Dr Mattson- PTSD vs MDD

so real quick before I before I get too far into the presentation a couple of things I do want to clear up a few of the abbreviations just make sure everyone’s on the same page so post-traumatic stress disorder is PTSD major depressive disorder is MD D and the Minnesota multiphasic personality inventory second edition restructured form is the MFP I to RF I shortened those you should have seen the title slide before I did so that’s that’s why those are shortened I just want to make sure that that we’re clear on any of the abbreviation in addition with the handouts I have given a scale listing just for anyone who is unfamiliar with the restructured form of the MMPI I should mention that this list is not exhaustive there are validity scales and interest scales that have been left off primarily because they’re they’re not pertinent to the information specifically that we’re going to be talking about today and I would ask you to consult the RF manual if if you have further questions about those one other housekeeping piece before we get started as we move through here I’m going to present a little bit of previous data with the MMPI 2 but the crux of the what I’m going to be presenting on is on the restructured form so you’ll you may hear me refer to the the MMPI to RF simply as the RF and when I refer to the MMPI 2 I’ll say the MMPI 2 so if there’s a question if you’re concerned don’t hesitate to ask so I should mention that I’ve left some time at the end of the presentation for questions but if there’s something that is more pressing on your mind as we move through don’t hesitate to ask so so I’ll go ahead and get started so some history kind of about the the creation of the origination of this product project the DSM as a diagnostic tool has undergone some pretty sharp criticism primarily and in large part due to significant Gnostic comorbidity the problem has largely been attributed to a classification system and in fact the DSM itself is very open about the idea that it uses a rational clustering of symptoms and not an empirical one I believe the exact phrase is a phenotypic presentation and so the this comorbidity is problematic and we particularly see it present in the anxious and depressive disorders and just some information there about some research that’s been done on on the comorbidity of these so particularly troubling is that first one on there where the current and lifetime comorbidity between depressive and anxious disorders has ranged from 57% for current and 81% for lifetime comorbidity and and for a group of clinicians or a as either even as a profession in general having comorbidity between you know supposedly distinct diagnostic classifications is quite problematic so two of the diagnostic classes that are most commonly hit with highest comorbidity rates are major depression and post-traumatic stress this is largely due to symptom overlap which makes the differentiation between these more difficult we can look at things like problem sleeping problems thinking clearly and and those are not specific to a certain disorder the problem as a profession is that diagnostic differentiation is is really needed because we know that anxious disorders and depressive disorders are for our purposes major depression and PTSD have very different treatment courses treatment prognosis outcome things like that and I think it is in our best interest and certainly in our clients best interest to try and come up with the best way that we can to measure and differentiate these disorders so acknowledging that we need better ways of examining this information there have been three kind of major pushes in doing this the first is a revision of the PTSD symptom structure the second one is an empirical restructuring of the internal izing disorders and the third one is an elaborated model of temperament or what is kind of come to be known since then is an elaborated hierarchical model of effect I want to touch on these three things but I’m going to do so kind of briefly so as to not stray too far away from the point of the presentation but hopefully

in showing these it’ll it’ll make a little bit more sense as to why differentiation is needed and why it has become so problematic the first thing that I want to look at is some revised symptoms structure of the PTSD as a diagnosis and the DSM for the text revision classifies PTSD as an anxiety disorder and it has three rationally derived clusters where we talked about that as being one of the criticisms of the DSM and you can see those listed Leonard Sims and several of his colleagues came out with a four factor model and they essentially kept the DSM clusters intact except you’ll notice that avoidance is no longer numbing and avoidance it’s just avoidance and they added a cluster for nonspecific dysphoria moving on to an empirical restructuring of the internalizing disorders one of the first to really look at this was Bob Krueger and then later David Watson and this kind of obviously moves from left to the right and and we’re gonna be kind of following it up once again this model and this diagram is just adapted dave watson in his 2005 paper after internalizing he actually has distress fear and bipolar as a separate cluster but obviously for space purposes like I couldn’t include that so for our purposes we will just be looking at the distress and then fear to show how they’ve kind of been separated you can see there that the models include major and PTSD under the distress disorders this makes sense based on some of those empirical criteria and even some of the symptoms structure of major depression in PTSD interesting to note is also that generalized anxiety also