Is There a Future for Precision Mental Health Care for Anxiety Disorders in Children?

Transcript

• 00:02Hello, everybody. Hi. Good afternoon.
• 00:05Hi, everybody. I'm I'm I'm
• 00:07I'm Wendy Silverman.
• 00:09I'm, I directed Ellie Leibovitz,
• 00:11the anxiety program
• 00:13here at the Charles Studies
• 00:14Center, and I'm
• 00:16so thrilled that a world
• 00:19famous child anxiety researcher from
• 00:22Australia
• 00:23is here.
• 00:24And, her name is the
• 00:25professor Jenny Hudson. I know
• 00:27Jenny for a long time
• 00:28because of her excellent work,
• 00:31And,
• 00:32she's
• 00:33known all over the world,
• 00:34and she's especially
• 00:36a a rock star all
• 00:37over, but a super rock
• 00:39star in Australia,
• 00:41where she is now I
• 00:41have to take my glasses.
• 00:43Where she is a fellow
• 00:45of the
• 00:45Australian Academy of Social Sciences
• 00:48and the what would be
• 00:49the equivalent of NIMH, the
• 00:51National Health and Medical Research
• 00:53Council.
• 00:54And she is also very
• 00:56heavily involved in Australian policy
• 01:00with with regard to child
• 01:01mental health. She chairs in
• 01:02Australia the government policy,
• 01:05committee of child mental health
• 01:07research plan.
• 01:09And she's also now participating
• 01:11in the Young Minds Matter,
• 01:13our future, which is also
• 01:15a Australian government department of
• 01:17health,
• 01:18initiative.
• 01:19So besides her incredible
• 01:21work in the public
• 01:23policy and sphere, she as
• 01:25I said, she's
• 01:26incredibly renowned in her work
• 01:28in child anxiety
• 01:29and developing,
• 01:31innovative interventions that involve parents
• 01:34and children and
• 01:35is now moving in the
• 01:37direction of,
• 01:39digital mental health. So I
• 01:41could go on and on,
• 01:42but and I I also
• 01:43just say she's
• 01:44fabulous. And if I may
• 01:46reveal a secret, she had
• 01:47her first pastrami sandwich in
• 01:49New York City
• 01:50last night,
• 01:52at Katz's Delicatessen
• 01:54in the Lower East Side.
• 01:55So,
• 01:57for those of you who
• 01:58may not know that, I'll
• 01:59take you also to, Katz's
• 02:01Delicatessen
• 02:02in the Lower East Side
• 02:02of Manhattan. Okay. So and
• 02:05and she's here to still
• 02:07you know, eating a sandwich
• 02:08like this is no easy
• 02:09task because she's still able
• 02:11to be with us now,
• 02:12after this unbelievable
• 02:14pastrami sandwich.
• 02:16I'm not gonna talk about
• 02:17the and the other stuff.
• 02:18She could tell you about
• 02:19that.
• 02:26I can highly recommend the
• 02:27Wendy Silverman tour of the
• 02:29Lower East Side. It is
• 02:31excellent.
• 02:32We had a fun time,
• 02:33over the weekend, so thank
• 02:35you.
• 02:36And a big thank you,
• 02:37Wendy, for having me here
• 02:38today. It's a pleasure to
• 02:40be able to present to,
• 02:42this incredible group of people
• 02:44here at Yale
• 02:45Child Study Centre. I've, admired
• 02:48Wendy's work for a very
• 02:49long time as well as,
• 02:51Ellie's
• 02:52and, Alan Kasdan as well.
• 02:54So it's very, it's a
• 02:55privilege for me to be
• 02:56here and share with you
• 02:57the work that I have
• 02:59been doing.
• 03:01So I'm kind of thinking
• 03:02about what I wanted to
• 03:03present today,
• 03:06and thinking about kind of
• 03:06the work that I've been
• 03:07doing. This one question that's
• 03:09been on my mind that
• 03:10I thought I would
• 03:11spend the time kind of
• 03:12going through the data
• 03:14and also,
• 03:16I suppose, exploring this idea
• 03:18whether or not there is
• 03:19a future for precision mental
• 03:21health
• 03:22in anxiety disorders in children.
• 03:25We've had there's been a
• 03:27lot of talk around personalized
• 03:29care and also precision
• 03:31mental health. So I use
• 03:33this kind of slightly different
• 03:34term. So personalisation,
• 03:35I think, can be
• 03:37any adaptation really that a
• 03:39person might make to therapy.
• 03:42You know, maybe you're tailoring
• 03:44something to a a specific
• 03:46child because of their presentation
• 03:48or maybe their interests.
• 03:50But
• 03:51that's kind of personalising
• 03:53it's not really driven by
• 03:55data.
• 03:56So when I say precision
• 03:57mental health or precision mental
• 03:59health care, precision care, it's
• 04:01about
• 04:02personalization that's driven by data,
• 04:05which,
• 04:06they're they're both quite different
• 04:07things.
• 04:08So
• 04:10I do think that that
• 04:11kind of personalization,
• 04:13there's always gonna be a
• 04:14future for that because that's
• 04:15how clinicians work. As a
• 04:16clinician, you always
• 04:18tailor your
• 04:19approach and, you know, you're
• 04:21not gonna approach a seven
• 04:23year old in the same
• 04:24way that you would approach
• 04:25a sixteen year old boy,
• 04:26seven year old girl, for
• 04:28instance.
• 04:28And you'll you'll have a
• 04:30different style in the way
• 04:31that you interact with the
• 04:32child and with the family.
• 04:34But that's based on kind
• 04:35of your clinical judgment. It's
• 04:37not necessarily based on getting
• 04:39better outcomes.
• 04:40So for me, the idea
• 04:41of precision mental health care
• 04:43is is more about personalising
• 04:46that's going to improve treatment
• 04:48outcomes.
• 04:49So what is it that
• 04:50you will do,
• 04:52that will be better for
• 04:53the seven year old as
• 04:54opposed to the sixteen year
• 04:55old that's based on data
• 04:57that is gonna improve outcomes?
• 05:00So the the slightly different
• 05:01things, but I just wanna
• 05:02make it clear they were
• 05:03the different definitions that I
• 05:05use. And it's the precision
• 05:07medicine,
• 05:08the precision mental health care
• 05:09that I'm not a hundred
• 05:10percent sure about. I'm not
• 05:12sure whether there is a
• 05:13future for it. So I'm
• 05:14gonna take you through
• 05:17as I would never have
• 05:18said that maybe five or
• 05:19ten years ago, as I
• 05:21thought, for sure, this is
• 05:22where where where we were
• 05:24headed. But I've become a
• 05:25little bit more skeptical,
• 05:28and I'm not yet convinced.
• 05:29I'm still hopeful, though. I'm
• 05:31still very hopeful that there
• 05:32there is a future, but
• 05:33I wanna take you through
• 05:34the data,
• 05:35that I've been working on
• 05:36and others around the world
• 05:37have been working on as
• 05:38well. I've even slipped in
• 05:39a slide of Wendy's work.
• 05:44Yeah. That's right.
• 05:46Paying you back.
• 05:48But, yeah, just to see
• 05:49where where we're at. What
• 05:50do we know in terms
• 05:52of data driven decisions
• 05:53about what is gonna improve
• 05:55outcomes for children and young
• 05:57people?
• 05:59Alright. And then you can
• 06:00make your own decisions about
• 06:01it as well. I'd be
• 06:02keen to hear from everyone
• 06:04here what you what you
• 06:05think about it, whether or
• 06:06not there is a is
• 06:07a future.
• 06:08Am I too jaded already,
• 06:10or,
• 06:12am I not looking at,
• 06:13yeah,
• 06:14the possibilities?
• 06:16Alright. So before I get
• 06:17started, though, just wanted to
• 06:19make sure that, you're aware
• 06:20that I have no financial
• 06:21interest to disclose that are
• 06:22relevant to the what I'm
• 06:24talking about today.
• 06:29You already understand about anxiety
• 06:31disorders and how common,
• 06:34impairing they are because I'm
• 06:35sure you hear Wendy talk
• 06:36about it all the time
• 06:37and, Ellie,
• 06:39but I just wanted to
• 06:41take you before I started
• 06:42talking about the personalization a
• 06:43little bit more about,
• 06:46the fact that anxiety disorders
• 06:48are still overlooked. And people
• 06:50might not necessarily
• 06:51think that anymore,
• 06:52but
• 06:54when you see what's happening
• 06:55in the community and particularly
• 06:57within children
• 06:58and in Australia, they are
• 07:00still very much overlooked,
• 07:02in terms of being considered,
• 07:04a mental disorder or something
• 07:06that's impairing and something that
• 07:07we need to take seriously.
• 07:10And I think in Australia,
• 07:12this comes about because of
• 07:14a common belief
• 07:15that mental disorders don't start
• 07:17until adolescence. I don't know
• 07:19if that's something that
• 07:21is is a common. You're
• 07:22here in the child study
• 07:23center, so you're kinda focusing
• 07:25on kids.
• 07:26But in Australia, there is
• 07:27a lot of focus on
• 07:29adolescents, and that is where
• 07:31a lot of interventions and
• 07:32a lot of government money
• 07:33has been spent on improving
• 07:35interventions for adolescents.
