So, I'm Paul Appelbaum, I'm a professor of psychiatry at Columbia, and I'm the chair of the DSM Steering Committee for APA. And today we're going to be focusing on a new diagnosis, which was added to the DSM just in the last two years, and first published officially in DSM-5-TR, which came out last year as you all know, and that's prolonged grief disorder. And we're going to be hearing from some of the key people who were involved in the process about the disorder itself and about the process of adoption. But I'm going to start off with a more general introduction to the question of how a new diagnosis gets added to the DSM. And then we'll shift towards to focusing more on PGD, prolonged grief disorder, itself. And so this is what I'm going to do, and when DSM-5 was published in 2013, the Board of Trustees of APA made a decision to move to a different process of revision. You all know the old process, right? Every 10, 15 years or so, a whole new edition of the DSM came out. There were multiple workgroups appointed to revise each section. Lots of changes were made, and a new edition was published. And in between those editions, essentially no changes were made, even when they might arguably have been warranted by the introduction of new evidence. And in some respects, that process reflected the first of the bullets here, which is that stability in psychiatric diagnosis is, in general, a good thing. It promotes consistency in our clinical practices. It makes training a new generation of practitioners easier. And it promotes stability in research as well. Older data sets can be added to newer data sets. They're using the same criteria and enhanced power results. However, the Board of Trustees recognized, to its credit, I think, that advances in the field might periodically warrant revision of the DSM, but that those advances were likely to occur at different rates for different disorders. And it made little sense to hold up the introduction of a new disorder or changes to the diagnostic criteria for an existing disorder while waiting another 10 years until a new edition of the DSM appeared. And with the introduction of online publishing and the online version of the DSM, we suddenly had a mechanism for updating the DSM in a more timely fashion. And since 2014, we've developed and put into place a process of iterative revision of the DSM, and that's what I'm going to tell you a little more about. Today, proposals are invited from anyone who's interested for changes to the DSM. And these are the specific kinds of changes that can be considered, although we're going to focus on the addition of a new diagnostic category because that's the topic of this session. Proposals are not just, I have a great idea. We need a new diagnostic criteria set for my pet disorder. But they must be accompanied by substantial evidence that it would accomplish, adoption of a new proposal would accomplish all of the following. It meets the DSM criteria for being a mental disorder. That should go without saying. It has strong evidence of validity. It has at least moderate evidence of reliability in the use of the proposed diagnosis. There's substantial clinical value attached to it. For example, it will help identify a group of patients who are not now receiving appropriate treatment but would if the disorder could be more precisely specified. It doesn't overlap with existing diagnosis, diagnoses, and it's not better thought of as just a subtype of an existing diagnosis. And overall, it will have a positive benefit-harm ratio. For example, low false positive rates, not very much likelihood of harm due to social or forensic considerations. This is a screenshot of the submission portal at dsm5.org. You can also reach it through psychiatry.org, which is the APA website. And when you click, it just takes you through the process of making a proposal and providing the appropriate evidence to support that proposal. And sometimes we get submissions which include 100 pages of text and tables and evidence in support. What happens when we receive such a proposal? Well, initially the steering committee, which has 12 members, reviews the proposal and asks on its face, does the evidence that's been provided for the proposal seem likely to meet the criteria for approval if we assume that everything that we're being told is in fact accurate? The answer to that question is yes. The proposal is then referred to one of five review committees for more in-depth exploration and if necessary, proposal of revisions to either the criteria, ultimately to the development of the text for that criteria. And if they recommend approval of the proposal, it comes back to the steering committee, which gets another crack at it if the steering committee is comfortable with the recommendation. It's then posted for public comment for either 30 or for larger, more substantial proposals, 45 days. And the field is notified of its availability for comment. So you will see a notice in your weekly bulletin, email bulletin from Psych News. We'll also directly contact subspecialty societies that have a particular interest in a given diagnosis to let them know that they should be letting their members know that this is available for comment. When comments are received, it comes back to our process. The steering committee typically asks the review committee to take another look at it and decide whether anything needs to be changed, after which the steering committee undertakes final review and answers the question, does substantial evidence exist to demonstrate that the proposal has met the criteria for approval that I outlined for you a few moments ago? If the answer is yes, the proposal is referred for review and approval to the APA Assembly and the APA Board of Trustees, which must both approve the inclusion of new diagnostic criteria sets or any other major changes to the diagnostic criteria for the DSM. I mentioned that we have five review committees. They each cover broad areas of psychopathology. There are standing committees, but they have the capacity to invite consultants for particular proposals. So if they need more in-depth knowledge of a field, they have a way of obtaining them. And they're deliberately chosen to include people with a broad scope rather than narrow area specialists so that they can see the proposals in the broader context that clinical and research psychiatrists would be likely to view them. So that's the process. We've got a couple of minutes if there are any questions at this point for us to entertain them and then I want to tell you a little about how that process worked for this particular proposal on prolonged grief disorder. So if there are any questions at this point, feel free to come to the mic, either mic, and if not, we'll just push ahead. So there is now out there in the literature, recently published this month, in the Journal of Nervous and Mental Diseases, a history of what we're going to be talking about today. A team led by Peter Zakhar has been doing a series of these historical pieces about various psychiatric diagnoses and how they came to be. So a more complete version than we can possibly give you today of this story is included in that article and I commend it to you. We were all interviewed for that article, as were many other people who were involved in the process. So the background to prolonged grief disorders ultimately being included in the DSM-5. During the process of developing the DSM-5, published in 2013, there was a proposal made for the inclusion of a pathological grief following loss of a loved one diagnosis in the DSM, but because the two major research groups that were working in this area, both of whom you'll hear from today, were not in alignment on the criteria for such a diagnosis, the DSM task force decided that the wiser course at that point was to put it in the chapter on conditions for further study. And at that point, this condition was called persistent complex bereavement disorder. In February 2018, using the process that I showed you, the DSM Steering Committee received a proposal that was submitted by a team led by Kathy Shear, who you're going to hear from, and Chip Reynolds, which was one of the two research groups that I referred to a moment ago, to incorporate as a formal DSM diagnosis a slightly modified version of that persistent complex bereavement disorder, which was in that chapter on conditions for further study. The Steering