falls under this distress domain as well and then is our temperament models so in particular research on the anxious and depressive disorders has looked at a two factor model negative and positive effect or negative and positive activation or negative emotionality and positive emotionality but Martin’s cell bomb in 2008 kind of took some of the earlier work that had been done and looked at an elaborated model and what they found was that they kept the negative and positive activation or effect and they they added a third overarching dimension that was later classified as demoralization initially it was known as happiness and unhappiness and I would love to have a discussion about this with anybody that would like to I think this is very interesting stuff but as I said for our purposes we’ll move on so demoralisation is broadly associated with the distress disorders so this would include major depression dysthymia generalized anxiety and PTSD whereas negative effect is a general component of the fear disorders so that kind of lower branch on the model that I just showed and then positive effect is a unique marker of major depression and social phobia specifically and what I’ve got and I’m going to see if this works if you guys don’t mind hanging out with me for just a second I want to try and kind of show you exactly what their model looked like although now having clicked on it I may have just become the victim of a technology error so no no this is on my flash drive but we’ll find out very good question so a lot of people had assumed in the past that positive and negative effect were just opposite sides of the same dimension and we actually view them as two separate axes so if you view positive effect is like a y-axis and negative affectivity as kind of an x-axis the demoralisation factor would want run from the top left to the bottom right and so when we talk about positive effect being a unique marker of depression we’re looking at a low positive effect so if you think about for those of you familiar with the MMPI to RF you know restructure clinical scale to is low positive emotion that’s measuring that low positive effect so and that’s important to think about when we look at where that overarching factor

of demoralization lays it lays kind of bisecting from the top left to the bottom right so it lay it lies in between low positive effect and high negative effect and so when we think of a combination of those two that’s where that demoralisation facet comes in and I’m afraid this is not gonna work so good just roll with it right that’s what we do oh goodness well I was going to try and show you the exact model that shows how demoralisation is associated with all these but now I mean I’ll be able to do all right well that’s right I don’t know what I should do that’s right an escape does not work either so everyone should be really familiar with this slide by the time we get done in here all right so what does this all mean right I’ll just go ahead and well we will press on so yes I found it so I actually do want to show this you guys it does have a point not to but we’ll take it so I don’t know how is that somewhat clear for you guys to see so where we’re at is if we look kind of on the right-hand side towards the top we see distress and we’ve got depression generalized anxiety and PTSD for this particular paper they didn’t study dysthymia so that’s why it’s not included there and then at the bottom we see for the fear we’ve got specific phobia agoraphobia or sorry social phobia agoraphobia and specific phobia and so on the Left we have the dysfunctional negative emotion which is that high negative effect which is associated with the fear disorders we see demoralisation which is associated with the distress disorders and then we see low positive emotion or that low positive effect which is a unique marker of depression in social phobia then you can see kind of the inter correlations how these are all tied together but I wanted to kind of show you that so you didn’t just have to read it and try and picture it in your mind that that this is the model that was proposed so now let’s hope I can go back and everything will be perfect all right we’re that we made it okay so what does this all mean I realized I’ve had you know cover 20 plus years worth of research in a few slides so essentially the reason why all of this is important is we use a diagnostic tool that assumes that for the most part anxiety and depression are separate constructs and yet the empirical research that has been done shows that that may not be the case and that maybe there’s other things that are associated with the variance between the anxious and depressive disorders or maybe that a better classification system is needed so we see things like the addition of a general dysphoria component to PTSD we see a empirical restructuring of the internal izing disorders that includes PTSD and generalized anxiety along with major depression and we also see that in terms of temperamental factors that both higher factors and lower factors so this demoralization and positive and negative effect we see that both higher and lower markers are associated in differentiating these two conditions so the reason why it’s important is we got to figure out a way to measure or to assess the difference between anxiety depression better so I want to first cover some MMPI to research as you know the MMPI 2 is a precursor to the restructured form so typically what we’ve seen in the past is that major