• 07:37With the assumption that children
• 07:39don't experience
• 07:41disorders like anxiety or depression,
• 07:42that it doesn't start until,
• 07:44till adolescence.
• 07:46And this really comes about
• 07:47from,
• 07:49Ron Kessler's
• 07:50original,
• 07:52age of onset paper,
• 07:54which says the median age
• 07:56of onset. So this is
• 07:57kind of national prevalence data
• 07:59from, the US and around
• 08:00the world,
• 08:02looking at you know, asking
• 08:04people asking adults
• 08:06when at what age did
• 08:07they first start to experience
• 08:08anxiety symptoms, or what age
• 08:10did they first start experiencing
• 08:11depressive symptoms.
• 08:13And when you take the
• 08:14median age of onset of
• 08:15all of the disorders, so
• 08:16it's kind of all the
• 08:17high prevalence disorders,
• 08:19you get this median age
• 08:20of onset of fourteen.
• 08:22But people misinterpret
• 08:24what a median age of
• 08:25onset is.
• 08:27Like, for me as a
• 08:28parent, if I was to
• 08:29hear that in the literature
• 08:31that in the community that
• 08:34mental disorders start at fourteen.
• 08:35That's when I'd start to
• 08:36focus on, okay, I need
• 08:37to I've got a fourteen
• 08:38year old, this is when
• 08:39I need to be concerned
• 08:40about, when I need to
• 08:41look out for mental disorders
• 08:43or mental health symptoms.
• 08:45But that's it's a median
• 08:46age of onset. Right? So
• 08:48by the time,
• 08:49a child reaches fourteen,
• 08:51if you were going to
• 08:52develop a a disorder,
• 08:53fifty percent of those people
• 08:55will have already experienced
• 08:57clinical
• 08:58symptoms.
• 08:59But
• 09:01what's even more important about
• 09:02this age of onset
• 09:04data
• 09:05is that it it's pulled
• 09:06together
• 09:07of all of
• 09:09mental disorders. So when you
• 09:11pull it out
• 09:12a little bit more
• 09:17There we go. I just
• 09:17need to press a bit
• 09:18harder. So the yellow line
• 09:20is anxiety disorders, and that
• 09:21the median age of onset
• 09:23is actually eleven years. And
• 09:25depression, mood disorders all lumped
• 09:26together, is thirty years. And
• 09:28fourteen is only there because
• 09:31you're pooling data from eleven
• 09:33and thirty. So it's actually
• 09:34not very helpful as a
• 09:35parent to think, okay. Well,
• 09:37this is when I need
• 09:37to start paying attention.
• 09:39And if you dive even
• 09:41deeper to the anxiety disorders,
• 09:43age of onset,
• 09:45starts
• 09:46for specific phobia is five
• 09:48years, separation is seven years.
• 09:50So
• 09:52interventions that start in adolescence
• 09:54are completely missing the boat.
• 09:55You know, these problems have
• 09:57already been around for a
• 09:58really long time.
• 10:00But in Australia, this age
• 10:01of onset of fourteen has
• 10:03been used to argue for
• 10:04interventions being delivered
• 10:06in adolescence. So that's where
• 10:08we have our services.
• 10:09And if you're a child
• 10:10presenting with anxiety and depression,
• 10:13you don't get help until
• 10:14you're a teenager.
• 10:16Unless maybe there's suicidal risk
• 10:18or self harm, then you
• 10:19might get to see community,
• 10:21care. Otherwise, you pretty much
• 10:23wait until you're an adolescent
• 10:25because that's when the services
• 10:26are. And a lot of
• 10:27kids age out of the
• 10:29child services while they're waiting
• 10:30because there is such a
• 10:31huge demand because mental disorders
• 10:34do not start in adolescence.
• 10:35They're starting in childhood, and
• 10:37then we're missing the boat.
• 10:39So
• 10:41the the other thing about
• 10:43this data is that it's
• 10:45all from
• 10:46adults. There's no children or
• 10:48adolescents included in this age
• 10:50of onset data.
• 10:52There is prevalence data around
• 10:53that asks about age of
• 10:54onset for kids, so we're
• 10:55trying we're doing a meta
• 10:57analysis at the moment,
• 10:58systematic review,
• 11:00pooling all of the data
• 11:01for
• 11:03children and adolescent prevalence studies
• 11:05to see if there's a
• 11:06difference in age of onset.
• 11:08Obviously, there will be because
• 11:09we won't have those kids,
• 11:10who are developing
• 11:11problems in adulthood.
• 11:13But if we get closer
• 11:14to the time and asking
• 11:16parents as well, when did
• 11:17these problems first begin? I
• 11:19think we're gonna get a
• 11:20very different picture.
• 11:25Alright.
• 11:26So the work that we've
• 11:27real really been doing,
• 11:29is trying to improve the
• 11:31lives of children and teenagers
• 11:34with anxiety disorders, not just
• 11:35teenagers, but both kids and
• 11:37teens.
• 11:38And we've been doing a
• 11:39lot of work on clinical
• 11:40trials, trying to develop treatments
• 11:43that work,
• 11:45psychological treatments and medication treatments
• 11:47as well, trying to reduce
• 11:49the symptoms and the long
• 11:50term impact that anxiety symptoms
• 11:52and disorders might have on
• 11:54a on a young person
• 11:55and missing out on education,
• 11:57missing out on social interactions.
• 12:01But
• 12:03the big question that we've
• 12:05been really searching for,
• 12:07not just in anxiety disorders,
• 12:09but across mental health research,
• 12:10is being able to answer
• 12:12this question,
• 12:13that was posed by Gordon
• 12:14Paul quite a long time
• 12:16ago, which is what treatment
• 12:18works,
• 12:19by whom,
• 12:20for this individual with that
• 12:22specific problem,
• 12:23under which set of circumstances.
• 12:25So this is really what
• 12:27we've been striving for and
• 12:28the holy grail, really, in
• 12:30in terms of
• 12:32personalised care or improving treatment
• 12:34outcomes. We want a treatment
• 12:36and we want to be
• 12:37able to help children in
• 12:38a way that really matches
• 12:40their needs,
• 12:41and choosing the right treatment.
• 12:44But the kind of underlying
• 12:46assumption of this particular
• 12:48question or series of questions
• 12:50is that personalization or precision
• 12:52is possible, that it will
• 12:54deliver different outcomes.
• 12:56So that's what we've been
• 12:57on this search for is
• 12:59trying to find out is
• 13:00there a treatment that might
• 13:01work differently for different children
• 13:03or does one work better
• 13:04than another?
• 13:06Under what circumstances?
• 13:07So this is what we've
• 13:08really been striving for for
• 13:10quite some time.
• 13:13But personalization itself, like with
• 13:15this question, is not really
• 13:16new. It is something that
• 13:20use my arrow instead. Yeah.
• 13:23The mouse does not like
• 13:24my fingers.
• 13:26Alright. So personalization, as I
• 13:27said before, is not really
• 13:28new. It's something that clinicians
• 13:30have been doing forever,
• 13:32but the the idea of
• 13:33science driven personalization
• 13:35is quite new.
• 13:36In that,
• 13:38at the moment even though
• 13:39clinicians make these choices about
• 13:41okay I'm gonna try this
• 13:42instead of that or I'm
• 13:43gonna, add this particular treatment
• 13:45component at the moment
• 13:48we don't have the science
• 13:49to drive those decisions.
• 13:51We have made some inroads
• 13:53into this and I'm going
• 13:53to take you through some
• 13:54of the work that we
• 13:55have been doing around this
• 13:57space but,
• 13:59yeah, I don't think we're
• 14:00there yet.
• 14:02So really I'll just backtrack
• 14:04a little bit because the
• 14:05most of the work that
• 14:06we have been doing within
• 14:08the child mental health space
• 14:09and within anxiety disorders is
• 14:11a one size fits all
• 14:12approach. It's very different to
• 14:14the what treatment works for
• 14:15whom, under what circumstances.
• 14:18It's if you present with
• 14:19the,
• 14:20with an anxiety disorder, it
• 14:22can be any one of
• 14:23them, generalized anxiety, separation anxiety,
• 14:26specific phobia,
• 14:28social anxiety,
• 14:31even OCD, that the treatment
• 14:33would be very similar.
• 14:34We have this one size
• 14:35fits all.
• 14:38It's different a little bit
• 14:39from adult treatment if an
• 14:40adult might be more likely
• 14:41to present with a,
• 14:43anxiety disorder, like social anxiety
• 14:45disorder, and you might get
• 14:46more disorder specific treatment there.
• 14:49But really, it's, in the
• 14:50child space, it's more a
• 14:52one size fits all. And
• 14:53this makes sense because of
• 14:54high comorbidity between,
• 14:56different symptoms and different disorders,
• 14:59and it seems economical to
• 15:01have this one size fits
• 15:02all approach.
• 15:04And it's true really for
• 15:05both psychological and pharmacological,
• 15:07interventions.
• 15:10So this one size fits
• 15:11all is kind of CBT
• 15:13in this lumped package, and
• 15:14we know that
• 15:16it's,
• 15:17it's been established as a,
• 15:19efficacious treatment for anxiety disorders.
• 15:21There's been many, many trials.
