Yale Psychiatry Grand Rounds: "Leveraging Experience Sampling to Advance Health and Wellbeing Among Victims of Intimate Partner Violence."

Transcript

• 00:00About,
• 00:02two women who share their
• 00:03deeply personal
• 00:05experiences
• 00:06with us to make this
• 00:08research possible.
• 00:10Obviously, they're coming in to
• 00:11talk about partner violence,
• 00:13very sensitive topic, and it
• 00:14can be difficult to participate.
• 00:16So a a huge note
• 00:17of thanks and gratitude to
• 00:18women who are willing to
• 00:19share those experiences.
• 00:21Along with staff,
• 00:23too many to include on
• 00:24this slide, but many are
• 00:26noted here.
• 00:27In the middle, I've had
• 00:28exceptional mentors
• 00:30along the way in the
• 00:31twenty three plus years that
• 00:32I've been here at Yale,
• 00:34as well as incredible colleagues.
• 00:36Community partners are on the
• 00:38right along with funders.
• 00:39So as Stephanie mentioned, in
• 00:41terms of my program of
• 00:43research, I focus on individual
• 00:44and system level factors that
• 00:46affect the well-being of mostly
• 00:48women who experience intimate partner
• 00:50violence. And so the individual
• 00:52level factors are noted here,
• 00:53those that promote resilience and
• 00:55recovery
• 00:56and those that increase risk.
• 00:58Today, I'll focus more most
• 01:00probably on, post traumatic stress
• 01:02and substance use, so I'll
• 01:03talk about HIV,
• 01:05and other, factors as well.
• 01:08At the systems level, I
• 01:09focus on the capacity of
• 01:10systems
• 01:11to meet the unique needs
• 01:13of women who experience intimate
• 01:14partner violence because their needs
• 01:16are largely unique from, women
• 01:19who have not experienced
• 01:21IPV.
• 01:21And then last in terms
• 01:22of intervention development,
• 01:24I did not I say
• 01:25this a lot with folks
• 01:26folks that I work with.
• 01:27I did not grow up
• 01:28as an intervention
• 01:29development,
• 01:31scientist. This is recent for
• 01:32me, so I didn't grow
• 01:33up, like, learning RCTs,
• 01:35but in the last five
• 01:36years or so have become,
• 01:38more skilled at developing and
• 01:40testing interventions. And we'll speak
• 01:41to you at the end
• 01:42a little bit about that.
• 01:44And so
• 01:45first,
• 01:46I'm gonna introduce micro longitudinal
• 01:48designs and their benefits.
• 01:50I used experience sampling in
• 01:51the title. That's one method,
• 01:53microlongitudinal
• 01:54method. This talk is gonna
• 01:55be a little bit different
• 01:57than some talks. It's not
• 01:58a here are my methods,
• 02:00here are the design, here's
• 02:01the statistics, the measurement, the
• 02:03outcome with all of,
• 02:05lots of numbers and formulas.
• 02:07That's really not this, this
• 02:09talk.
• 02:12Instead, I'm really gonna integrate
• 02:14the micro longitudinal
• 02:15designs,
• 02:17and the benefits of them
• 02:18with study
• 02:20findings, to talk with you
• 02:21both about this design and
• 02:23its benefits,
• 02:25as well as what it's
• 02:26helped us come to learn
• 02:27about women who experience intimate
• 02:29partner violence.
• 02:30And so elements central to
• 02:31my work of note, I
• 02:33collaborate with community members. Sometimes
• 02:36it's always women with lived
• 02:38or living experience
• 02:39as well as folks that
• 02:41serve them. It's critically important
• 02:43to me that I center
• 02:44the voices of the women
• 02:46that we work with
• 02:47and that it reflects to
• 02:49the extent possible their lived
• 02:51experiences.
• 02:53And so to start with
• 02:54what is intimate partner violence,
• 02:55I promise to be quick.
• 02:56You might think in twenty
• 02:57twenty four we don't need
• 02:59this discussion,
• 03:00but we actually do for
• 03:01some folks.