Committee took a look at it and referred it to one of our review committees, the Internalizing Disorders Review Committee, and you're going to hear next, when I sit down, from Roberto Luis-Fernandez, who is the chair of that review committee. The review committee conducted an initial review and concluded that, as is usually the case, they would like some more data. That is, that on its face, the proposal wasn't an open and shut case for the adoption of this diagnosis as formulated. And there was also, frankly, concern that there were two major groups working in this area and only one of them was represented in the proposal itself. The second group was led by Holly Pergerson, and you're going to hear from Holly today as well. So the review, which took 11 months, I think, 10, 11 months, and the request for additional data was a triggering event, if you will, for all the frustrations that I think everybody who worked in this area felt from the inception of the process of DSM-5 forward. So there was 15 years of frustration here about the failure to persuade the DSM authorities that this warranted inclusion in the manual as an official diagnostic category, and we reached a stalemate. The proposers declined to provide additional data or revise their proposal, and it wasn't clear how we were going to move forward here. We were faced with a paradox, which is almost everybody agreed that pathologic grief was a real thing. It was a real phenomenon. Everybody had seen it in some patient or other and sometimes in many patients, and many people agreed that it was deserving of DSM recognition. But moving to construct a valid diagnostic criteria set in the face of continuing differences between the major research groups in the area seemed somehow an insurmountable obstacle. So we needed to find a way forward. So I reached out because I'd known Chip Reynolds, one of the proposers, for many years since we were junior faculty together. Many years ago in Pittsburgh to discuss how to keep this process going and bring people together in some sort of consensual way. And Chip analogized to an NIH consensus development conference where people with different perspectives are brought in. There's a panel that listens to them, and that panel makes recommendations as to how best to resolve the differences in the field. And we decided to try a modified version of that. And so to the credit of the APA administration, they agreed to fund a workshop in June of 2018, which was held in New York, fortunately pre-pandemic, with representatives from the DSM Steering Committee and APA leadership. Saul Levin himself came to this meeting. And the leaders of the two research groups, both Kathy and Holly and their closest collaborators, plus a child research group at UCLA that has worked on the child and adolescent version of this condition. And we had set up this criteria development panel, this independent panel of experts, which was just chaired by David Brent, who's head of child psychiatry at the University of Pittsburgh and an expert on these issues. Each side, each group presented its data, and that was followed through the course of the afternoon by a facilitated discussion. And by the end of that afternoon, we were at the point where there was substantive agreement around the table on a tentative set of diagnostic criteria, which didn't look like anybody's favorite set of criteria, but represented some compromises on all sides. Crucially, because this was a set of criteria that was agreed to by experts sitting around the table, both groups agreed that they would take these proposed criteria and go back and test them to ensure that, based on their data sets, and there were multiple data sets involved, that in fact they showed evidence of validity and, as best we could ascertain, reliability as well, sufficient to warrant adoption. And by the end of the summer, shortly after Labor Day, we had the findings from both groups back and both groups' findings converged. They were supportive of the identified criteria from that consensus process. They identified the same symptom cutoff, the same number of symptoms required independently, and that gave us substantial confidence in what had been agreed upon. The panel that I described to you, led by David Brent, reviewed the data, made some suggestions for resolving some open issues, for example, what duration of symptomatology or how long after the death, if symptoms are present, should be required. It was different for kids than for adults. In the end, 6 months for kids, 12 months for adults. And all parties agreed to the final form of the proposal. So the final version of the criteria set, based on the confirmatory analyses that both groups did, was reviewed and approved by the Steering Committee and posted for public comment. Those comments, and we got a lot of them, largely supportive, but some making suggestions about wording that might be less than perfectly clear, were sent back to the Review Committee, which massaged some of the wording, just to clarify it, with agreement from the research groups. Steering Committee approved it. It was referred to the APA Assembly and the Board of Trustees, and in the fall of 2020, both those groups approved the diagnosis, and Prolonged Grief Disorder became an official diagnostic category, which is now part of the DSM. So, that was the process, and if there are questions now, we can take them. We're also going to have time for questions and discussions at the end of the presentations, but I see we've got a question. Yeah. Thanks so much. I wonder if there's any connection. This work happened after DSM-5 and the issue of the bereavement exclusion from major depression. That was taken out. Is there any connection here, or sequence, or does that fit into the story at all? There were changes in the DSM-5TR to try to recognize that we now had a Prolonged Grief Disorder diagnosis. There's still a bereavement exclusion for major depression, but the text now indicates in multiple places the availability of a new diagnosis, which could be applicable depending on the context of the presentation of the patient. So there were multiple changes made elsewhere in the manual, which is often true when a new diagnosis is introduced. Okay. So I'm going to turn this over to Roberto, and I'm going to go back to my laptop where I've got my introductions. My introduction's written out. Roberto, who is a colleague of mine at Columbia, where he's professor of clinical psychiatry, also directs the New York State Center of Excellence for Cultural Competence, and is research area leader for anxiety, mood, eating, and related disorders at the New York State Psychiatric Institute. As I told you previously, he is the chair of the Internalizing Disorders Review Committee for the DSM. Roberto. Thank you very much, Paul. It's a pleasure to be here to discuss this process. One more point that I would throw into the mix is that this was the first substantive addition of a new disorder in the DSM revision process that Paul has outlined. So part of the delay and the complexity of figuring this out was that we were the trial balloon, so to speak, of the diagnosis with this new approach, and so had to work out lots of things as we went along. I don't have any disclosures. I wanted to focus my presentation, our thinking about this disorder, with this goal, if you will, how I approached the review process. I'm speaking personally, though I in some ways represent the committee, but this was the way I framed the problem of how and whether to include a prolonged grief or pathological maladaptive grief disorder category. It's how can we develop a reliable, culturally valid set of diagnostic criteria that distinguish maladaptive from adaptive grief in order to avoid over-medicalization and help identify individuals who may benefit from treatment. And the over-medicalization element was something that I took and the committee took very seriously, because everybody on earth will undergo this stressor. Everybody on earth will have a close friend or loved one who passes away, and so we wanted to make sure that our criteria, that the criteria set that was finally approved, distinguished between people who were having an extreme and maladaptive form of grief from the regular grief that everybody in the world would at some point go through. We didn't want to diagnose normal grief. Nobody wanted to do that, so we had to make sure that the criteria