• 15:24And when you look at
• 15:25what those treatment components are,
• 15:27there's lots of different components
• 15:29that make up
• 15:30cognitive behavior therapy. And if
• 15:32you,
• 15:33hear about a a CBT
• 15:35therapy, it can have any
• 15:36one of these components,
• 15:37more emphasis on one over
• 15:39another, but they tend to
• 15:40have a psychoeducation,
• 15:42cognitive restructuring,
• 15:44gradual exposure,
• 15:46where the child faces gradually
• 15:48faces their fears,
• 15:50parent training as well, where
• 15:51parents are encouraged to reduce
• 15:53accommodation
• 15:55or,
• 15:56reduce their overprotection
• 15:58or kind of rushing in
• 15:59to help,
• 16:00and also a relaxation component
• 16:02that goes onto that. But
• 16:03that's kind of the standard
• 16:04treatment,
• 16:06package.
• 16:07That'd be about right for
• 16:08your program at the moment,
• 16:09Wendy, or your standard CBT.
• 16:11Sounds familiar.
• 16:13And we know that CBT
• 16:15works. Like, when you compare
• 16:16it to a wait list,
• 16:17yep. We get really strong
• 16:18effects. Compared to doing nothing,
• 16:21we know CBT works. When
• 16:22we compare it to an
• 16:23attention control,
• 16:24it works better, but the
• 16:26effects are a little bit
• 16:27weaker.
• 16:28When we compare it to
• 16:29treatment as usual, so another
• 16:31treatment, another psychological treatment, the
• 16:34effects are a little bit
• 16:35weaker again. They're there, but
• 16:37not as consistent,
• 16:38and,
• 16:39there's a little bit more
• 16:41limitation in those findings and
• 16:42not as strong.
• 16:44So keep this in mind
• 16:46as I go through these,
• 16:47the next sets of,
• 16:50data as well.
• 16:52But importantly, I want to
• 16:54draw attention to the fact
• 16:55that these effects are all
• 16:57on average. So we've got
• 16:58an average treatment package,
• 17:00and we look at means
• 17:01and standard deviations. On average,
• 17:04how does everybody
• 17:05how does the typical child
• 17:07within the sample
• 17:10fare?
• 17:11And so we have we're
• 17:12kinda lumping
• 17:14all of the children that
• 17:15present,
• 17:16and all of the variation
• 17:18that we do experience and
• 17:20we do see in treatment
• 17:21outcome studies, they're all lumped
• 17:22together.
• 17:24But when we look at
• 17:25that average, we know that,
• 17:27fifty percent fifty to sixty
• 17:29percent of kids are likely
• 17:30to improve.
• 17:32But we don't really have
• 17:33any,
• 17:34methods for predicting or, previously
• 17:37hadn't had any methods for
• 17:38predicting that variability.
• 17:40How do we know who's
• 17:41gonna respond?
• 17:43So
• 17:44and our our,
• 17:46our treatment approaches really
• 17:48will
• 17:49and in terms of guidelines,
• 17:50I know we've got treatment
• 17:51guidelines in Australia, but I
• 17:53know you've also got them
• 17:54here,
• 17:54that,
• 17:56recommend CBT
• 17:57for a child presenting with
• 17:59an anxiety disorder.
• 18:00But let's just say we've
• 18:01got two different children.
• 18:03We've got a,
• 18:04Keira, a nine year old.
• 18:06She's,
• 18:08they're both, both of these
• 18:10children are
• 18:11medication naive. They've never had
• 18:13any,
• 18:14medications
• 18:17before.
• 18:19Keira is a nine year
• 18:21old. She comes from her
• 18:22dad found her dad's family
• 18:23comes from a Polynesian island.
• 18:25Mum was born in India.
• 18:27She's very anxious about her
• 18:28schoolwork and getting things right.
• 18:30She's a bit perfectionistic.
• 18:33She meets criteria for generalized
• 18:35anxiety disorder and social anxiety
• 18:37disorder.
• 18:38Her parents,
• 18:39particularly her dad, is very
• 18:41anxious,
• 18:42and she's also started to
• 18:43experience some mild depressive symptoms.
• 18:46Jay, on the other hand,
• 18:47is a sixteen year old
• 18:48boy. He hasn't previously had
• 18:50huge amounts of anxiety. He's
• 18:51been coping okay. He's just
• 18:53started driving and started experiencing
• 18:55regular
• 18:56panic attacks that are now
• 18:57coming out of the blue.
• 18:59He comes from a a
• 19:00single parent household,
• 19:03and he's been having these
• 19:04panic attacks regularly and seeking
• 19:05treatment.
• 19:07At the moment, our treatment
• 19:08guidelines would say that the
• 19:11CBT
• 19:11treatment is the same.
• 19:14All all we know from
• 19:15science is that you would
• 19:16apply the same approaches. But
• 19:17as a clinician, you'd probably
• 19:18think, well, I do something
• 19:20differently with each of these.
• 19:21But yet at the moment,
• 19:23the science says
• 19:24CBT or is that that's
• 19:26the recommendation.
• 19:29Alright. So we've been trying
• 19:30to move away from this
• 19:31one size fits all approach
• 19:33and trying to understand whether
• 19:36a personalized evidence based or
• 19:37a precision medicine
• 19:39approach
• 19:40might work a little bit
• 19:41better. So the goals of
• 19:43personalization
• 19:44really, you've gotta be able
• 19:45to make an accurate diagnosis
• 19:46or an accurate picture or
• 19:48profile. Doesn't necessarily have to
• 19:49be diagnosis if you don't
• 19:51come from a a medical
• 19:52model.
• 19:53Being able to predict what
• 19:55that individual risk is, being
• 19:57able to predict how likely
• 19:58the child might be to
• 20:00respond
• 20:01suboptimally to treatment.
• 20:03But the ultimate goal of
• 20:05personalization
• 20:06is really to achieve a
• 20:07more effective
• 20:10response.
• 20:12There are a number of
• 20:13ways to approach personalization or
• 20:15precision medicine in this case.
• 20:17You can adapt treatment for
• 20:19subgroups. So if you've got
• 20:20a child presenting with social
• 20:21anxiety like Kira, in the
• 20:23example before,
• 20:25you may
• 20:26deliver a different treatment and
• 20:28adapt it for children
• 20:30with,
• 20:31particular in particular subgroups.
• 20:35Modular therapies are another kind
• 20:36of precision approach where depending
• 20:39on the presentation, you might
• 20:40add, or subtract different modules,
• 20:44of treatment,
• 20:45components.
• 20:46And there's also more individualized
• 20:49metrics and predictive analytics that
• 20:51are using kind of more,
• 20:53recent techniques to try and,
• 20:56predict ahead of time based
• 20:58on previous sets of data,
• 21:00how likely is it that
• 21:01this child is to respond
• 21:03to treatment.
• 21:05So using
• 21:07predictive analytics
• 21:08or previous datasets to be
• 21:10able to predict
• 21:12with the goal of being
• 21:13of being able to get
• 21:14a better treatment response.
• 21:16And sorry. I should have
• 21:17said this is, the work
• 21:18of a PhD student who's
• 21:19now
• 21:20in the process of becoming
• 21:21a postdoc, Liz Alberti, who
• 21:23also had a pastrami sandwich
• 21:25yesterday and loved it.
• 21:28She's
• 21:29unfortunately, the the
• 21:31the pool of Yale was
• 21:32not quite as strong as
• 21:33New York City, so she's
• 21:34staying in New York City.
• 21:35Sorry to say, everyone.
• 21:37But,
• 21:38she's having fun driving around
• 21:40Central Park today instead. It's
• 21:42her first trip.
• 21:44Anyway, she's been doing this
• 21:45work as part of her
• 21:46PhD in trying to look
• 21:47at,
• 21:48how we can shift the
• 21:50field
• 21:51more towards,
• 21:52using science to drive personalization.
• 21:55So the different methods that
• 21:56you might want to use
• 21:57to get towards
• 21:59personalization or precision medicine, you
• 22:01want to understand what the
• 22:02predictors are, so which children
• 22:04are likely to respond to
• 22:05treatment.
• 22:06And then also, really importantly,
• 22:08what are the moderators?
• 22:09Just knowing who's gonna have
• 22:11suboptimal
• 22:12responses
• 22:13is not helpful if we
• 22:14don't know anything
• 22:15if we don't know what's
• 22:16gonna work better for them.
• 22:18And then mediators are also
• 22:20important,
• 22:21less so for the baseline,
• 22:23predictors,
• 22:25and predictive metrics, but it's
• 22:27important in being able if
• 22:28we're targeting
• 22:29treatments that
• 22:30actually work on the mechanisms
• 22:33that that's perhaps what we
• 22:34can target,
• 22:35to personalize care.
• 22:39Alright. So
• 22:41I wanna focus just a
• 22:42little bit on the predictors
• 22:44aspects. That's some of the
• 22:45work that we've been doing,
• 22:46looking at what predicts treatment
• 22:48outcomes. So
• 22:49can we identify
• 22:51ahead of time so when
• 22:52a child presents to treatment,
• 22:55can we identify who's which
• 22:57one of them which which
• 22:58of the children is not
• 22:59gonna respond favorably?
• 23:01So rather than this,
• 23:04you know, kind of going
• 23:05and treating everybody the same,
• 23:06it's before that treatment starts,
• 23:10what is it who's gonna
• 23:12respond to that treatment?