• 03:02So the media depictions are
• 03:04what a lot of people
• 03:05come to understand as intimate
• 03:06partner violence. And those media
• 03:08depictions tend to focus on
• 03:10victims
• 03:11who are fragile
• 03:13and helpless.
• 03:14They tend to be depicted
• 03:16as women only.
• 03:17Those women are abused by
• 03:19angry, mean men that you
• 03:20could point to in a
• 03:21room if you walked in.
• 03:23Women are severely physically abused
• 03:25in media depictions. We see
• 03:27black eyes, broken bones,
• 03:29someone cowering in the corner
• 03:30in fear for their life
• 03:32on a daily basis.
• 03:33And absolutely,
• 03:35do victims and abusers that
• 03:37look like this exist?
• 03:39Absolutely. There's no question. I
• 03:40could do presentation
• 03:41after presentation
• 03:42on
• 03:44opportunities to help individuals and
• 03:45communities because we're looking for
• 03:46that very specific type. So
• 03:47what is intimate partner violence?
• 03:48It's really
• 04:00type. So what is intimate
• 04:01partner violence?
• 04:03It's really much more about
• 04:04a pattern
• 04:05of behaviors. Of course, it
• 04:07includes physical abuse, but there
• 04:08are many more strategies and
• 04:10tactics than that. It includes
• 04:12sexual abuse, verbal and emotional
• 04:14abuse, And the intent really
• 04:15is to threaten, harm, or
• 04:18intimidate,
• 04:19to control someone, isolate, or
• 04:21monitor them. So it's a
• 04:23range of strategies with a
• 04:25range of of motivations and
• 04:26intent.
• 04:27We'll skip physical IPV in
• 04:29terms of describing.
• 04:31Sexual IPV, we often think
• 04:33about sexual assault or abuse
• 04:35in the context of a
• 04:37dating partner or an acquaintance.
• 04:39We're way past the don't
• 04:40run, you know, don't jog
• 04:42in the park at night
• 04:42because some stranger is gonna
• 04:43jump out from the bushes.
• 04:45That, of course, happens.
• 04:47But we think more about,
• 04:48sexual assault in the context
• 04:50of acquaintance relationships.
• 04:52We don't often think about
• 04:53it in the context of
• 04:54partner violence, but one third
• 04:56of women who experience
• 04:58partner violence are made to
• 04:59do sexual things that they
• 05:00do not want to do.
• 05:02So it's highly prevalent in
• 05:03this population.
• 05:05In terms of psychological abuse,
• 05:08I would love to present
• 05:09to you for hours on
• 05:10psychological
• 05:11abuse. It is critically important,
• 05:13understudied, and under focused on
• 05:15in interventions.
• 05:17It's it's the spectrum of
• 05:18behaviors. It's name calling, put
• 05:21downs, humiliation
• 05:23to the end of the
• 05:24spectrum of monitoring where someone
• 05:26goes, what they do, what
• 05:28they can wear, keeping them
• 05:30from family and friends. And
• 05:31then there's financial and economic
• 05:33abuse, limiting the amount of
• 05:34money someone has access to,
• 05:36taking their paycheck, or or
• 05:37the opposite, not allowing them
• 05:39to work, and to be
• 05:40self sufficient.
• 05:42So let's talk about IRL
• 05:44or in real life.
• 05:46In a therapy session, I'm
• 05:47a psychologist as as doctor
• 05:49O'Malley mentioned. In a therapy
• 05:51session, if a client describes
• 05:52struggling with a recent episode
• 05:53of depression, I as a
• 05:55clinician might say, so in
• 05:57the last six months, how
• 05:58often did you use drinking
• 05:59to cope? Never, sometimes, often,
• 06:02always?
• 06:03In the last six months,
• 06:04how often do you cope
• 06:05by daydreaming about better times?
• 06:07Never, sometimes, often, always?
• 06:10For those of you on
• 06:11Zoom, you can't see that
• 06:11we're in a small room,
• 06:13but some of you are
• 06:13probably trying to figure out
• 06:14how to make it to
• 06:15the exit thinking if I
• 06:16have to listen to this
• 06:17person for the next hour,
• 06:18if that's how they do
• 06:19a therapy session. Of course,
• 06:20we would never ever say
• 06:22that in a therapy session.