was, and everybody would, I mean, I'm not saying mine was, this was my idea and nobody else's. This was the focus of the review that everybody, was an element of everybody's thinking. Now, as Paul has mentioned, and I will go through this relatively quickly and mostly emphasize the parts that, emphasize some parts that Paul mentioned more generally, we all agreed in the review process that there was a lot of research already done on maladaptive grief. Nobody doubted that part of it, and we also generally agreed that there was a need for a new diagnostic category, and some of the reasons for it are there in the subtitle, there in the subheading, that we, by including it in the official diagnostic classification system, people would be more aware of the diagnosis. It would be more looked for, better detected. There was a diagnostic gap between people who met criteria for major depression and people who had a traumatic-like grief experience. There was a gap there in the middle of people with a set of experiences who didn't meet necessarily criteria for MDD. One study estimated that between 40 and 50 percent of people with pathological grief or maladaptive grief did not meet major depressive criteria. And if the death wasn't traumatic or if it had some certain constellation of symptoms, PTSD would not be either the appropriate diagnosis. Sometimes a person may have both, but it was quite possible that they might fall through the cracks. And there was a big incentive for getting the diagnosis right and including a diagnosis in the manual, because there are, there were studies, a lot of them done by Kathy's group and other people, that showed that treatment mattered, the diagnosis of grief mattered for treatment, that there were specifically tailored forms of grief therapy that would be more successful in people with a grief presentation. And, for example, major, sorry, antidepressants would not be particularly successful in people with grief without major depression, for example. And so there was a need for a diagnostic category. At the same time, as Paul mentioned, and I won't belabor the point, there was a debate about the exact types of symptoms that would be included. And as Paul mentioned, there were two general categories of ideas about concepts. Just very briefly, one category, and I hope I'm not oversimplifying and the experts are here and they can tell me if I do, but one category generally favored the notion that it was like grief but too long, that generally the phenomenology of the pathological condition was similar to normal grief, it just lasted a very long time. And there was another camp at the beginning that argued that there were more phenomenological differences that could be characterized between maladaptive and normal grief above and beyond duration. This was one of the major debates, and issues like that, there were many more details, needed to be hammered out, and Paul already mentioned the Solomonic decision of DSM-5 about splitting the baby, so to speak, and trying to create a hybrid category between these. Nobody much liked that one, so far as I can tell. And then another thing that happened that this one, I should be added to the story, is that in between DSM-5 and 5TR and the revision of our committee, ICD-11 approved a prolonged grief category that was essentially made it official, an official part of the nosology, and it had a six-month duration. And one element that was emphasized, that was a big focus for me about over-medicalization, had to do with the response to that diagnosis of a six-month duration of a prolonged grief disorder. In some quarters, there was a nice study from Germany of clinicians finding that about half of the clinicians in Germany who responded to this survey thought that the six-month criteria was too short, that it tended to have the potential for over-medicalization, and that it may be ignoring some cultural, national variation in grief, normative grief duration. Forty-nine percent versus 11 percent were against and for the six-month duration. So these experiences were part of the hopper, part of what was in the picture. Now, in terms of these steps that Paul mentioned, just to give a little bit more detail, the review committee agreed that there was sufficient evidence for the construct of maladaptive grief on a number of domains. One is that it did meet the DSM criteria for the definition of a mental disorder, that there was sufficient validity and reliability evidence for the construct in general, that there was great clinical utility, positive benefit-harm ratio, and a non-overlap with a key possible alternative diagnosis of major depressive disorder. So there was a lot of consensus around these points. At the same time, the review committee felt that we needed more evidence to finalize the exact criteria. And this was, for example, like in the realms of item selection, exact duration. I already mentioned the difference between a six-month and a 12-month duration requirement, the caseness threshold, how many symptoms were required to be considered a case, the psychometrics of the exact criteria set, the predictive validity, and some cultural and development-related details or specifics about the type of the criteria set that should be used. And so, as Paul mentioned, I won't again go into more detail, the breakthrough came from bringing everybody together. It was a very one long day, but a very successful day, I think. People were very willing to hear each other and reach an agreement. And we hammered out a criteria set for further study, as Paul mentioned. And from my perspective about over-medicalization, we built in some safeguards against over-medicalization. For me, that was a key element of the discussion, and I think everybody shared it. So we agreed on this 12-month duration period for the time since the death for adults, and a six-month duration for children. This was one of the later developments that came from the work of the child and adolescent folks. There was a fairly high threshold for symptom frequency and severity, so that there needed to be a persistent grief response since the death most days, to a clinically significant degree, and the symptoms had to be present, the symptom threshold had to be present nearly every day in the last month before presentation. So there are four components there of phenomenology that people felt were sufficiently severe. It indicated a more extreme reaction, so that we would tend to reduce the risk, not that over-medicalization is ever reduced, but that it would help to reduce the risk. Then there was criterion D, about clinically significant distress or impairment, which is common across disorders, but very important particularly in this one as well as others. And criterion E, that the duration and severity of the bereavement reaction clearly exceeds expected social, cultural, or religious norms for the individual's culture and context. And of course, that requires the clinician to understand what is the individual's culture and context, which is not a simple thing. The cultural formulation interview, for example, in section three, is attempting to address that kind of question. It is a clinical, cultural decision, but it's something important to figure out whether it fits with what is considered normative in that society or group. Then as Paul started to discuss, we reviewed the psychometrics. We were very lucky that both groups had some excellent data to which this new agreed-upon consensus, if you will, criteria set, could be tested against. And I'll just briefly mention some of the main findings that were written up after the process, very usefully. That there was the symptoms, which I'll show you the exact criteria in a minute at the end. The symptoms that were chosen for the criteria set were unidimensional in factor analysis. There was a single factor to which they all belonged. There was a high internal consistency. For example, in three sites, the Cronbach's alpha independently went between 0.8 and 0.9, which is very good. This is internal consistency data. The cutoff that was selected of three or more symptoms was selected precisely because empirically in nine data sets, in the majority of those nine data sets, they optimized sensitivity and specificity of the condition in these data sets using previously obtained data with those symptoms and characteristics. And then that there was a good, this is just more psychometrics, a