• 23:14And we're actually in a
• 23:15pretty good place
• 23:17in the last five years
• 23:18in terms of answering that
• 23:20question. The moderators not so
• 23:21much but predictors we we
• 23:23have a much better understanding
• 23:25of.
• 23:27One of the things that's
• 23:28complicated this piece of work
• 23:29is around the the,
• 23:31the issues around methods,
• 23:33that there are a lot
• 23:34of different methods to look
• 23:36at treatment effects.
• 23:37So we have,
• 23:39we also use in in
• 23:41child work, we use the
• 23:42clinicians report, we use parent
• 23:44report, and we use child
• 23:45report,
• 23:47and you can get three
• 23:48different treatment effects,
• 23:50based on those three different
• 23:53reports.
• 23:54So it's then how do
• 23:54you work out what treatment's
• 23:56gonna work,
• 23:57if you're using different methods?
• 23:59You know, which one are
• 24:00you gonna
• 24:01give preference to a priority
• 24:03to? We also have kind
• 24:04of different methods of looking
• 24:06at
• 24:07at the effect. We can
• 24:09use the end point. So
• 24:10remission has the disorder that
• 24:12they started treatment with. Is
• 24:13that no longer there?
• 24:15And that's an end point
• 24:16where they they get to
• 24:18at the end of treatment.
• 24:19Or you can also look
• 24:20at in terms of change
• 24:21in severity.
• 24:23And that change in severity
• 24:25really is determined by how
• 24:26high or how severe they
• 24:28were in the first place.
• 24:29So sometimes
• 24:30when you look at predictors
• 24:31of outcome or even treatment
• 24:33effects, those two different those
• 24:35things will be different in
• 24:36terms of,
• 24:37in terms of
• 24:40effects. Alright. So
• 24:42now ten years ago, we
• 24:43set out on this piece
• 24:44of work looking at, well,
• 24:45what does the literature say
• 24:46so far? What do we
• 24:47know about predictors of outcome?
• 24:50And we did a systematic
• 24:51review at the same time
• 24:52as Michael Fassum did a,
• 24:54systematic review, and we pretty
• 24:55much came to the same
• 24:56conclusion.
• 24:57It's usually devastating when you
• 24:59see that somebody else has
• 24:59done a systematic review at
• 25:01exactly the same time as
• 25:02you, but always reassuring that
• 25:04you come up with the
• 25:05same as the same answer.
• 25:06That's the whole purpose of
• 25:07a systematic review. But,
• 25:09that there were very few
• 25:11consistent predictors.
• 25:13Not that they were not
• 25:14there, it's just that they
• 25:15weren't consistent.
• 25:16There was a few hints
• 25:17in the literature.
• 25:19Social anxiety disorder predicted poorer
• 25:22outcomes and, in fact, generalised
• 25:23anxiety predicted better outcomes.
• 25:27Greater symptom severity predicted greater
• 25:29change but poorer endpoints.
• 25:32Comorbid disorders
• 25:34like, having
• 25:35comorbid depression,
• 25:37in particular, so it wasn't
• 25:39comorbidity
• 25:40generally,
• 25:41predicted
• 25:43poorer treatment response. Although there
• 25:44were some studies in this
• 25:46review that was not just,
• 25:48depression but
• 25:49tended to be associated with
• 25:51poor treatment response,
• 25:54and parent symptoms as well.
• 25:55If a parent also presented
• 25:56with anxiety or depression that
• 25:58the the child had poorer
• 26:00endpoints.
• 26:01But whenever we found a
• 26:03study that showed
• 26:05social phobia, social anxiety predicted
• 26:07outcome, we found another study
• 26:08that didn't show it. So
• 26:09it was so inconsistent that
• 26:10we really we concluded that
• 26:12there was no consistent predictors.
• 26:16But there were so many
• 26:17limitations with the the work.
• 26:19There was you know, we
• 26:21we only ever power our
• 26:22clinical trials to show a
• 26:23treatment effect of, does treatment
• 26:25a work better than treatment
• 26:27b? We never power it
• 26:28enough to be able to
• 26:29look at,
• 26:30predictors or moderators. So all
• 26:32of the the studies in
• 26:33this
• 26:34review tended to be small
• 26:36sample sizes.
• 26:37There was limited power, hence
• 26:39the reason why some studies
• 26:40showed that depression predicted outcome
• 26:43and other studies didn't. If
• 26:44a one
• 26:46sample of a hundred only
• 26:47had five children with, comorbid
• 26:50depression,
• 26:51then,
• 26:52the likelihood
• 26:53of, that study being able
• 26:55to show an effect
• 26:56is very different from, you
• 26:58know, study that had much
• 26:59greater numbers. So we had
• 27:01limited power,
• 27:03different assessment methods, as I
• 27:05talked about, different definitions of
• 27:07outcome.
• 27:08And also,
• 27:09you know, if you look
• 27:10at one variable relating to
• 27:11another and it's only related
• 27:14to it's only related to
• 27:15outcome because of its relationship
• 27:17with another variable,
• 27:18you know, you're gonna get
• 27:19inconsistency depending on whether or
• 27:21not that variable has been
• 27:22controlled for. So it was
• 27:23a lot of mess in
• 27:24the literature. And so we
• 27:25thought, well, what about if
• 27:27we try to address this
• 27:28issue of
• 27:30of limited
• 27:31power and increased our sample
• 27:32sizes? So,
• 27:35Wendy was involved in this
• 27:36work as well where we
• 27:37were actually interested more in,
• 27:39genetic predictors of treatment outcome,
• 27:41which I will mention briefly.
• 27:43But we needed for genetic
• 27:44studies, we needed much larger
• 27:46samples. So it was fantastic
• 27:48because we got much bigger
• 27:49samples,
• 27:50and then we're able to,
• 27:52to have potentially be
• 27:54more power to be able
• 27:55to detect,
• 27:56predictors.
• 27:57We also pulled by that
• 27:58stage. We'd had a number
• 27:59of clinical trials, so we
• 28:00were able to pull the
• 28:01data together. So it's like
• 28:03when you pull data, can
• 28:04we get more of
• 28:07understanding of predictors through through
• 28:09combining
• 28:10RCTs?
• 28:11So this is a study,
• 28:13I did two thousand and
• 28:14fifteen just with our samples
• 28:16that we'd collected within our
• 28:17own lab.
• 28:18And
• 28:20we were able to really
• 28:21consistently show with this much
• 28:23larger sample that children with
• 28:25social anxiety disorder on the
• 28:26left
• 28:27had much poorer outcomes. So
• 28:29it's not that they
• 28:31were not responding. They were
• 28:32responding to treatment, but they
• 28:34just weren't responding as well
• 28:36as children with other presentations.
• 28:38And interestingly,
• 28:39OCD here often people separate
• 28:41that as a separate you
• 28:43know, have a different type
• 28:44of treatment, but we consider
• 28:45it
• 28:46part of the anxiety disorders
• 28:48and showed that actually our
• 28:49CBT works, very well, much
• 28:52better than it does for,
• 28:53social anxiety.
• 28:56We also looked at comorbidity
• 28:58as well and showed that
• 28:59children with comorbid mood disorder
• 29:01had poorer
• 29:03lower remission rates, poorer endpoints.
• 29:06So pooling again this data
• 29:07together was able to consistent
• 29:09like, kind of have a
• 29:10more robust finding and something
• 29:11that we can be more
• 29:12confident in.
• 29:15We also,
• 29:17looked
• 29:18oh, this is the genetic
• 29:19study. So this is an
• 29:20even bigger sample size of
• 29:21fifteen hundred. We looked at
• 29:24this is including,
• 29:26data,
• 29:28from
• 29:29Yale and Florida,
• 29:32combining
• 29:33the data sets from around
• 29:35the world.
• 29:36We had
• 29:37samples from Oxford, from Groningen,
• 29:40from Amsterdam,
• 29:41Denmark, Norway, and Sydney,
• 29:44pulling them together
• 29:45and showed again
• 29:47with social anxiety disorder.
• 29:50This is
• 29:52a change or a change
• 29:53in severity
• 29:54that the slope of the
• 29:55line is it's it it's
• 29:57still improving,
• 29:58but not as well as
• 30:00children with other,
• 30:02primary anxiety disorders.
• 30:06We also looked more closely
• 30:08at parental mental health as
• 30:09well
• 30:10and pulling the data together.
• 30:13And
• 30:14parental psychopathology was also something
• 30:15that was quite inconsistent in
• 30:17the literature. Sometimes it showed
• 30:19an effect and sometimes it
• 30:20didn't. And when we pulled
• 30:22this data together, it helped
• 30:24really understand
• 30:25why that might be is
• 30:26that pre to post, you're
• 30:28getting similar effects regardless of
• 30:29whether or not the parent
• 30:31also experiences anxiety, both mothers
• 30:33and fathers.
• 30:35But
• 30:36the further you go
• 30:38away from the treatment
• 30:39that the clinic,
• 30:41the more that the parental
• 30:43psychopathology
• 30:44impacts on
• 30:45the clinical severity rating of
• 30:46the child's anxiety. So when
• 30:48there's no parental psychopathology,
• 30:51kids continue to,
• 30:53have a lower level of
• 30:55anxiety post treatment on average.
• 30:57But on average,
• 30:59kids who had parents that
• 31:01were more likely to have
• 31:02anxiety and depression,
• 31:04their their symptom levels went
• 31:05back up again. So this
• 31:06pooling of data was really
• 31:08exciting for us. We it
• 31:09came
• 31:11with having a stronger understanding
• 31:13of what predicts poorer outcomes.