• 06:24But that's how we assess
• 06:26things when we aggregate.
• 06:27Right? And we have to.
• 06:28There are many studies that
• 06:29have to be done that
• 06:30way, but that's not how
• 06:31people live their lives. In
• 06:33a therapy session, we would
• 06:34ask what was going on
• 06:35before, during, and after the
• 06:37episode of depression, who they
• 06:39spent time with, what strategies
• 06:40they tried to use to
• 06:42cope, how they worked out.
• 06:44So we would ask about
• 06:45day level experiences. We wouldn't
• 06:47aggregate.
• 06:49Research needed to get with
• 06:50the times. Now I said
• 06:51this twenty years ago before
• 06:53I did my first micro
• 06:54longitudinal study.
• 06:55So in walks a micro
• 06:57longitudinal
• 06:58design known by many names.
• 07:00Experience sampling is in my
• 07:01title, intensive longitudinal
• 07:03designs, real time data collection.
• 07:06What is a micro longitudinal
• 07:08design?
• 07:09So essentially, it's capturing lived
• 07:12experiences or data from the
• 07:14perspective of a researcher,
• 07:16with great frequency
• 07:18in near real time and
• 07:19in someone's natural environment. So
• 07:22for example, it's an assessment
• 07:24a day or multiple times
• 07:26a day completed at home
• 07:28on a smartphone, for example.
• 07:30So this is a hypothetical
• 07:32data example. These are not
• 07:33real data points based on
• 07:35real data.
• 07:36But essentially,
• 07:37if we were developing an
• 07:38intervention to reduce PTSD symptoms,
• 07:40most of us,
• 07:42again, not an intervention developer
• 07:44per se, but but I
• 07:45know enough to know we
• 07:46would do baseline and follow-up
• 07:47assessments. Right? So here's our
• 07:48baseline assessment of PTSD symptoms,
• 07:51and ideally, we would see
• 07:52a decrease that was maintained
• 07:54at three and six months.
• 07:58But what could we do
• 07:59with a micro longitudinal
• 08:00design, and what could we
• 08:02understand? Right? If we're if
• 08:03we're targeting something in an
• 08:04intervention like PTSD symptoms,
• 08:07we know that at the
• 08:08day level,
• 08:09someone doesn't reach a level
• 08:11of a PTSD symptom, and
• 08:12there are many, and then
• 08:14stay at that level for
• 08:15the next six months. Of
• 08:16course, that doesn't happen. PTSD
• 08:18symptoms, depression symptoms, name name
• 08:20the mental health disorder symptoms
• 08:22fluctuate.
• 08:23And so a micro longitudinal
• 08:25design can help us understand
• 08:27this fluctuation
• 08:28at the day level or
• 08:29even within a day. And
• 08:31so you can see this
• 08:32says day one one day
• 08:34one two one three. This
• 08:35is for example four surveys
• 08:37a day. So one one
• 08:38is day one survey one
• 08:40day one two etcetera.
• 08:43So what are the benefits?
• 08:46Well, first, it reduces recall
• 08:48bias. When you're asking someone
• 08:50a question in near real
• 08:51time about what happened yesterday
• 08:53or this morning, it's certainly
• 08:54better than asking them what
• 08:56happened last week. I can't
• 08:58remember what happened last week.
• 09:00But we ask people to
• 09:01recall over long periods of
• 09:03time. And so this method
• 09:04reduces recall bias, so more
• 09:06accurate information.
• 09:08It improves ecological validity because
• 09:10it's collected in someone's own
• 09:11environment. They're not coming into
• 09:13our office here at Yale.
• 09:15They're sitting on their couch.
• 09:16They're on the bus. They're
• 09:17at work.
• 09:18It allows us to identify
• 09:20proximal and contingent
• 09:22relationships. So to speak with
• 09:23greater certainty about how things
• 09:25are associated
• 09:27in real time because actually,
• 09:30we found that we can't,
• 09:32misrepresent
• 09:32the experiences
• 09:34of individuals
• 09:35when we aggregate.