depression is associated with clinical scale 2 which is depression and the content scale dep which is depression unfortunately findings regarding PTSD have been more equivocal and and you know in particular we see that PTSD profiles are often quite pervasive they often elevate a majority of the measured scales in fact pretty much any study you pick up looking at the MMPI – and PTSD one of the first statements in the discussion will include some context of the PTSD group elevated a majority of the scales we also see in particular clinical scales – 7 & 8 are the most commonly elevated but we also see scales 1 in 6 have been associated with trauma in combat veterans peacekeepers and female victims of domestic violence this can also be due in part to some work that Mark Miller has done looking at differences between subtypes of PTSD so he’s postulated the internalizing and an external izing subtype of PTSD that present in different ways so in looking at the MMPI 2 we see that clinical scale code types which is traditionally kind of the way that interpretation on the MMPI 2 has been taught that a 7 8 or 8 7 profile is consistent with diagnosis of PTSD anxiety depression and that code types of to seven or seven to are broadly associated with general classes of depressive and anxious disorders so which one is which I know right so also looking back at scale eight elevations with PTSD we see that in PTSD samples scale eight has been associated with some of the some of the markers related to depression including social alienation problems with thinking including which may include hallucinations as well we generally don’t see it be associated with what we would look for in true markers of psychosis so you know if you remember scale eight on the MMPI two is labeled as the schizophrenia scale we actually see that scale elevate with PTSD not because of schizophrenia but because of these other factors that are mentioned here so in the most focused examination of differentiating major depression and PTSD Greenblatt and Davis they actually looked at differentiating PTSD major depression and schizophrenia using the MMPI two for our purposes we’re just going to be referencing their findings where they differentiate a major depression and PTSD so what they found was that both groups elevated clinical scales to seven and eight which we kind of expected and then the PTSD group also had additional clinical elevations on clinical scale six and the bizarre mint Asian content scale clinical scale six is paranoia so some post hoc findings that were done after their scales of interests were reviewed is they found that clinical scale 9 which is hypomania and content scale anger and social discomfort were also elevated in the PTSD group so i want to kind of pause here for a second and think about some of these MMPI two findings and i think it’s pretty clear that in regards to PTSD i don’t know that we have a really good idea about what constitutes a PTSD profile and and part of the reason for showing you the last several slides with all of these different mm pi2 scales that are elevated and all these problems is that was one of the issues behind restructuring the MMPI to to come up with the restructured form is we need a better way of measuring some of these markers it’s confusing to me as well and I’ve been you know I’ve been working on this project for a long time it’s confusing even to me to look at the MMPI to findings in relation to PTSD and really wrap my head around them so recognizing that excessive inter correlations between the scales on the MMPI – that there’s a lot of item overlapped and construct overlap between the different scales we needed a better way and there we go so for our purposes the RF represents an excellent opportunity as the RF has both high and low markers of temperament that we talked about as well as a hierarchical structure that fits well with current models of psychopathology the RF is also anchored by the RC scales instead of the

traditional clinical scales and the RC scales in pretty much every empirical examination have outperformed the clinical scales in terms of reliability and validity now that’s of course not 100% the case but but for the most part has been consistent so to date there has actually been relatively little information that has been collected on the specific problem scales or the revised version of the SCI v scales for the RF there has been some research done on the restructured clinical scales and traditionally the RC D RC 2 and RC 7 have been associated with PTSD and major depression and once again the RC scales and research have been shown to better predict PTSD and other correlated conditions including major depression so where does that leave us so this current study what we wanted to do was examine the ability of the RF to differentiate PTSD and major depression so for differentiation purposes we have we have two requirements the first is that the mean scores between the two groups had to be a significant t-test finding so we had to have some significance between them and we also wanted a mean T score difference of greater than two equal to five points or a medium effect size so for our hypotheses to kind of orient you to this a little bit try and presented in in the easiest way possible on the far left side here these are scales that we believe that both groups will elevate and we mean elevation in the traditional MMPI sense of T scores 65 or above in the middle we have these are the scales that we believe our groups will differentiate on specifically that the PTSD group will score higher with and on the far side the scales that we believe that the depressive group will be differentiated from the PTSD group with you’ll notice that all of the scales that we believe that both groups will elevate are all within the same hierarchy and you can see that on the scale listing that I