• 31:15So being at the point
• 31:16where we can say, okay.
• 31:17When a child presents for
• 31:18treatment,
• 31:19we have a good understanding
• 31:21of who's gonna
• 31:22do less well.
• 31:24So if you have a
• 31:25social anxiety disorder, comorbid mood
• 31:27disorder, and if the parents
• 31:29present with,
• 31:31anxiety or, depression.
• 31:35That's really helpful, but
• 31:38we wanna know what to
• 31:39do with it. Right? Like,
• 31:40how do we we don't
• 31:41wanna not offer them treatment
• 31:42just because we know they're
• 31:43not gonna do as well.
• 31:44We wanna know what to
• 31:46do, what else we can
• 31:47do. So can we actually
• 31:48use this,
• 31:49knowledge to inform treatment decision
• 31:51making?
• 31:53And that's really, I suppose,
• 31:55when we need to look
• 31:55at moderators and what what
• 31:57can
• 31:58what can improve outcomes.
• 32:00We we started on exploring
• 32:02just whether or not we
• 32:03can use this as a
• 32:05precision
• 32:05tool
• 32:07to,
• 32:08you use perhaps in the
• 32:09clinic, not necessarily that will
• 32:11shape your treatment decisions, but
• 32:13in terms of
• 32:15shaping your understanding of treatment
• 32:17outcome
• 32:18of kind of pulling all
• 32:19of these predictors together. So
• 32:21we with this data set,
• 32:22the genetic data set I
• 32:24was talking about before, we
• 32:26looked at the idea or
• 32:27kind of explored the idea
• 32:28of
• 32:30developing
• 32:30a a a risk index
• 32:33that might help us
• 32:35be able to make those,
• 32:36decisions and be able to
• 32:38know who's not gonna respond,
• 32:40well to treatment.
• 32:41So I'm not quite at
• 32:42the moderators yet. So we're
• 32:44changing treatment still back at
• 32:45predictors,
• 32:46just to clarify. So we
• 32:48we created,
• 32:49we we're just using regression
• 32:52methods,
• 32:53and we looked at all
• 32:55the predictors in our data
• 32:56set
• 32:57and,
• 32:59chose those variables that were
• 33:01strong predictors, and then we
• 33:03developed a score. So if
• 33:04a child presented with social
• 33:06anxiety, they got a risk
• 33:07score of one. It was
• 33:08kinda rounded beta weights. They
• 33:09got parental anxiety. If the
• 33:11the either parent had anxiety,
• 33:13we'd give them a risk
• 33:14score of one.
• 33:16And low mood was actually,
• 33:17in this particular model, had
• 33:18a a high beta weight.
• 33:20We're also looking at, genetic
• 33:22predictors as well. I'm not
• 33:23gonna go into that too
• 33:24much today because it was
• 33:25a really exciting
• 33:26initial finding. But as with
• 33:28all genetic studies,
• 33:30or not all of them,
• 33:31almost all of them, it
• 33:32never replicated.
• 33:34So we we did show
• 33:35that,
• 33:36the SS allele looking at
• 33:38differential
• 33:38susceptibility predicted
• 33:40better outcomes that if you
• 33:42were more responsive to your
• 33:44environment,
• 33:45then you responded better to
• 33:46CVT, but we haven't been
• 33:47able to replicate that. So,
• 33:49waiting till the data can
• 33:51catch up to that one.
• 33:52But, anyway, this was just
• 33:54an example of how we
• 33:55added the predictors,
• 33:57the potential risk together to
• 33:59create a score. And then
• 34:00we looked at using that
• 34:01score in another dataset how
• 34:04how likely it is that
• 34:05the child will respond to
• 34:07treatment. So on the left
• 34:08hand side of the graph,
• 34:10children with a score of
• 34:11zero to two,
• 34:13so that's
• 34:14not very many of these
• 34:15predictors. So they had either
• 34:17none of them or maybe
• 34:17just one or two. And
• 34:19you can see response rates
• 34:20are much higher. They're closer
• 34:22to eighty percent. So if
• 34:23you only had this small
• 34:24number of risks risks,
• 34:26then your,
• 34:28your
• 34:29chance of responding to CBT
• 34:31was much higher. Whereas, it
• 34:33was kind of this cumulative
• 34:34risk. The more of these
• 34:35risk factors that you had,
• 34:37the less likely. So if
• 34:39you had a a score
• 34:41of more than five,
• 34:43your chances of remission were
• 34:45really low, kinda dropped to
• 34:46under, you know, forty percent.
• 34:48So we're getting quite a
• 34:50lot of variability in treatment
• 34:51outcome
• 34:52based on baseline predictors.
• 34:57But
• 35:02the really important thing is
• 35:03what works better for them.
• 35:04Right?
• 35:05If we know they're not
• 35:06gonna do well,
• 35:07we need to know
• 35:09what else to do. Do
• 35:10we do CBT plus something?
• 35:12Do we do
• 35:13you know, modular therapy? Do
• 35:15we adapt it for for,
• 35:17these different presentations
• 35:19for social anxiety?
• 35:23But I'm gonna kinda spoil
• 35:24the
• 35:26surprise. I kind of already
• 35:27have, really, that our efforts
• 35:28to identify better treatment alternatives
• 35:31for likely non responders have
• 35:34been really limited.
• 35:35We haven't really got to
• 35:36the point where we can
• 35:37say one treatment works better
• 35:39than another. We've I've been
• 35:41spending the last
• 35:43fifteen
• 35:44years trying to
• 35:46develop a treatment that might
• 35:47be better,
• 35:49and it has not led
• 35:50to anything very fruitful,
• 35:52which is really disappointing. A
• 35:54lot of grant money that's
• 35:55gone into looking at does
• 35:56this treatment work better than
• 35:58that, but we haven't got
• 35:59to that point.
• 36:01There's a couple of hints
• 36:02though, so I haven't lost
• 36:03sight entirely.
• 36:05I'm gonna take you through
• 36:06the journey anyway.
• 36:07So, just
• 36:09I'll take you through
• 36:11the number of
• 36:13clinical trials that have had
• 36:15a negative outcome or not
• 36:16the outcome we were expecting,
• 36:18not negative, I should say.
• 36:19Alright. So we started on
• 36:20social anxiety disorder. So if
• 36:22we know that children with
• 36:23social anxiety disorder are not
• 36:25likely to not as likely
• 36:26to improve, maybe we need
• 36:28a disorder specific approach for
• 36:30those kids. Maybe this one
• 36:31size fits all is not
• 36:33the best
• 36:34best approach.
• 36:35So we looked at the
• 36:36adult literature, and there was
• 36:37some hint of this in
• 36:38the adult literature as well
• 36:39for social anxiety disorder that,
• 36:43approach
• 36:44that focuses more on behavioral
• 36:46experiments,
• 36:47giving video feedback,
• 36:49more safety behaviors,
• 36:50attention training, and exposure to
• 36:52costs.
• 36:53But that,
• 36:55so kind of, what I
• 36:57mean by exposure to costs
• 36:58is,
• 37:00facing, like, an exposure where
• 37:02a child has to deliberately
• 37:03get laughed at in front
• 37:04of others or,
• 37:06you know, kind of that
• 37:07exposure
• 37:08to the dreaded the dreaded
• 37:10situation. But that idea
• 37:12that we're trying to make
• 37:14targeting the mechanisms of social
• 37:15anxiety more specifically.
• 37:20And
• 37:22Sue Spence did a study
• 37:23on this also from Australia
• 37:24showing,
• 37:25kind of, it was an
• 37:26online version transdiagnostic
• 37:28versus disorder specific. So kind
• 37:30of a one size fits
• 37:31all to the disorder specific
• 37:33approach and found very little
• 37:34difference between the two.
• 37:37No difference, but there was
• 37:38a highlight of dropout because
• 37:39of the was the online,
• 37:41attrition in this
• 37:43particular study, and she didn't
• 37:44have the video feedback aspect.
• 37:47So we,
• 37:48misguidedly
• 37:49embarked on a new clinical
• 37:50trial where we, which we
• 37:52targeted those a little bit
• 37:53more specifically,
• 37:55but,
• 37:56there was some effects that
• 37:57follow-up,
• 37:59kids were slightly
• 38:01better,
• 38:02with the disorder specific treatment,
• 38:04but it wasn't robust across
• 38:06across measures,
• 38:08and it was a very
• 38:10weak effect. And if you're
• 38:11making decisions at a policy
• 38:13level, what treatment to deliver,
• 38:16yeah, there wouldn't be much
• 38:18justification
• 38:18for this disorder specific treatment.
• 38:22That's the same.
• 38:24So that was a little
• 38:25bit disappointing that the idea
• 38:26of disorder specific treatment might
• 38:28not be the solution for
• 38:30social anxiety disorder.
• 38:32I wanted to also present
• 38:33some work that has not
• 38:35come out of, my lab,
• 38:36but, John Weiss and Bruce
• 38:38Chorpedo have been using this
• 38:39modular approach
• 38:41to depending on how the
• 38:42child presents, if I had
• 38:43to present with anxiety, they
• 38:44give them an anxiety module.
• 38:46If they present with depression,
• 38:47they get a depression module.
