has been overly simplistic. That you know, this dichotomous on-off state may in fact fail to capture a lot of the nuance and you know, the idea that it's it's more of a trait than a state, right? That that also is is probably incorrect given the potential fluctuations and variation over the course of a day.

SPEAKER_00 18:06

Right.

SPEAKER_02 18:07

Let me ask again, just to be very blunt about it. Are our tools up to the task? Or do we need a radical rethinking of how we've approached this in order to get a better view of the phenomenon?

SPEAKER_03 18:26

What I would never want is for people to stop asking about suicide. I guess sort of that's where I'm coming from, especially as I alluded to before, that there are these people who very much take out can't ask because we don't know what we would do with that. And so I would never say not to ask because I think that's better than nothing. But in short, yes, I think we need a rethinking, like you're talking about, sort of what we mean when we mean suicide risk. You know, if somebody's reported suicidal thoughts once 20 years ago, they're clearly in a very different point than somebody had it last night while scrolling social media, but doesn't have it now in your office or something like that. So I think really that we really need to understand more about suicide risk as dynamic, and our assessment tools need to take account of that. And I feel we're in a really good place because of the advances in technology, but I don't think we're there yet. There's still a lot of work to do.

SPEAKER_02 19:25

So speaking of a lot of work to do, if I may, what are what does the future look like from your perspective? If we move in the right direction, if we try to go beyond the limitations of our current conceptual conceptualizations and and you know transcend the limits of our tools, what does that look like? What are the innovations that are going to help unpack this in a meaningful way? What shape are they going to take?

SPEAKER_03 19:56

Yeah. I mean, in my ideal scenario, at some point we have, I don't know if it's a battery or a package in a clinical trial or in clinical research, that you're you're you're collecting this data in lots of different ways. Of course, you're having the clinician providing a warm, safe environment for them to report. But then, you know, there's their self-report. Are there implicit tasks that we could do? So part of my research is focused on the life-death implicit association task, which was developed by Matt Knock up at Harvard, which is just a reaction time task between the self and death and the self and life, and has overall mixed findings, but promising enough. We don't really have any objective markers for suicide risk. Um, and then we've been working on other attentional bias probes for life and death that sort of are still in the works. But like, what if you could come up with an assessment plus a 10-minute battery? And then, you know, the clin, you can sort of the clinician can look at all of that information and sort of overall understand what they need to do with their patient. But then in the course of the trial, we could know, okay, this is what the patient was reporting over time. And then this is sort of what our algorithm predicts their overall suicide risk was. And so I think it's not, it's gonna not be either or, it's going to be and, you know, sort of what are we gonna supplement our instruments with?

SPEAKER_02 21:22

Um right. So the future is is a a bionic augmentation, if you will, right. Of what of what the current state of affairs has been for quite some time. How do we get there? Not a simple question, I I will admit. And and you know, one one that requires us to sort of think about where we are, where we want to be, and all of the tremendous barriers that lie in between. But for what it's worth, how do we get there?

SPEAKER_03 21:54

So I I think thinking about suicide again, that you know, if you're a clinician, you see many suicidal patients. Patients, but that suicide death itself is a very much a low base rate phenomenon. You know, it's really hard to develop a data set where you're going to be able to sort of look at markers of suicide death with any sort of validity. And even attempts you can find individuals who are at high risk, but you're still going to, you know, it's like 10% or something are going to go on and engage in the repeated suicidal behavior. So I think a lot of it comes down to data sharing and sort of tracking across groups, you know, in terms of making these next generation measurement instruments. It's really going to be about sort of again creating these networks or combining data to really understand that. I mean, I think there is the place, like I said for so this innovative new task development, but then that's going to have to be validated at scale to really figure that out. And again, there has to be the willingness on the part of the researchers, on the institutions and organizations to support this kind of work because it is risky. It is hard. There are a few, you know, sleepless nights when you're worrying about your people. And so there has to be some sort of agreement support understanding of the importance of this work, but that it's also bumpy. It's it's it's hard work, but it's really critical. Like the suicide rate isn't going to go away just because we stop doing these trials. If anything, it's going to get worse.

SPEAKER_02 23:34

Thank you. And you know, that that certainly that certainly resonates. I think there are a lot of similar challenges in other areas that that will likely probably follow a similar path if we can get there. Just before we finish up today, one more question for you, if I may. You know, what advice would you give you know individuals who are who are designing protocols for clinical trials or other sorts of studies? Uh, what advice would you give them in in neuroscience when suicidality isn't the primary focus, but you know it's a phenomenon that you're going to encounter? What are the three things you really want them to do? You know, they're they're going to have to do certain things. There are certain instruments they're going to have to include for risk, for risk and reporting purposes. But aside from the necessities, what are the three things you really want to see them start doing?

SPEAKER_03 24:33

Walk, I guess, walking through what's going to happen when you get a positive response of what is exactly like who is going to do what? You know, we you every group usually has sort of a naysayer, you know, who's going to take the most negative view, you know, sort of think through like, okay, what's the worst possible scenario? So really walk through what that's going to be. Really sit down and whatever assessment measure you look at, just really look at it yourself and sort of think about how you would approach this, especially without maybe all of the sort of mental health training that we have or the assessment training you have. Like, was this done? Is this going to be very surprising to folks? Are they really going to understand sort of what is meant by that? There's lots of different measures that talk about suicide risk in lots of different ways. And so, sort of taking a moment to think about that. And then consulting experts, whether that's, I mean, ideally somebody with lived experience or somebody else that has a bit of expertise in that. I'm very used to people, you know, at NIH, you know, do diabetes or sort of worried a little bit about their suicide or depression risk assessment. And so, you know, a half an hour phone call can go a long way. And so don't be afraid to sort of seek out sort of expert guidance on this. Because what you don't want to do is have the worst possible thing happen on a Friday afternoon. You know, somebody reporting something and they've already left the office. Like, what are you going to do with that? And so having people who've been there who've really thought about this deeply can be very helpful.

SPEAKER_02 26:10

Thank you, Elizabeth. I really appreciate your expertise, your time, and I've enjoyed the discussion. Once again, everybody, I've been talking today with Dr. Elizabeth Ballard, associate scientist at the NIMH intramural research program. Thank you again, Elizabeth.

SPEAKER_01 26:29

Mind the methods brought to you by the International Society for CNS Trials and Methodology.