good test-retest reliability after seven months using the criteria in one of the data sets. And in another data set, a good intra-class correlation between that in terms of the internal test-retest reliability. Another set of data involves the predictive validity of these criteria sets at five to 12 months after diagnosis. So you have a diagnosis at a certain point, say at 12 months, and then you follow them for, or you assess them again five or 12 months later. And the cutoffs that we were recommending as a committee based on the previous data I was discussing was confirmed by, prospectively, by looking at clinician assessment of caseness. So independent of the disorder criteria, the clinicians agreed it was a case of maladaptive grief, functional impairment, suicidal ideation, quality of life, treatment response, state trait anxiety, and comorbid diagnosis of major depression and or PTSD. So these were all predictive validity elements that were satisfactory, were good with respect to the diagnostic criteria that were selected. And out of the nine data sets or more that we assessed, there were several from multiple countries, including these five that you see here, U.S., Turkey, the U.K., Taiwan, and the Netherlands, rather distinct societies. So there was some confidence that this was an internationally acceptable data set. So the next steps Paul mentioned, it was put up for public discussion for 45 days. I printed out the responses. We have 53 pages of single-spaced responses. Some were very detailed. I went through the comments very carefully, prepared tables, submitted to the committee. We discussed all of them. Some were accepted. Some were very useful. This is a little unclear. So we would fix that or other little bit more substantive comments. There was discussion about the name of the disorder and so on. We incorporated as many as we could. Some people, of course, object to the whole idea of pathologization of grief. Those were also recorded. But the majority of people were in favor of the idea, who responded, of course. It's a self-selected group. We're in favor of the inclusion. Now I really came away with a sense, confirmation of the value of an iterative review process that is super important to get multiple perspectives, including public review. That was, I found, very helpful to me. And then there was, of course, the major sets of institutional sign-offs that Paul knows better than I do about the Board of Trustees and the Assembly and everybody involved about making sure that this fulfills the requirements for the organization and in some ways the representatives of the profession. This is my last slide. Oh, I should also mention, I'm a cultural psychiatrist, so I mentioned this particularly, that there is a fair amount of attention to cultural variation and that kind of points, not only in that criterion that you see there as Criterion E, but also in the text of the disorder. In the writing of the chapter, Description of the Disorder, in the book, there's a fair amount of attention to the question of cultural variation. So these are the criteria that are now official. Exposure to the death of a person who was close to the bereaved individual for a certain period of time. Adults, it's 12 months, six months for youth. Since the death, since the death, there was the development of a persistent grief response with one or more of the following two gait symptoms present most days to a clinically significant degree and nearly every day for the last month, either intense yearning or longing for the deceased, which is the main symptom, the main gait symptom in almost every study. Yearning is the key defining feature. Some people, some studies estimated at like 15 percent, don't have yearning so much, so it's the second catch there is about more general preoccupation. Most people have yearning, at least in the databases I remember. And then there was the set of descriptive symptoms that you also had to have in addition to the gait question. These were a threshold of three or more. Like I said, that was empirically calculated from the data set, maximizing, optimizing sensitivity and specificity. Three symptoms, most days, clinically significant degree, nearly every day for the last month. And you can see the eight of them there that are all related to the death, one of which has a specific child adolescent variation embedded in the symptom description, and there's more in the text. But they involve all kinds of different separation distress, reactive distress, role or identity conflict. There are several dimensions of grief reaction that are listed. And then there's the distress or impairment criterion and the cultural variation and cultural normativity criteria. That's it for me. Any questions? Should we take questions now or later? I think we should hold them. Till the end? Hold it till the end. Very good. I will then do that. We'll have time for your questions and comments at the end. Thank you, Roberto. So as I indicated previously, this process that we're talking to you about was started when we received a proposal which was formulated by Kathy Shear and her colleagues. Kathy is the Marion Kenworthy Professor of Psychiatry at the Columbia School of Social Work and at the Vagelos College of Physicians and Surgeons at Columbia and Director of the Center for Complicated Grief at Columbia. Kathy? Can I do this from here? You should see them, I believe. Okay, great. So thank you. Thank you, Paul. And it's really a pleasure to be here and to be able to share this perspective. I think this diagnosis has probably caused more kind of controversy than maybe any other in DSM, probably because grief is such an intense experience. So I was the lead of a proposer in this round, as Paul described, of the proposals, but I was by no means the main person involved in this because all of us really worked for over a year, actually, to prepare the document that Paul sort of introduced you to. And this group, as you can see, has expertise in a lot of, basically most of us had worked in mood anxieties or trauma and we had a senior biostatistician, Bill Avori, and the biostatistician that worked with us in our studies in New York, Christine Mauro. So why did we think that a diagnosis was needed? And it was because basically we were seeing people who were clearly experiencing distress and impairment and they could be reliably identified, which I'll say a little bit about in a minute. I'm talking about right now the proposal that we made in early 2018. And also we knew that grief was not a focus in most mental health training programs and many clinicians still were thinking, even at this point after we had published three treatment studies, there's a lot of publication in the literature about this, which Holly may talk about as well. And they were still being diagnosed as depressed and we had shown, which I'll show you in a minute, that high quality treatments for depression were not helping them and no efficacious treatment initially was available, so that's how we got involved. And eventually we and other groups around the world have shown that there are now available treatments that are efficacious for this condition, which is, we thought was very important, the combination of good treatments for depression not being helpful and in this. And also, in addition, there was extensive, by this time, published data confirming the importance of the diagnosis. So I want to introduce you to three people that we saw early on. These videos are kind of activating, so just to alert you to that. My involvement in the treatment was around the death of my mother. However, my husband died about two or three years later, and I think that really exacerbated the situation. But my mom was almost 85 years old, had really only been ill for less than a month, and she passed away. We were very, very close. I could not imagine living without her. I never really recovered, because her death left this gaping hole inside of me, and although my husband was still alive, he wasn't my mom. So I sort of just kind of wandered around through life. I recall looking out my bedroom window and being a little angry because everybody else in the world seemed to just be going on their merry way as though nothing has happened. I'm thinking, how can you do this? Don't you know that the world has lost somebody, you know, wonderful? My 19-year-old daughter Jackie died in a car wreck on August 21st