gave you on the right hand side I listed the Association the hierarchical Association for the second column they’re kind of the second hypothesis and all of these were based on previous research that had been done on the MMPI – we kind of carried that idea forward and we want to test whether or not these scales will also differentiate the PTSD group from the major depression group and then on the far right we actually don’t really expect that the that the major depression group will be that differentiated from the PTA sorry we didn’t view that the depressive group would be differentiated from the PTSD group by higher scores on much except you’ll notice that our c2 and our c2 is a little bit of an oddity in terms of the clinical or the restructured clinical scales and that it does not have any associated sub scales and you’ll you can see that on the scale listings under that AI D if you follow the IEEE Eid hierarchy down you’ll see that there’s none that are associated with our C – so our participants were psychiatric inpatients at a VA Medical Center or a large urban County Medical Center these are both in Minneapolis we had that just shy of 3,000 participants at the beginning and for those interested the final two slides of the presentation are a reference list in the the RBC 2003 article actually describes that sample in full detail the patients were given the MMPI – as part of the intake procedure and the RF scores were extracted from that there’s actually been research that shows that taking mmpi to RF scores from an mmpi to administration yields similar results as if you were just to give someone the RF to begin with so no concern about you know pulling that information over from the NPI – we had three separate exclusionary criteria that we used to kind of will down to our final group the first was since we were primarily interested in individuals with PTSD and major depression we decided we only wanted those people who had a primary diagnosis of one of those two things so if you didn’t have a primary diagnosis we excluded you what we did next was a bit of back-and-forth in the discussion room about how we were going to handle this what we decided on was be if you

had a primary diagnosis of major depression if you had a secondary or tertiary diagnosis of any anxiety disorder we excluded you if you had a primary diagnosis of PTSD and any unipolar mood disorder we excluded you as well we tried to leave some diagnostic comorbidity because that’s what the real world is but we also wanted to try and emphasize as what these differences between these two diagnoses may be so we’ll get into it a little bit more in the limitations but that that is going to be kind of a problem with generalizing these results just so you’re aware and then the last thing we did was using kind of standard validity criteria as listed in the RF manual we excluded those people who invalidated the MMPI to RF protocol we ended up with a final sample of 257 people and you can see the descriptive information there we had two hundred and sixteen in our major depression group and we had 41 in the PTSD group there were some differences as the major depression group tended to be younger and had a higher proportion of female but those effect sizes were small in fact one of them was you know borderline not even small anyway there were no differences in race or education and if anyone’s interested afterwards I can talk to you about some of the you know specific ranges and percentages within that sample but for the sake of time in this moment I will not so the MMPI to RF is that like I said a hierarchical structure what that means is it kind of starts few scales at the top and kind of branches out as it goes down so at the top we have higher-order scales so on your sheet you can see that these are eid THD and bxd and you can see all that information the higher-order scales are broadly measures of these these huge domains of psychopathology the RFC scales are considered mid-level constructs and they essentially measure the core concepts from the clinical scales from the MMPI to as we branch down further we get to the specific problem scales that are narrow band scales that measure kind of very facet level information and those those low-level temperament markers that we are talking about the revised size 5 scales are not part of the hierarchy of the MMPI to RF but they are associated with with various branches of it so to speak and the sci-fi scouts have been shown to provide some information on personality psychopathology so for our results we first calculated means and standard deviations we then conducted our t-test to compare these means and then we calculated our Cohen’s D values the idea was that an effect size of Cohen’s D value is a measure of effect that we could more accurately quantify the difference between the two group means scores so you know if we had the depressive group on demoralization and the PTSD group on demoralization those scores might be different and the effect size is going to help us know how significant that difference is okay so into the tables I want to kind of orient you to the table so on the left hand side I’ve kind of outlined and and riri laid out some of our diagnostic or differentiation criteria so along the bottom here and you can see one here like on th D if there’s an asterisk it means that the t-test finding was significant and then if you see a red arrow it means that it met the differentiation criteria for the PTSD group scoring higher so that differentiation criteria once again was significant t-test finding and a medium effect size so 0.5 or above for the Cohen’s D value and then if you see a blue arrow that is a group that met the differentiation criteria for the major depressive group being higher so from here we can see that behavioral externalizing dysfunction bxd and then our c7 our c8 and our c9 all met our criteria for the PTSD group scoring higher th sorry THD did not but if you see up here or look on your handouts 0.49 for the effect size so it was oh it was so close so close to also meeting that criteria as well so an additional thing so along the top you see it says somatic cognitive and then internalizing those are the kind of different classes of the