• 38:48It's that match treatment if
• 38:50you're familiar with it.
• 38:52And
• 38:53here,
• 38:54you can see the effects
• 38:55for the modular treatment on,
• 38:58internalizing symptoms are much better.
• 39:01But, again, it's not a
• 39:02consistent finding. It's not across.
• 39:04It's only in this symptom
• 39:06measure. It's not in remission
• 39:07rates.
• 39:08And, it hasn't been able
• 39:10to be replicated either.
• 39:12So I'm not necessarily convinced
• 39:14that this modular approach,
• 39:17is working. You could,
• 39:18consider adding a depression module
• 39:21for,
• 39:22you know, with a standard
• 39:23anxiety protocol, it may lead
• 39:25to better outcomes. But at
• 39:26this stage, I'm not necessarily
• 39:28convinced by the data. And
• 39:29I can see Wendy shaking
• 39:30her head. She's not either.
• 39:35So we also tried Okay.
• 39:36Well, if parents who have
• 39:38anxiety,
• 39:39and those children are less
• 39:40likely to do well, what
• 39:41if we add treatment for
• 39:43parents at the same time
• 39:44and give parents more skills,
• 39:46to be able to manage
• 39:47their own anxiety? Would that
• 39:48have better longer term outcomes
• 39:50for kids? Do you know
• 39:52what the answer's gonna be?
• 39:54No. It does not.
• 39:56Alright. So additional treatment.
• 39:58There was maybe a weak
• 40:00effect at post, but it
• 40:02was washed out by three
• 40:03months later.
• 40:05And it yeah. It Again,
• 40:07with quite a good sample
• 40:09of two zero nine,
• 40:10a hundred and something in
• 40:11each condition,
• 40:13there was no
• 40:14strong evidence that we should
• 40:16be giving one treatment to
• 40:18a child over another.
• 40:23Cathy Cresswell did a very
• 40:25similar trial to the one
• 40:26that we did looking at
• 40:28treatments for anxious parents, tested
• 40:30all sorts of different,
• 40:32methods either focusing on the
• 40:33parenting or focusing on parent
• 40:35child interaction
• 40:37versus a standard CBT, and,
• 40:38again, no difference.
• 40:41So this kind of search
• 40:42for,
• 40:44an enhanced treatment or a
• 40:46better treatment,
• 40:47we're just not finding it.
• 40:49So and particularly with the
• 40:50answer to prayer and anxiety,
• 40:51we haven't really solved that.
• 40:52There's, you know, perhaps perhaps
• 40:53some hint around the modular
• 40:54treatment, but we haven't really
• 40:56solved the problem of what
• 40:58to do with it if
• 40:59a family presents with,
• 41:01the parent experiencing anxiety as
• 41:03well as the child.
• 41:07I swore I wasn't gonna
• 41:08come back to this genetic
• 41:09stuff, but I
• 41:10forgot I had this slide
• 41:11in here. So,
• 41:13I'll just briefly touch on
• 41:14it because we did this
• 41:15hasn't been replicated.
• 41:17So,
• 41:18yeah,
• 41:20hold your judgment of it.
• 41:22But, we actually found,
• 41:24a moderation based on polygenic
• 41:26risk scores.
• 41:28In one sample, we looked
• 41:29at,
• 41:30differential susceptibility.
• 41:32So we looked at genetic
• 41:33markers
• 41:34or a polygenic risk score
• 41:36for susceptibility.
• 41:37So it's those children who
• 41:39were more likely to respond
• 41:40to their environment. So I
• 41:41don't know if you know
• 41:42about the orchids and dandelion,
• 41:45child, that that idea where
• 41:46differential susceptibility is about, that
• 41:48there's kids that are very
• 41:50responsive to their environment. They
• 41:51can thrive in the right
• 41:53conditions, but in the wrong
• 41:54conditions, they will not thrive
• 41:56like most of the orchids
• 41:57in my that I have
• 41:59attempted to,
• 42:00to raise.
• 42:02They tend to just never
• 42:03survive. Whereas the dandelion is
• 42:05the idea. It doesn't matter
• 42:06what kind of environment. They
• 42:07can grow inside the road,
• 42:09anywhere. They'll be they'll thrive.
• 42:10So that's the idea of
• 42:11differential susceptibility. So we were
• 42:13interested in, first of all,
• 42:14whether or not that would
• 42:15predict treatment outcome,
• 42:18as a predictor. But then
• 42:19we also started to play
• 42:20around with, you know, was
• 42:22there any moderators. So
• 42:24as I said, we haven't
• 42:25been able to replicate the,
• 42:27prediction
• 42:28and also haven't been able
• 42:29to replicate this finding, but
• 42:31it's still very interesting from
• 42:34a a personalization
• 42:35or,
• 42:36trying to understand
• 42:37can we experience can we
• 42:39detect
• 42:40differential effects.
• 42:42So this was the large,
• 42:45genetic samples, fifteen hundred children.
• 42:48They either received individual CBT,
• 42:50they might have received group
• 42:51CBT,
• 42:52or parent led CBT.
• 42:54So they kind of vary
• 42:56in intensity in a way.
• 42:57Individual is one therapist with
• 42:59one family or one child.
• 43:02Group is kind of less
• 43:04intense in that you've got
• 43:05one therapist to five or
• 43:07six kids. And then parent
• 43:08led is that there's no
• 43:09therapist. That's really the parent
• 43:11as a therapist.
• 43:14And we actually found what
• 43:15we would have expected
• 43:17in that,
• 43:19those kids who are high
• 43:20on differential susceptibility,
• 43:22so the orchid children,
• 43:24that's I know that's a
• 43:24terrible label. I suppose it's
• 43:26better than a diagnostic label,
• 43:27isn't it? Orchid children,
• 43:30were more
• 43:31had better remission rates if
• 43:33they received individual therapy,
• 43:36compared to
• 43:38children who were
• 43:40high in differential
• 43:42susceptibility, but if they received
• 43:43parent led CBT. So if
• 43:45they didn't have the the
• 43:46treatment that wasn't as intensive
• 43:48and was led by the
• 43:49parent,
• 43:50then they were
• 43:51they had lower likelihood of
• 43:53remission.
• 43:56Potentially,
• 43:57a genetic
• 43:58moderation,
• 44:01but it hasn't been replicated
• 44:03yet. And, actually, I don't
• 44:04know whether,
• 44:06at least in my lifetime,
• 44:07it would be replicated because
• 44:08it's it's a very difficult
• 44:10study to
• 44:11run and,
• 44:13yeah. It's a it's a
• 44:14hint that there may be
• 44:15something there.
• 44:18Alright. In terms of improving
• 44:20outcomes,
• 44:22we
• 44:24there's a study not me.
• 44:26Sorry.
• 44:27Walkup and colleagues
• 44:28did in two thousand and
• 44:30seven looking at, well, what
• 44:31about if we added medication
• 44:32to CBT? Could we get
• 44:33better outcomes?
• 44:35And it was very exciting
• 44:36in the in the literature.
• 44:37It was pretty robust finding
• 44:39that,
• 44:40across all measures, pretty much
• 44:42that if you got CBT
• 44:44plus sertraline,
• 44:46that those kids had much
• 44:47better outcomes compared to CBT
• 44:49alone or sertraline alone,
• 44:53and a lot better than
• 44:54placebo. So that's pretty convincing.
• 44:55And I think in the
• 44:56in the US, at least,
• 44:57there was a big shift
• 44:58in treatment that kids were
• 45:00getting this combination of of
• 45:02treatment.
• 45:03But when I saw these
• 45:04findings, I wasn't particularly convinced
• 45:06about it because
• 45:07if you,
• 45:08as a as a parent,
• 45:10as a child,
• 45:11going in to see a
• 45:12psychiatrist and a psychologist, you
• 45:14know you're getting the best
• 45:15treatment. And
• 45:16compared to kids just receiving
• 45:18a treatment from a psychologist
• 45:20or receiving treatment,
• 45:21a medication treatment.
• 45:22So those families that were
• 45:24receiving double care were actually,
• 45:27that they they weren't,
• 45:29they were aware of their
• 45:31superiority of their treatment. So
• 45:33maybe they had an impact.
• 45:34So I went about doing
• 45:35a,
• 45:37oh, yeah, not adequately controlled.
• 45:39So I did a it's
• 45:40a smaller study, but it
• 45:42was just two groups,
• 45:43CBT and sertraline and CBT
• 45:45in a sugar pill, and
• 45:46no difference.
• 45:48This study doesn't get cited
• 45:49because it's nowhere near as
• 45:52palatable to the drug companies
• 45:54that we and also, you
• 45:56know, not particularly useful in
• 45:57terms of implications
• 45:59that we just need to
• 45:59give kids sugar pills. That's
• 46:01not the answer. But the
• 46:02idea that maybe
• 46:03the expectations
• 46:04of medication
• 46:05might work to enhance our
• 46:08outcomes. Alright. So there's also
• 46:10some other hints in the
• 46:11literature about what might improve
• 46:13treatment outcomes for kids. We
• 46:14know that the more exposures
• 46:16that happen in a treatment,
• 46:17so the,
• 46:19therapist led exposures
• 46:21and also more exposures at
• 46:22home,
• 46:23and also the more challenging
• 46:24exposures. So rather than just,
• 46:27small tasks where a child
• 46:28has to raise their quest
• 46:30raise their hand in class,
• 46:31but actually doing something really
• 46:33challenging, like deliberately making a
• 46:35mistake in front of the
• 46:36school assembly,
• 46:37that the more challenging the
• 46:39exposures are, the better outcomes.