• 09:38And it identifies more precisely
• 09:40targets for intervention. So here
• 09:42you'll see an overlay, again,
• 09:43hypothetical example
• 09:45of substance use,
• 09:47on top of the PTSD
• 09:50symptoms. And you might take
• 09:51a quick look and say,
• 09:52oh, well, it looks like
• 09:53substance use in yellow.
• 09:55Does it does it happen
• 09:56more often after PTSD symptoms
• 09:59kind of fluctuate, kind of
• 10:00peak upward?
• 10:02And so a micro longitudinal
• 10:04design can test that, which
• 10:05would help us better understand
• 10:07these really specific, precise relationships.
• 10:10But the first question is
• 10:11it safe and feasible to
• 10:12do this with women who
• 10:13are experiencing intimate partner violence?
• 10:17So I had,
• 10:19developed a career development award
• 10:21to answer that very question.
• 10:23Is it safe and feasible
• 10:24to apply this method
• 10:26with women in the community
• 10:27who are currently experiencing intimate
• 10:30partner violence?
• 10:32It was a ninety day
• 10:33design driven feasibility study where
• 10:35we would randomize women to
• 10:37one of three ways to
• 10:38give us daily reports. One
• 10:39was with automated telephone surveys
• 10:41where they would use the
• 10:42numbers on their keypads to
• 10:44enter responses to prerecorded questions.
• 10:47The second was paper diaries.
• 10:48Folks, this is almost twenty
• 10:49years ago. There was no
• 10:51we did not have iPhones
• 10:53like we do
• 10:54now. And so questions were
• 10:56preprinted.
• 10:57They would have,
• 10:59envelopes that were addressed to
• 11:00us and stamped. We'd capture
• 11:01the date that they sent
• 11:02them in. And then also
• 11:04monthly retrospective interviews because if
• 11:06we could, in fact,
• 11:07get accurate information
• 11:09retrospectively,
• 11:10why burden people with responding
• 11:12multiple times a day?
• 11:14But first, is a safety
• 11:16study even safe? And Yale's
• 11:18IRB said, no. It is
• 11:19not. So I got funded
• 11:21to do my career development
• 11:23award, and our IRB said,
• 11:25no. No. No. No. Absolutely
• 11:27not. They were concerned that
• 11:29the paper diaries that women
• 11:31would take home could increase
• 11:32risk because of the questions
• 11:34that were preprinted.
• 11:36And so,
• 11:37I was frustrated
• 11:39and concerned. Though, as Stephanie
• 11:41mentioned, I've been doing research
• 11:42with women for twenty years.
• 11:44I've been working with women
• 11:45who experienced partner violence since
• 11:46I was twenty. I'm fifty
• 11:48two. And I knew that
• 11:50this was safe to do.
• 11:52But you can't just say
• 11:53to the IRB, trust me.
• 11:56So after I stopped hyperventilating,
• 11:59Kathy Carroll told me to
• 12:01take a breath, and in
• 12:02her infinite wisdom said, propose
• 12:04a safety study, Tammy. Tell
• 12:06the IRB
• 12:07you'll do this with ten
• 12:09women for two weeks. You'll
• 12:10have a focus group. You'll
• 12:11come back and you'll tell
• 12:12them what you've learned. And
• 12:13if it's not safe, you
• 12:15won't proceed.
• 12:17And so I did breathe,
• 12:19and we did a mini
• 12:20safety study to pilot this
• 12:23and ultimately learned that the
• 12:25majority of women were not
• 12:26at all concerned
• 12:28about having the paper diaries
• 12:30at home. And of those
• 12:31that were, they concealed them
• 12:33as I believed they would
• 12:35be able to. So IRB
• 12:36approval was subsequently granted,
• 12:39and I was allowed to
• 12:40move forward with this once
• 12:41daily study for ninety days.
• 12:44So this safety and feasibility
• 12:46study, as I explained, we
• 12:47did randomize people to one
• 12:49of three groups.