of the year 2000. She was out that evening and was told to come home at midnight. She got home about quarter to 12. She was real excited, asked me if she could go out and say goodbye to one more friend, because we were taking her to college on that Tuesday morning. And she left, and I went to bed, and we got a phone call at about 3.15 in the morning asking us if our daughter was home. I jumped out of bed and went to her room, and she wasn't there. There wasn't not a whole lot that I remember about that morning, other than sitting in total disbelief and didn't want to eat, didn't want to do anything, spent most of the day crying. It felt like somebody had tore my heart out. It was my little girl. Well, as time went on, I found that I wasn't progressing or being able to focus myself on work, on anything that I'd like to do. All I could find myself doing is going to work, getting some work done, going in the back room crying a lot. The more I thought about it, I realized that I wasn't doing anything with my life. It's been 20 years. My husband died at a relatively young age. He was 53. And I got stuck. I just couldn't move on from that. I just grieved and grieved and grieved. And I thought that somehow that it was my fault that I hadn't been able to diagnose his tumor. So then when he died, there was just this giant hole that just seemed impossible to fill. For years, I had horrible nightmares. And I'd wake up in the morning, and I'd think, no, this did not happen. He's still here. I didn't care whether I lived or died at that point, to be honest. But I had this child. So I would lay on the couch and cry every night. And then I would get up in the morning, and I would go to work and do my job. And then I would come home, and I would lay on the couch and cry. And I did that for years. That's all I could do. I was paralyzed. So these were people who were presenting for treatment, as you can see, back in the late 1990s and the early 2000s. And I'm a treatment researcher. And I was very interested in being able to identify these people. But I really basically was interested in trying to figure out how to help them. And I want to acknowledge, before I go on, two very important early contributors, and early and continuing contributors, I will say. The first one is Marty Horowitz, who actually proposed including a form of PGD in DSM-IV as a stress response syndrome, refocusing us on that idea that maybe this isn't really depression. Because, I mean, I had learned that in my residency training, that grief was a form of depression. I think a lot of people probably today still think so. But Marty really said, no, this is really a stress response syndrome. Talk more about that later if you want. And then Holly really gets enormous credit for this whole entire enterprise, because I think without her, we would not be sitting here today. I don't think we would have a DSM diagnosis of prolonged grief disorder. And she's done a huge amount, which I know she's going to tell you more about. But in the beginning, when we were working together, she had developed a simple, reliable symptom measure, which is absolutely essential for doing research. And so that's one of the many things that she did, was to introduce the world to the inventory of complicated grief. So a lot of the survey studies that actually support the diagnosis, as well as our treatment studies, were really possible because of that instrument. And just parenthetically, I wanted to point out, because this always comes up, but this syndrome is essentially the same as what we've always called unresolved grief, complicated grief, traumatic grief, and then eventually persistent complex bereavement disorder. They're really all the same. I'm pretty sure they're all the same group of people. It's just how you exactly identify them. That's a whole other story. That's what we've been grappling with for many years. So what we did, as I said, is develop a treatment. We started with a pilot study. Using actually that idea of a stress response syndrome, I reached out to Edna Foe. I was an anxiety disorders researcher before I started working in grief. And I knew Edna personally, and those of you familiar with her might know that she really developed and extensively studied in the community, as well as in the clinic, prolonged exposure for PTSD. It was a highly efficacious treatment. So she came and did some workshops, and we kind of used that as a basis to start with. And we did an initial study and found that both data and lived experience, because that's always how we've operated, is by, we do pretty rigorous clinical research with lots of data collection, but we also carefully listen to our patients, as I'm sure everyone in this room does, and sort of document and talk to them about, they provide various feedback like this, as do our therapists. We also only use therapists who are actively doing therapy in the community. That's just how we've operated. So the pilot study was quite promising, and we then got two sequential NIMH-funded studies to compare this treatment, which I'm calling PGT, although at the time we called it CGT. I'm just trying to minimize the confusion that we have around the terminology. There's enough confusion around other things. So we compared this grief treatment to IPT, which those of you who know IPT would know that it has a grief focus. It's a wonderful treatment for depression. We had IPT experts at the University of Pittsburgh, and they led the IPT. Ellen Frank was the person who supervised, and actually the IPT therapists were all trained before the study. The PGT therapists obviously were not. But in any case, we found basically in the pilot study, we found almost about twice the response rate for PGT compared to IPT, and that's what we continued to see. So in our first study, which was mostly middle-aged adults, a few older, but mostly middle-aged, we had a 51% response to PGT and a 28% response to IPT. In our second study, which was all older adults, we had a stronger response to PGT and about the same response to IPT, as you can see here. And then we did a third study, which was a foresight study, which included Chip Reynolds, Naomi Simon, and Sid Zizook. They were all co-first authors of our publication in GEMMA Psychiatry in 2016. There we found that PGT was significantly more effective than citalopram, in this case there was no IPT in that, that's an error on this slide, than citalopram. And we had, in this case, an overall 70% response versus 40% response. But we also found that citalopram, when administered with prolonged grief disorder therapy, was significantly better in reducing symptoms of depression than PGT with placebo. In other words, citalopram was working perfectly well for the depression and it actually augmented the outcomes of depression, but not the outcomes, the grief outcomes. Finish up, yes, okay. That was true of suicidal thinking. I think I will actually stop at that point. I think that's a good point, a good place for me to stop, because that's basically what I wanted to say. Yes, I can. Okay. Thank you, Kathy. Our last presenter is Holly Prickerson. Holly was there at the inception of the concept that ultimately became prolonged grief disorder. She's now the Irving Sherwood Wright Professor in the Department of Medicine at Weill Cornell and co-director of the Cornell Center for Research on End-of-Life Care. And as you heard, one of the main goals of this process that we've been describing to you was to bring Holly and her group into this discussion and come up with a consensual approach to PGD. Holly. So first of all, it's a pleasure to be here, and I was asked to talk about my perspective on the process. So I assumed, I guess rightly, because Paul was okay with what I'm about to say, but I had a different reaction as someone who's not a psychiatrist about my perspectives on the process. So I have 10 minutes, and so I'll provide my piece. I have no conflicts of interest, which I actually think is a really important statement, especially given the concerns of over-medicalization that Roberta was referring to before. So my perspective is that I feel like psychiatry and the DSM have an ambivalent relationship with evidence and the data. And so the case for the data and the evidence are PGD was born from data. PGD was identified and validated by evidence confirming psychiatrist's observations, including Freud, Lindemann, Bowlby, Parks, and Horwitz. So