specific problem scales so we see on here that suicidal death ideation or the Sui scale differentiated the major depressive group from the PTSD group this was the only scale on the profile that did that and one thing I do want to want to say is you see a negative value for the Cohen’s D that just simply means that it was kind of in the reversed order that we had them entered so negative 0.5 is essentially a positive 0.5 if we were looking with the major depressive group being higher we can also see that anxiety and anger proneness the a X Y and ANP also met our differentiation criteria aggression the AGG and activation AC T met our differentiation criteria and like several other scales social avoidance and shyness significant T test finding but once again they didn’t meet both criteria so that’s why there’s an asterisk but no arrows and then looking at the size five scales we see that aggressiveness revise the aggr met our our differentiation criteria whereas negative emotionality did not so in the discussion just to remind everyone what we did is we sought to examine the ability of the MMPI to RF to differentiate post-traumatic stress from major depression consistent with the research that had been done using the MMPI to both the PTSD and the major depressive groups elevated a number of scales but when compared to major depression like we said before the PTSD group had more elevations over more scales in the profile and those elevations tended to be more severe alright so what I’ve done here is this is the this is the same table that I had showed before and what I’ve done is I put an asterisk by our hypotheses that were supported I have grayed out the hypotheses that were not supported so first off looking at the first column we see that rc7 on the RC scales did not meet our hypothesis it was elevated in the PTSD group but it was not elevated in the major depression group helplessness hopelessness hlp and stress and worry were not supported and those skills actually didn’t elevate in either group and then none of the sigh five scales for that particular hypothesis were supported moving to the middle group we can see that and I want to point out the THD the thought dysfunction and the psychoticism revised all the way down here at the size five scales those were not supported and it could just be simply that those scales are so broad in nature that while they may have had some items that were consistent with what we’re looking for just the broad nature of them kept that elevation from occurring we also see that in general it looks like the PTSD group met a number of our hypotheses and essentially that even though both groups elevated a number of scales that PTSD is is fairly fairly well differentiated from depression on the basis of increased negative emotionality some dissociation specifically associated with rc8 and then externalized behavior things like acting out hyper vigilance things like that so what does this all mean well first off in terms of those scale elevations neither major distress or major depressive disorder or post-traumatic distress neither one of those two groups elevated a scale outside of that a ID hierarchy this is good right we we shouldn’t expect these disorders to get a clinical elevation outside of the hierarchy that they’re associated with of course this is except for a few somatic scales when we can look at that and actually make some pretty easy conclusions from that we see that PTSD is best differentiated from depression due to negative emotionality externalized behavior and dissociation whereas major depression is only differentiated from PTSD and our study by a specific facet of demoralisation or generalized dysphoria we also see that a number of the scale elevations with PTSD highlight the heterogeneous nature of

this and Dave Watson in his paper actually made the comment that PTSD appears to be a weak marker of the distressed disorders and and no doubt you know he there are several criteria for PTSD that fall clearly within that those fear disorders the phobias and things like that so PTSD while it empirically seems to belong with the distress disorders I think we see from the number of elevations that this group produced on in this study that maybe it’s a weak marker at best so in regards to findings on the suicidal deaf ideation scale there’s a study back in 2009 where they actually found that women who had been diagnosed with major depression were 18 and a half times as likely to report some past suicidal thoughts and three and a half times as likely to have a previous suicide attempt when compared with PTSD we had hypothesized that major depression will be differentiated from PTSD by rc2 and I mentioned earlier that it was a bit of an oddity the reason for this is that when we look at some of the research that’s been done on our c2 being a measure of low positive emotion it’s actually a strong enough scale by itself and that’s one of the reasons that doesn’t have some of those sub scales that’s why we hypothesize that it would be a unique marker of depression like some of the timber models had suggested it was not supported and and that’s