• 12:50They also did a baseline
• 12:51interview, and they participated every
• 12:53day for ninety days telling
• 12:55us about what happened the
• 12:56day before, about partner violence,
• 12:58substance use, and then other
• 12:59key factors like face to
• 13:01face contact with their partner
• 13:02or hospitalization that could limit
• 13:04the occurrence of partner violence
• 13:05and substance use. One of
• 13:07the few slides I'll show
• 13:08you with sample demographics just
• 13:09so you get the sense
• 13:11of the sample,
• 13:12racially and ethnically diverse.
• 13:15Most were unemployed.
• 13:17One note that I will
• 13:19make
• 13:20across many of the slides
• 13:22that I have that have
• 13:22demographics.
• 13:25The average duration of a
• 13:27relationship
• 13:28at the time they come
• 13:29in for an interview
• 13:30in all of my studies
• 13:32is between
• 13:33five, and now it's ten
• 13:34years. It used to be
• 13:35five and seven years. That's
• 13:37just the time they come
• 13:38in. So these are not
• 13:39quick in and out relationships.
• 13:41These are established
• 13:43long standing relationships.
• 13:45And in terms of the
• 13:46IPV and substance use in
• 13:48the sample, the inclusion criteria
• 13:49were physical IPV, so you'll
• 13:51see a hundred percent experience
• 13:52that in the thirty days
• 13:53before baseline.
• 13:54But you'll see a range
• 13:55of IPV severity. Severe physical
• 13:58victimization, which is being, beaten
• 14:00up, having a knife or
• 14:01a gun used on you,
• 14:04kicked, etcetera,
• 14:05You can see that. You
• 14:06can see IPV injuries,
• 14:08sexual coercion or attempted rape,
• 14:10rape,
• 14:11and then you can see
• 14:12the range of substance use.
• 14:13They had to have had
• 14:14at least a sip
• 14:16of alcohol,
• 14:18prior to the baseline interview,
• 14:19the three months before, or
• 14:21it could have been just
• 14:22a hit, off a joint
• 14:24or something like that. Some
• 14:25substance use, but not substance
• 14:27use problems.
• 14:28We did learn from the
• 14:29study that micro longitudinal
• 14:31methods are safe. This was
• 14:33a safety study first and
• 14:34foremost. Nearly a hundred percent
• 14:36said their safety was not
• 14:37put at risk.
• 14:38One woman said her partner
• 14:40became suspicious about why she
• 14:42was on the phone so
• 14:42much, but it never caused
• 14:44any problems.
• 14:45Depending on condition between two
• 14:47thirds and three quarters
• 14:49reported that their partners knew
• 14:51they were participating in a
• 14:52study, and of those, another
• 14:54two thirds to three quarters
• 14:56knew the focus was on
• 14:57relationship conflict.
• 14:59Flies in the face of
• 15:00what we know could see
• 15:01faces. Flies in the face
• 15:03of what we know about
• 15:04media depictions. Would never think
• 15:05that could be true if
• 15:06you hadn't worked with this
• 15:07population.
• 15:09Speaks to self selection.
• 15:11Consistent with survivor defined practice
• 15:13and victim centered services, women
• 15:15can make their own choices
• 15:16about what is safe for
• 15:17them, even in the context
• 15:19of ongoing abusive relationships.
• 15:22And yes, these methods are
• 15:24feasible. Women found it easy
• 15:25to participate. They said the
• 15:27ninety day duration was just
• 15:29right. I'm sure that had
• 15:29to do with the fact
• 15:30that we compensated, of course,
• 15:32for daily participation.
• 15:35They said the condition specific
• 15:37training
• 15:37was appropriate because you do
• 15:39have to do some training
• 15:40at the baseline interview, make
• 15:42sure they know how to
• 15:42understand those daily questions.
• 15:44And we had approximately seventy
• 15:46two percent completion of surveys,
• 15:48which is high,
• 15:50for these kinds of studies.
• 15:52So sorry for the bright
• 15:53background on this slide, but
• 15:55this method became foundational to
• 15:57my program of research.
• 15:59And since that time, I've
• 16:01conducted multiple microlongitudinal
• 16:03studies, have been asked to
• 16:05be a coinvestigator
• 16:06on multiple others, and this
• 16:08is a list of many
• 16:09of them. I will not
• 16:10have time to talk about
• 16:11them all.