data, they had observed similar things, but it wasn't until data actually showed that they were right about a lot, not everything, but about a lot, that things started, the field started moving forward. You needed to measure things. Chief Reynolds' data, trial data of tricyclic antidepressants and interpersonal therapy for bereavement-related depression revealed that the grief symptoms didn't resolve with the same clarity as the grief symptoms. So that led to many questions, questions, why was it the case that those symptoms weren't resolving if depression seemed to be effectively treated by tricyclics and IPT? Did it even matter? That was actually the first question. Why are you concerned? Why be concerned about grief if, as long as our primary target, our primary outcome of depression is being treated, why worry about grief? Grief is normal. Grief is the normal reaction to loss. So data were used to set out and test several first basic questions. One of those questions was, are these symptoms different? In the context of bereavement-related loss, are there symptoms of yearning, pining, intrusive thoughts, identification symptoms, feeling the person's presence? Are those different than being sad, a generic low mood? Are they different than being anxious and worried? And this is one of many, many factor analyses that have replicated over and over again with data showing that the symptoms cohere and they're different than the symptoms of bereavement-related depression and anxiety, and even inversely related to some of these symptoms. So it's inversely related to a lot of the anxiety symptoms, which I think suggests the distinction between traumatic stress responses and this syndrome of more of a, I consider it more of an attachment disturbance. I think Roberto referred to some of the data, some of the, we concatenated and compiled five international data sets of community-based bereaved people. And in this publication of World Psychiatry in 2019, we showed that in the Yale data, the Utrecht data, the Oxford data, it was not only significantly correlated with a whole host of bad outcomes that I'll be mentioning soon, but it also had what they call incremental validity. Over and above competing diagnoses, it was predicting increases in suicidality and bad physical health. Data were used to debunk myths about, promulgated in the media about problems with a new diagnosis for grief. So I don't know how many of you saw the New York Times article, How Long Should It Take to Grieve? Psychiatry's Come Up with an Answer. That was really a sensational unfair title. Any normal person in the community hears a title like that, the automatic response is going to be, who are these psychiatrists? Who are these experts? Why are they telling me how long I should miss this person? This is outrageous. And that wasn't helpful to the public's perception of what the disorder was or the intensity or rarity of that disorder. Then there were similar headlines about the grief pill is coming, we're going to make you forget about your loved one, the big pharma is trying to over-medicate this disorder to take your grief away. So data were used to debunk these myths promulgated in the media and elsewhere. What are the myths that data were used to debunk? One was that PGD would diminish family support and stigmatize mourners, and it medicalizes grief to push unwanted treatment, that that's the agenda, is to push unwanted treatment. In the Yale Bereavement Study data, we had collected data from about 150 bereaved, largely widows, and it wasn't culturally diverse, granted a relatively small sample, and what we found was that 90% of them in that survey, when asked, said that they thought their family would actually be more understanding of what they were going through. If they met criteria for this particular disorder, 96% said that they would be more and relieved to know that they had a recognizable problem, 96% said that they'd be relieved to know that they weren't going crazy, and most importantly, 100% of those who met criteria said they would be interested in getting treatment for that disorder. Data debunked the myth that the criteria are so easy to meet. Everyone has grief. They're going to diagnose everybody. They're pathologizing a normal, typical response to loss. A representative of a very well-conducted study in Germany found 4% prevalence among the bereaved sample, only 1% among the non-bereaved. The false positive rate, compared to clinicians agreeing upon what is a diagnosis and isn't, one out of 100, very, very low, would be falsely, wrongly considered to have that disorder when they didn't. The diagnosis requires severe distress and impairment, so pathology is baked into the diagnosis, and as I showed a slide that I'm sure you couldn't see, the numbers and what it was showing, but the basic idea is that it had demonstrated predictive validity in multiple international data sets showing increases in suicidality over above competing risks, increases in depression, increases in bodily pain, worse general health, worse vitality, worse social and role function, and even in heart attacks. There's been studies in Australia showing higher rates of heart attacks associated with high levels of grief. Another myth is that the evidence was thin and there was a rush to judgment and that there wasn't enough strong evidence to support adding this to the DSM. As of last August, there were over 600 peer-reviewed articles, 10 systematic reviews in the past five years, ample evidence demonstrating from multiple researchers that this is a thing and it's a bad thing and that existing treatments largely don't work for it. So despite how data was available, from my perspective, I'm sure not other people's in this room, but my perspective was that DSM-5 equivocated. In 2010, there was ample evidence of all these things, of predictive validity, reliability, sensitivity, specificity. There was enough evidence for all the things that Paul and Roberto were saying were requirements. There was substantial evidence at that point, from my perspective. They punted and they decided that there was a perceived lack of consensus between two groups and politics trumped the evidence that was available at that time. At the 2019, I guess it was 2018 meeting, Reynolds presented a single case study, N of 1, as evidence to support PGD's inclusion in DSM-5. It was lauded as an inspired choice. As a researcher, an N of 1 is insufficient. I know clinicians probably differ. It helps characterize the disorder, but in terms of evidence supporting that a new entity should be added to the diagnostic nomenclature, that seems insufficient. Even for this symposium, I was invited initially as another expert on PGD who was asked to join the process to offer her perspective, quote, unquote. To me, it feels like I'm being invited to a party that I and the data started. From my perspective, psychiatry is an ambivalent relationship with data. Evidence is needed to confirm diagnostic formulations, but it's been marginalized by political specialty interests when it should be the other way around. That said, I'm very grateful that the DSM ultimately did require empirical confirmation before PGD was included. Evidence did inform the ultimate criteria set, which is now an official new mental disorder, and the data in the future will be needed to confirm the wisdom of PGD's inclusion in the DSM-5. But given the ample evidence to date, that's likely to happen. I also wanted to thank Dr. Appelbaum and Kenlor for the invitation both to present to the DSM workgroup in 2018 and to Dr. Appelbaum for the invitation to present here today. So what is needed to change the process of how the disorder got into the DSM? The process could require very specific elements of diagnostic rigor. It could require sensitivity, specificity, reliability, validity, cross-cultural sensitivity, acceptability as a precondition before even submitting it to be considered. And without sufficient levels, sufficient sample sizes, sufficient power, it should be a non-starter. And while there's no avoiding politics, checks and balances to avoid undue influence might be warranted. I'm eager to hear reactions. Thank you. Okay. So I am going to moderate a discussion now, which we invite all of you to join. And first, I'm gonna ask if anybody on the panel has comments about the presentations that they would like to make, and then we'll open it up to all of you. And