not entirely unsurprising because there has been some research that showed that that demoralization may be a better measure of this and also we know that the internalizing subtype of PTSD is more associated with a lot of characteristics of major depression so what we may have been seeing is maybe our group was more heavily weighted towards the internalizing PTSD and that increase in some of those endorsements kind of removed rc2 from being a differentiator in terms of limitations the big one that I said we’d come back to we will get to in just a second what I do want to mention first is in the diagnostic assigning z’ that were done at the inpatient or the VA Medical Center those were done by clinicians during the intake process they did not have access to the MMPI to data but they also did not have access to any sort of diagnostic screening instrument that DAV’s or the skit or anything like that so there is some question about the reliability of their diagnosis particularly in response to some of the criticisms that have been lobbed at the dsm in addition our PTSD sample only had 41 individuals and there were some group differences I’m not too concerned with the 41 people in the PTSD group this may actually just simply be a reflection of kind of normal base rates within the population you know we know that major depressive disorder is one of the most common mental health diagnoses and it may simply be that the difference between 2 16 and 41 may just be a reflection of the base rates the last one and the one that I think is is most problematic but maybe gives us a glimmer of hope towards the future is the fact that we created relatively clean groups and for anyone interested you can go back and look at bob Krueger’s 1999 article that’s in your reference list and he actually has a really interesting discussion about what it means to create clean groups and the fact that really we don’t have any evidence that there such thing as psychopathology existing in a vacuum we created a little bit of a vacuum to create our groups and we acknowledge that as a limitation but in terms of kind of hopefully looking towards the future maybe we can create better groups moving forward maybe there’s maybe there’s a better way to measure this and now that we know some of the things to look for maybe we can use that so in future directions we want to create more real world groupings maybe what we do is we keep the people with the primary diagnosis and just leave everybody else and see if that has an impact we also want to replicate these findings you know this sample was an inpatient sample how is it going to look without patients how’s it going to look with military how’s it going to look with sexual assault survivors PTSD is studied in a number of different areas and just because we found some interesting results in ours doesn’t mean that it will necessarily move to other areas and then also if our findings remain consistent as we study this more we’d like to work on developing some classification accuracies or working on some algorithms to maybe come up with a way to better differentiate you know

this profile that i’m looking at is it depression or is it PTSD and if we have consistent findings we may be able to be to find a way to better emphasize that so that is kind of the end of the study are there any questions and I did put my contact information there on the bottom in case anyone has a question that they think of later or in a moment of Epiphany as you’re brushing your teeth if you have a question you can email me to the I absolutely agree and you know they went through several processes to go from the two to the RF you know identifying the core constructs and seed scales and and all that good stuff but that was kind of a question that has been raised before by other people you know Roger green is one of them and he’s raised the idea of if the scales weren’t that great to begin with all you’ve done is cut down the scales what makes him so good now and I would agree that on the surface I believe that’s actually a very valid criticism and maybe there’s more maybe we need to go back to the creation in the 1940s and really start from the ground up but you know consistently in most avenues of research the RC scales have shown to do better at predicting what we think that they’ll predict and they’ve shown you know improve construct validity and discriminative elyda t so you’re right it’s certainly not perfect and I don’t think there’s an expectation that it would be maybe this is just getting us closer closer to home so maybe it’s a maybe it’s a new way point on the way to the final goal so to speak so yeah yeah yeah you know I think five arrived with its own fanfare and I don’t know that any of it was was really all that good so I think that’s the concern and I think and actually you know Dave Watson’s the 2005 article that I referenced where he restructures the internalizing disorders that actual title the paper is moving towards an understanding for dsm-5 and he lays out some very rational arguments for why you know we need to relook at this and then yes m5 came out and basically said no so it’s interesting you know that you know much of the classification differences between the anxious and depressive disorders remained unchanged and yet we’ve got 20 years of research that says there’s a problem here you know where there’s smoke there’s fire and we’ve got a house filled of smoke and no one’s running for the exit so I think that’s a problem so but I mean I think that’s the reason why the fanfare was so negative for the dsm-5 why it was not as well received so yeah I would have hated to try and do this with dsm-5 criteria any other questions