• 46:40So this was really helpful
• 46:43for me. I kind of
• 46:44shifted after seeing this literature
• 46:46thinking, okay. Well, we really
• 46:48need to focus more on,
• 46:50exposures and
• 46:54improving
• 46:56that
• 46:57component in the in the
• 46:58treatment.
• 47:02But when you look at
• 47:03what actually clinicians do in
• 47:04the community, very few receive
• 47:07or,
• 47:09they're less likely to favor
• 47:11exposure techniques. This is a
• 47:12bit of a hard graph
• 47:13to read, but this we
• 47:14asked the clinicians in Australia
• 47:16what techniques they use to
• 47:17treat children with anxiety disorders.
• 47:18We had a whole list,
• 47:19and then we, had a
• 47:21look at it. And, actually,
• 47:23I've cut it off at
• 47:23the top of the graph,
• 47:24but exposure was up there
• 47:26with psychodynamic
• 47:27techniques that people are likely
• 47:29to use in terms of,
• 47:31the
• 47:32one to five scale frequency
• 47:34of use. But people are
• 47:35much more likely to use
• 47:36psychoeducation,
• 47:37much less likely to use
• 47:39exposure.
• 47:40We did another study asking
• 47:41kids in the community who
• 47:43had anxiety,
• 47:44what type of treatment did
• 47:45you receive? And actually, very
• 47:47few children reported
• 47:49using this technique that we
• 47:50think is really critical.
• 47:52So there's limited research on
• 47:54the technique specifically in young
• 47:55people. So that's the some
• 47:56of the work that I've
• 47:57been doing, which I'm not
• 47:58gonna go too much into
• 48:00today because we're running out
• 48:01of time. I wanna make
• 48:02sure we've got time for
• 48:03questions, but I'm on this
• 48:04mission to try and make
• 48:05expo exposure cool again because
• 48:07nobody is using it,
• 48:09and,
• 48:10trying to
• 48:11use it more effectively. And
• 48:12so we're doing some work.
• 48:13We've done a developed a
• 48:15parent led digital intervention called
• 48:17Courage Quest,
• 48:18where parents can take their
• 48:20their kids through this quest
• 48:22of trying to,
• 48:26face their fears and use
• 48:27exposure. We're actually doing a
• 48:28factorial design. I'm not gonna
• 48:30spend too much time on
• 48:30it because it's a little
• 48:31bit of a sidetrack,
• 48:32but we're looking at five
• 48:33different features trying to test,
• 48:36what might make exposure work
• 48:38more effectively and to be
• 48:39able to give clinicians a
• 48:41little bit more information about,
• 48:43how to use exposure effectively
• 48:45and what might make it
• 48:46work better.
• 48:48Alright. That was a bit
• 48:49of a bit of a
• 48:50tangent. I wanna come back
• 48:51to,
• 48:53differential treatment effects and moderators.
• 48:55We did another systematic review
• 48:57that Lizelle's just submitted,
• 48:59for publication trying to look
• 49:01at
• 49:02moderators. Do we know what's
• 49:03gonna work better for these
• 49:04kids?
• 49:06There was fifty five studies
• 49:08that looked at moderators
• 49:09on thirty eight trials
• 49:11and
• 49:12lots of different control conditions,
• 49:15and you can predict the
• 49:17outcome. No definitive effects were
• 49:19identified.
• 49:22No clear moderation or differential
• 49:24treatment effects. So we know
• 49:26let's say say that social
• 49:27anxiety
• 49:29children children with social anxiety
• 49:30don't do as well. There's
• 49:32no effects that show
• 49:34what might work better for
• 49:35those children.
• 49:37There were a couple of
• 49:38hints in the literature again.
• 49:39We had four moderator variables
• 49:41of interest that were consistent
• 49:43across those fifty five studies.
• 49:46Interestingly, Hispanic ethnicity.
• 49:48I'm not necessarily convinced by
• 49:50that, though, because that really
• 49:51came from one trial,
• 49:54so I'm not so sure
• 49:55how robust it is, but
• 49:56at least,
• 49:57it came
• 49:59it emerged in more than,
• 50:01one study.
• 50:02Age was also a predictor,
• 50:04a moderator,
• 50:06social anxiety and parental psychopathology.
• 50:09But
• 50:11importantly, even though those variables
• 50:13came up as important moderators,
• 50:15in one study, there was
• 50:16a moderation effect one way,
• 50:18and in another study, there
• 50:19was a moderation effect in
• 50:20another way. So for social
• 50:22anxiety disorder, one study showed
• 50:23group worked better, and another
• 50:25study showed,
• 50:27group CBT didn't work as
• 50:29well as individual therapy. So
• 50:30we had moderation, but completely
• 50:32opposing effects.
• 50:34So each result really points
• 50:36to something different about the
• 50:38the direction.
• 50:39A study that didn't go
• 50:40into the
• 50:42systematic review that I just
• 50:43wanted to point to because
• 50:44this is still leaving me
• 50:45with a little bit of
• 50:46hope. And this is one
• 50:47of Wendy's,
• 50:48papers that she's just published,
• 50:51showing a moderation effect
• 50:53for two parent variables
• 50:55that,
• 50:56compared CBT versus
• 50:59CBT plus relationship training,
• 51:02was moderated by parental acceptance.
• 51:04So child outcomes were better
• 51:07for families where there was
• 51:09high parental acceptance,
• 51:11when CBT included more of
• 51:13a focus on, increased acceptance
• 51:16and decreasing control compared to
• 51:18CBT. So that gives us
• 51:19an idea of what could
• 51:21be done,
• 51:23what how we might wanna
• 51:24shape treatment,
• 51:26that if a child presents
• 51:27with high with parents with
• 51:29high acceptance, then this is
• 51:30the type of treatment that
• 51:31we might that might produce
• 51:32better outcomes,
• 51:34and a similar
• 51:35effect for,
• 51:37parental or negative reinforcement as
• 51:39well. So there's a hint
• 51:41there, but it still needs
• 51:41to be replicated.
• 51:43And,
• 51:47it's also kind of,
• 51:49I think the study wasn't
• 51:50designed right to to find
• 51:53these moderation effects, but they
• 51:54were there. So in designing
• 51:56studies
• 51:57for moderation is a very
• 51:59different approach to kind of
• 52:00the normal,
• 52:01treatment design.
• 52:04Alright. So can we enhance
• 52:07outcomes in answer to that
• 52:08question? At the moment, I'm
• 52:09not really convinced by it.
• 52:10I'd be curious to see
• 52:11after being being presented with
• 52:13all that data.
• 52:15To me, there's pretty limited
• 52:17differential treatment effects so far
• 52:19In being able to answer
• 52:20this question, what works by
• 52:22whom for this individual with
• 52:23that specific problem, apart from
• 52:25the fact that we're honing
• 52:26in on anxiety disorders,
• 52:28we've got very limited information
• 52:30about,
• 52:31what works.
• 52:32So this could be either
• 52:33there are no differential treatment
• 52:35effects, doesn't matter, evidence based
• 52:37personalisation
• 52:37doesn't lead to better outcomes,
• 52:39there is no holy grail,
• 52:41or it could be that,
• 52:44these effects
• 52:45are there they're there, but
• 52:47we just haven't been able
• 52:47to detect them because our
• 52:49methods haven't been rigorous enough.
• 52:50We still haven't designed our
• 52:52studies in a way that,
• 52:54can consistently,
• 52:56show these robust findings.
• 52:58I'm not quite ready to
• 52:59swallow this bitter pill of
• 53:01every every treatment works for
• 53:02everybody the same amount.
• 53:05I'm still hoping and holding
• 53:06on to the fact that
• 53:07we may actually have,
• 53:09there may be differential treatment
• 53:11effects there, but we just
• 53:12haven't,
• 53:13we haven't got there yet.
• 53:14So we've been doing some
• 53:15individual patient data meta analysis.
• 53:17I'm just gonna finish up.
• 53:18Yep.
• 53:19Where we're combining
• 53:21data from a whole lot
• 53:22of clinical trials together,
• 53:24so kind of advancing the
• 53:26previous work that we've done,
• 53:27but we're doing it in
• 53:27a systematic way now. We're
• 53:29developing this topic based living,
• 53:31repository of data. We're storing
• 53:34it on a secure server.
• 53:35At the moment, I'm very
• 53:37excited. The first time last
• 53:39week, I got my first
• 53:40data set of seven thousand
• 53:42one hundred and twenty three
• 53:43kids from around the world.
• 53:44And
• 53:46I haven't even put the
• 53:47arrow in the right spot
• 53:49for Yale. But, anyway, you
• 53:50there is Yale data in
• 53:52there.
• 53:53But yes.
• 53:55I don't know geography, clearly.
• 53:58I don't know where that's
• 53:59going to. But yeah. So
• 54:00that's what we're working on
• 54:01now,
• 54:02that seven thousand. And it'll
• 54:04hopefully give us a lot
• 54:04more variability in the studies.