if you'd come up to the mic, you can queue by the mics, and we'll take you in turn. Kathy? Yeah, I just wanna set one thing straight. I mean, we did participate with data in the diagnostic discussions, and in particular, we had data that the criteria that were proposed for DSM-5 identified only 60% of the individuals that we were seeing in our clinical trials. And maybe it shouldn't identify all of them, but 40% was too many. So this was not a political decision on the part of the DSM committee. It was based upon that troubling information. And we didn't have the opportunity to kind of work together, which is what we had been trying to do. We did propose also an alternative data set. So I just wanna make sure that's clear. It wasn't just basically telling stories. I think stories are all very, very important, but that wasn't all we were doing. Thank you. Holly, did you wanna say something? Why don't you, Roberto, could you pass the mic? In 2010, I don't believe there was an alternative diagnostic formulation or data to support it at the 2010 meeting. Hi. Please. So I know I'm probably supposed to kickstart this political discussion or whatever, but I just wanna thank all of you guys involved for pushing through the bureaucracy to make this happen. I think all of us probably have grief-stricken patients we're not entirely sure what to do with and think it probably has something to do with attachment and whatever. But my question was actually a clinical one. You mentioned your therapy that you've developed, and I was looking online to try to find out what are the training protocols for this. It seems like there's a level one through five. Maybe there's like an online on-demand course for a couple hundred dollars. Can you give a little bit of an example of how someone might be trained to treat this new disorder? Kathy? Sure. We do have, so I do direct the Center for Prolonged Grief Disorder at Columbia University School of Social Work. And centers at Columbia University are not supported by Columbia University, just for the record. So in order to have that center, we do have to support ourselves, which is the reason for the cost of the training. And this is also pretty typical. And you can find opportunities for training. We do regular two-day workshops, one-day workshops and two-day workshops. And we recently have developed an online self-guided program that you can get. This is all www.prolongedgrief.columbia.edu. You can find that on that website, as you mentioned. And we do have consultation, or the therapists who work with us provide supervision and consultation for people who want it. But there is a cost to all of this, because it's basically to support the people who work in it. I don't get anything at all from it, just to be clear. What is the difference between level one through five as far as your ability to do that? Actually, we don't certify people. That's a whole other process. What we give certificates of completion. And level one means you've completed a one-day workshop. Level two means you've completed a two-day workshop. Level three means you've completed a two-day workshop plus you've done the kind of supervision that we did in our studies, which is very intensive supervision. Other questions or comments? Yes, sir, please. I wonder if I missed the first part of the presentation when you referred to grief is because when someone loses a person. I have a number of seniors who've lost their pets and grieve for long periods of time. I'm just wondering whether this has been missed in the description. So the question is about prolonged grief after the death of a pet. I think, first of all, I don't think anyone would deny that someone could feel intense, horrible, very painful grief over the loss of a death, over the loss of a marriage, over the loss of a farm. We've defined grief as any loss of something that's very dear to you and wanting that thing back. It's longing for that thing that you feel was so self-defining and it could be that relationship. It's a lost relationship. It's the loss of a relationship with a pet. Could you qualify? I think regarding the DSM and the process, I think there's been a real sensitivity and concern for over-medicalization and over-diagnosis. And in order to be more conservative as sort of a first case, first instance example, it was decided before we start opening up Pandora's box of all types of losses, let's just try to be conservative about it and get the most obvious, typical case of someone who a psychiatrist might see and make it about the death of a loved one, a significant other person. But I think whether the criteria could apply to any other loss thing, I think the answer is yes. But I think that's not the DSM's position initially, you know, let's first see, my sense was let's just see how it goes with that. And that's been a little rocky, so I don't know if the pets are gonna be included anytime soon. Yeah, and you could analogize to the whole discussion that's going on now about PTSD for which very specific concrete precipitants have been identified in the DSM, but many people believe they're overly strict and that there is this disorder which gets called complex PTSD, which is the result of multiple traumas that accumulate over time. At the moment, that's not an official diagnosis, nor do we have a proposal to consider it, but you know, there's always a question of boundaries and how far a diagnosis should go to encompass other circumstances than those that are specified. And that's similar, I think, in both of these cases. So let's go to this mic, and then we'll go to the mic in the back. Thank you. Thank you for a very interesting presentation. You mentioned the need to measure the sensitivity and the specificity of this diagnosis, and I was curious about if you knew anything about how the gold standard has been established when you have a new concept and a new diagnosis. I imagine that it could be difficult to define this gold standard and you need something to measure it up against. And do you have any comments about how it has been solved in these studies? So that was the key challenge to be honest. You know, it's like, so what's the criterion standard? How do you know what a case is? For example, 60% are missed. For example, who decided what the case was in the first place? So the way we approached it was we had raters determine, in our studies, what we approached it was, it was a compromise between a quantitative and a qualitative evaluation. So raters had to make the diagnosis that in their opinion, you know, trained raters, that this person met a clinically significant level of the disorder based on their impression, their overall clinical impression. And we also required that there be some diagnostic threshold for the severity of the sub-threshold score. So it was a combination of clinical insight plus a minimal threshold score required. And together, those things were required to make a diagnosis of what a true case of the disorder was. Kathy? Yeah, I mean, I think that the issue of the gold standard is the key issue of any diagnosis, as the speaker noted, and it's not solved. And I honor Holly's efforts in that first study of, you know, it's a fairly small study. It was a good faith effort, but it didn't establish the gold standard. And that was really what the problem was, because we too had both data and clinical judgment, and it wasn't matching. And we wanted, you know, we reached out many times to try to discuss this, but as you can hear, I mean, Holly is very dedicated to collection of data in a way that she didn't see that we were doing, and maybe she's right. I mean, I'm not debating that, but we couldn't talk to each other, which I think is sort of too bad until we got into the meeting which Paul and his group arranged. And then, you know, and then we came up with what is, I'm sure, gonna be a first pass at this, because it's hard. I mean, but you saw the kinds of people. These were not special people. These were just kind of random people who were willing to talk to the videographers who came to Pittsburgh in 2002 or three after 9-11. That's how we got the funding for it. And to Kathy's point, one thing that's confirming and builds confidence in the diagnosis is I think most people, clinicians and people have been, researchers trained to see what it is, they know it. They know it when they see it. Probably everyone in this room could correctly, I mean, I shouldn't