• 54:06We've got
• 54:07a lot of different variation
• 54:09in treatments
• 54:11that, you know, we might
• 54:11not have been able to
• 54:12predict before, but now we'll
• 54:14have much more, information and
• 54:16be at a be at
• 54:16a better point where we
• 54:18can recommend
• 54:19treatment and be able to,
• 54:22match treatments hopefully in the
• 54:23future, but we're not there
• 54:25yet.
• 54:26So thank you very much
• 54:27for letting me take you
• 54:28through that. I'm much appreciated.
• 54:35What
• 54:36a well, I'll clear your
• 54:38presentation,
• 54:38Jenny. It was fantastic. So
• 54:40we we do have, like,
• 54:40five minutes for some questions
• 54:42that people would like to
• 54:43ask questions.
• 54:51Hi. Thank you for coming
• 54:53to speak to us.
• 54:55My question is that,
• 54:58in addition to,
• 54:59looking at
• 55:01differential responses to treatment and
• 55:03and prognostic
• 55:04factors,
• 55:06are others interested? And or
• 55:08do you think there's utility
• 55:09in,
• 55:12using these or related predictive
• 55:14measures
• 55:14to,
• 55:16for early detection
• 55:17or for,
• 55:19like, stratifying
• 55:21of,
• 55:22Of anxiety disorders.
• 55:23Yeah. So we haven't really
• 55:25been looking at risk of
• 55:27disorder in the first place
• 55:29because all of the samples
• 55:30already have a disorder. So
• 55:31we're looking at risk, and
• 55:33risk gets used in different
• 55:34ways. So risk we're looking
• 55:35at risk in terms of
• 55:36predicting
• 55:37poorer outcome.
• 55:39So
• 55:40I do think there is
• 55:41utility,
• 55:43in well, also in detection.
• 55:45But at the moment, the
• 55:46idea of what else to
• 55:48do with them, that's where
• 55:49we're stuck.
• 55:50But clinicians have been doing
• 55:52this
• 55:53for as
• 55:54forever long that there's been
• 55:56clinical practice. So we adapt
• 55:58treatments based on what we
• 55:59think should work based on
• 56:00our treatment formulation,
• 56:02and we expect that if
• 56:04we adapt
• 56:05based on risk,
• 56:07that
• 56:08there might be better outcomes.
• 56:10But at the moment, we
• 56:10just haven't got that bit
• 56:12to say that they're gonna
• 56:13do better.
• 56:14But I I do think
• 56:15there's value in the risk
• 56:17index and,
• 56:18that we that's our next
• 56:20step is to actually take
• 56:21it at least that far
• 56:22as developing a platform for
• 56:24clinicians to enter in that
• 56:26risk data to see, okay.
• 56:28Well, based on this massive
• 56:29dataset,
• 56:30how likely is it that
• 56:31the child that I'm about
• 56:32to see with these particular
• 56:34profiles,
• 56:35these particular variables, how likely
• 56:37is it that this standard
• 56:39CBT is gonna work?
• 56:41I wouldn't want to see
• 56:42it being used for,
• 56:44if
• 56:45okay. It comes out that
• 56:46the likelihood of response is
• 56:48very low,
• 56:49then I wouldn't want them
• 56:50to start some completely different
• 56:52treatment. Right? It's,
• 56:54yeah. So it's it's
• 56:56I I do think there's
• 56:57utility, but I think there's
• 56:58a bit of concern around
• 57:00then saying, okay. Well, CBT
• 57:01doesn't work, then let's do
• 57:03let's, you know,
• 57:05do some,
• 57:07something entirely different that there's
• 57:08no evidence base for.
• 57:11Yeah. It seems like it's
• 57:12at least low hanging fruit
• 57:13to say Yeah. I'll start
• 57:14CBT or start the SSRI
• 57:16earlier. Yes. Yeah. Yeah. Yeah.
• 57:18Definitely.
• 57:19Yeah.
• 57:21Yeah. And if
• 57:23yeah. Given that there has
• 57:24been some hint in the
• 57:25literature around that we get
• 57:27better effects for CBT and
• 57:28sertraline that if there is
• 57:30a poor risk
• 57:32for treatment,
• 57:33outcome being favorable,
• 57:35then, yes, starting the CBT
• 57:37earlier.
• 57:39As, yeah, the combined treatment
• 57:41earlier. Yeah.
• 57:43Thank you. That was really
• 57:44interesting. I was just wondering
• 57:45what you, could tell us
• 57:46about the differences between Australia
• 57:48and here in the US
• 57:50in anxiety,
• 57:52in kids. What what makes
• 57:54kids anxious and how differentially
• 57:56parents respond?
• 57:59Yeah. It's a good question.
• 58:01I'm gonna be talking anecdotally.
• 58:05Very similar. I have worked
• 58:06in the clinic in the
• 58:07US, in Philadelphia, as well
• 58:09as in Australia. But Philadelphia
• 58:11was a long time ago,
• 58:13and it was
• 58:14before cell phones.
• 58:16Very similar things, though,
• 58:18that they worry about. A
• 58:19lot of you know, depending
• 58:21on the type of presentation,
• 58:23but there's
• 58:24I've seen this increase in
• 58:26concern around academic pressure,
• 58:29particularly,
• 58:31yeah, even in primary school
• 58:33ages, kids worrying about how
• 58:35they're gonna perform in the
• 58:36future and more parental pressure
• 58:38than I think we've seen
• 58:39in previous year previous generations.
• 58:45And in terms of types
• 58:46of anxiety,
• 58:50we have a lot more
• 58:51spiders. Yeah. Yeah. Yeah. Put
• 58:53eyes.
• 58:54No. There's a koala. Yeah.
• 58:55No. I don't I mean,
• 58:57we don't we don't get
• 58:59I've never seen a phobia
• 59:00of a koala before.
• 59:03But we do have,
• 59:04you know, a lot of
• 59:05spider phobias, but not more
• 59:07than in other countries, but
• 59:08yet we've got a lot
• 59:09more risk.
• 59:10I found a couple of,
• 59:13deadly spiders in my one
• 59:14in my bedroom and one
• 59:15in my swimming pool,
• 59:17a while ago. But yet
• 59:19even though the risk is
• 59:21higher,
• 59:22more yeah. Exactly.
• 59:25Yeah. Yeah. But,
• 59:28not in terms
• 59:29of presentations being different in
• 59:31terms of prevalence of disorders.
• 59:39Oh, thank you. Thanks for
• 59:41such a beautiful talk, Jenny.
• 59:43I was fascinated by this
• 59:44finding, the JCAP one, where
• 59:45you show that, you know,
• 59:47kids with parents with severe
• 59:48psychopathology or psychopathology are similar
• 59:50at post, but then you
• 59:51see these longer term effects.
• 59:52And I'm wondering if you
• 59:54can speak to, you know,
• 59:55what do you think is
• 59:55driving that? What's the mechanism?
• 59:59Yeah. I
• 01:00:01my gut feeling about it
• 01:00:03is that parents
• 01:00:05who
• 01:00:06are more likely to be
• 01:00:07anxious
• 01:00:08and
• 01:00:09who have maybe learnt some
• 01:00:10of the skills than when
• 01:00:12they get back in their
• 01:00:12environment with the young person
• 01:00:15that,
• 01:00:17because they've had a really
• 01:00:18long time of being anxious,
• 01:00:20most likely
• 01:00:21that that in interaction
• 01:00:23kinda goes back to normal.
• 01:00:25And that particularly for the
• 01:00:26child,
• 01:00:27that they're exposed to this
• 01:00:29more modeling. Like, in those
• 01:00:31in that graph,
• 01:00:33that wasn't the parents didn't
• 01:00:35also receive treatment. So they
• 01:00:37they received it in terms
• 01:00:38of through their child, and
• 01:00:40they learned about the strategies.
• 01:00:41And so we do get
• 01:00:43even when we don't target
• 01:00:44it specifically, we do get
• 01:00:45a reduction in parental anxiety,
• 01:00:48even when it's not targeted.
• 01:00:50So there would have been
• 01:00:51a reduction in parental anxiety
• 01:00:52at post treatment,
• 01:00:54but that wasn't,
• 01:00:56you know, enough really to
• 01:00:57stop the child being exposed
• 01:00:59to increased anxiety and modeling
• 01:01:02and avoidance behavior. So I
• 01:01:03think they just go back
• 01:01:04into an environment where there's
• 01:01:06continued avoidance and,
• 01:01:08modeling of anxiety and accommodation.
• 01:01:12Well, thank you again, Jenny.
• 01:01:14I we've that's we
• 01:01:16we're ending now. But thank
• 01:01:18you all very much for
• 01:01:19attending. However, the trainees have
• 01:01:21an opportunity now to meet
• 01:01:22with Jenny. Right now, the
• 01:01:24room is over there in
• 01:01:25the Giselle. I believe that's
• 01:01:27what we have scheduled if
• 01:01:28that's,
• 01:01:29so that's what it is.
• 01:01:30And there might be some
• 01:01:31I will see you next
• 01:01:32door. You stay here for
• 01:01:33now. But if you get
• 01:01:34kicked out, go there. Is
• 01:01:35that okay? Oh, okay. Yeah.
• 01:01:36Let's do that. Okay. So
• 01:01:38thank you all so much
• 01:01:39for having me. Jenny with
• 01:01:40a wonderful talk. It's thank
• 01:01:41you so much. Thank you.