dismiss the criteria, but when you see it, you know it. And as one example of why I'm saying that is in our data, we had all these different criteria sets, different algorithms, different thresholds, and we were comparing them for predictability and criteria validity and all that. We had one assessment of the trained rater just at the very end of the interview saying, after all you heard, yes or no, do you think they have it? And if I'm being perfectly transparent, that rater diagnosis of yes, they had it, which correlated very highly with the criteria, but not perfectly, was the best predictor of future suicidality. So there's something that I think, and I think the ICD was really playing off of that knowledge, that clinical intuition, because their criteria are a lot more flexible and sort of, I guess, generous or weight more clinical impressions of whether this person has it or not. But I think, yes, it's true with a new diagnosis. There is no gold standard, no criterion standard against which you can evaluate a new diagnosis, but clinicians know it when they see it. So I think there's ample confidence that it is a thing. Roberta. Yeah, I just wanted to say that you are getting a flavor for the discussion over the last few years and why it took a certain process and a certain amount of time for it to reach a consensus, or at least enough of a consensus to put it in a criteria set. And I think I just wanted to frame it in terms of the difference between DSM and ICD, that because of our commitment as an organization since 1980 to a criteria-based descriptive system, we face certain difficulties in defining diagnoses that are not faced, frankly, by ICD. Because ICD, and I've worked on both systems, ICD is based on a narrative description without a strict set of counts or exact thresholds. And so as you say, there's a sense that if the clinician agrees that this meets the general description of the criteria, it's sufficient, whereas DSM has a criteria-based approach, four from column A, three from column B, that requires a certain level of analysis that took a while to reach. Yeah, ICD is freely available online. If you've never looked at it, I really encourage you to do so because it takes a very different approach, as Roberta was describing, to defining diagnoses more like DSM-2 with a couple of sentences that provide a prototype of what a person with this diagnosis will look like and ask clinicians to apply it. Yes, please. I've had some, some personal experience. Could you come closer to the mic, please? Loss, and there are two organizations that have been out there for over 40 years, SHARE, which is miscarriage, stillbirth, and neonatal death, and there's Compassionate Friends, where people lose children from any type of either trauma, cancer, it doesn't matter. And I wondered if you guys had, you know, collaborated with these groups that have been around for over 40 years. And I've seen people in those groups, you could see it, like she said, maladaptive grief. Not everyone had it, but there were definitely people notable, and I've seen it as a clinician in my practice, too. I mean, that's with children, of course, of any ages or even miscarriage. But I wondered if there was collaboration because these two organizations have been around and very helpful to people, but, you know, they're not sanctioned clinicians to treat the grief when it becomes maladaptive. I don't know if Kathy's had this experience, but I suspect you have. So there are organizations like ADEC and other organizations that really are not pleased with coming up with a diagnosis and medicalizing and stigmatizing and labeling people's grief as pathological. So I feel like they feel offended that we're somehow not appreciating what their reaction is and that we're coming from it from a very over-medicalized perspective. So my personal experience has been a lot of hostility from ADEC and some, not compassionate friends, but ADEC has not been receptive to the new diagnosis. And I think that some of that might be some professional turf issues because of treatment. And I'm not a psychiatrist, but I would think that maybe that they feel threatened by psychiatry. If you have a psychiatric label, then you would require a psychiatric diagnosis and a psychiatrist to treat it. So I don't know if that's a good thing. I don't wanna, I'm right out on that limp there, but my experience with those organizations, it would make perfect sense to work with them because we're not over-pathologizing. A lot of the people who are struggling the most are mothers of kids who've died from a car accident and they contact me denying the diagnosis, but clearly it's struggling and wanting help intently and nothing's helping them. So it's just a political challenge to deal with it, as I see it. Kathy, did you wanna say something? I mean, I basically agree. I mean, we have worked with, I met with the president of Compassionate Friends at one point, but basically they are not interested in working with us. And I think some of the organizations around grief and loss also came out of a very anti-medical stance from the beginning, really having to do with end of life. We all know that the medicine in general hasn't been great with end of life issues for a long time. And I mean, we're better now, I guess, but, and also we haven't acknowledged grief until recently. We haven't acknowledged grief until recently. So we have time, we just have a couple of minutes. One end comment. So SHARE was run by a clinician, a master's prepared nurse from Northwestern, and actually collaborated with my cousin who was a pediatric geneticist. So that was under the medical umbrella, just to make a comment. But thank you for this diagnosis because I think it's been a long time coming. So we've got time for a fast question and fast responses. So I had been doing work with people with HIV and bereavement back from the early nineties and did an NIMH supported study of a bereavement support group intervention where our criteria was significant distress related to the loss within the past six months. And we found a cascade of positive effects in terms of grief and distress, stress hormones, immune function, viral load, and clinical healthcare utilization there. And it was very frustrating because I'm also a clinician to be in the setting at the same time I was seeing patients who were presenting more with characteristics of prolonged grief disorder where you couldn't make the diagnosis. That you didn't have MDD, you didn't have PTSD, and yet you had a patient who needed care and you didn't feel you could appropriately designate them for the care. And in fact, that represented a barrier. And it seemed to me that it represented a barrier for many more years than perhaps was necessary according to the data. So let me conclude with some reflections here on what we learned as a DSM steering committee from this process. We learned that it's extremely difficult to include a new diagnosis in the DSM without consensus among the major people in the field who are providing the data on that disorder. We learned, second, that one way to try to achieve some sort of agreement, if not absolute consensus, is to bring people together, physically together, in the same room, and have them talk to each other. And we're doing something similar now with personality disorders, and we'll see how that all plays out. Another fraught and historically laden diagnostic category. And I think the third thing we learned from this is that we're not starting new in 2023 or 2018 when the proposal first came in, that there are histories here, and strong feelings attached to those histories that had nothing to do with us and well preceded us, but that nonetheless, as psychiatrists, you're all aware, can't help but play a role in how things play out. So these are complex situations, and we are very pleased with the way prolonged grief disorder turned out, and that we were able to achieve a data-based agreement on how to approach that condition, in some sense, it's a model for us for future efforts. So let me thank our presenters, Kathy Scheer, Roberto Luis-Fernandez, and Holly Bergersen, and let me thank all of you for joining us this afternoon. Thank you.

Prolonged Grief Disorder

DSM-5-TR

Paul Appelbaum

DSM Steering Committee

APA

diagnostic inclusion

clinical experts

empirical data

major depressive disorder

PTSD

Kathy Shear

Holly Prigerson

maladaptive grief