• 16:13I'm gonna start with the
• 16:14actual findings from that daily
• 16:16co occurrence or the feasibility
• 16:18study to talk to you
• 16:19about the daily co occurrence
• 16:20of partner violence. Again, this
• 16:22kind of goes against what
• 16:23we see in media depictions.
• 16:25We had two thousand seven
• 16:26hundred and seventy eight days
• 16:28of TADA. It's a lot
• 16:29of data with a micro
• 16:31method.
• 16:32And on those two thousand
• 16:33seven hundred and seventy eight
• 16:34datas days of data, we
• 16:36were looking to characterize
• 16:38day level experiences because back
• 16:40when I started doing this,
• 16:41we had no day level
• 16:42information. That was anecdotal.
• 16:45What did we learn?
• 16:47On sixty two percent of
• 16:49days,
• 16:50no partner violence occurred. No
• 16:52physical, sexual, or psychological.
• 16:54And again, that does not
• 16:55align with media depictions, but
• 16:57if you've worked with the
• 16:59victims, you know that that
• 17:00is that makes sense. The
• 17:02next most typical type of
• 17:04day was characterized by psychological
• 17:06abuse at twenty seven percent
• 17:08and then if you combine
• 17:10psychological abuse with the other
• 17:12forms of abusive behavior it's
• 17:13totaled about thirty eight percent.
• 17:16You can also see here
• 17:19that it's almost
• 17:20never the case that physical
• 17:22abuse occurs
• 17:24outside of psychological
• 17:25abuse and sexual abuse, never
• 17:27does. You can see the
• 17:28totals. Sexual IPV only point
• 17:31six days, eighteen days. Physical
• 17:33only eight days out of
• 17:34two thousand seven hundred and
• 17:37seventy eight days. And that
• 17:38makes sense when we stop
• 17:39to think about it because
• 17:41it's very rarely happens that
• 17:42someone walks into a room,
• 17:44does something physical. It's typical
• 17:46that things escalate.
• 17:47So what is what are
• 17:48the practical implications of understanding
• 17:51this? The finding that on
• 17:52most days, no partner violence
• 17:52occurs helps to explain the
• 17:53ambivalence.
• 17:54Partner violence
• 17:55occurs helps to explain
• 17:57the ambivalence that people have
• 17:57about ending abusive relationships. It
• 17:58doesn't happen all the time
• 17:59for most people.
• 18:01And there are people for
• 18:02whom it would
• 18:04be
• 18:07be safe to end the
• 18:08abusive relationship.
• 18:10Right? But this helps us
• 18:11understand that ambivalence.
• 18:12The breaks in between incidents
• 18:14of intimate partner violence
• 18:16likely promote and still hope
• 18:17that things are gonna change
• 18:19just like the abusive partner
• 18:20promises to do.
• 18:22These data also help us
• 18:23understand that psychological IPV warrants
• 18:26more attention
• 18:27in intervention and prevention programs
• 18:30because if we can reduce
• 18:31psychological IPV,
• 18:32we are likely to reduce
• 18:35experiences, events of physical IPV.
• 18:37And some interventions
• 18:38have moved on to incorporating
• 18:40emotion regulation
• 18:41that includes colleagues here, like
• 18:44Carla Stover,
• 18:45for the fathers for change
• 18:46intervention and some other interventions
• 18:48as well.
• 18:50So next, I'm gonna elucidate
• 18:51daily use of safety strategies
• 18:53with day level data.
• 18:55This is also a once
• 18:56a day ninety day study,
• 18:58but this is more recent.
• 18:59This was funded by the
• 19:00National Collaborative on Gun Violence
• 19:01Research
• 19:02to understand the experiences of
• 19:05explicit and implicit
• 19:06firearm
• 19:07threat.
• 19:09And so here, we're looking
• 19:10to characterize at the person
• 19:12and day level their experiences.
• 19:14There's so much more to
• 19:15do with this data.
• 19:17We have we just finished
• 19:19collecting data
• 19:21earlier this year last at
• 19:22the end of last year.
• 19:23Essentially, this looks like the
• 19:24other study. It's once a
• 19:25day for ninety days.
• 19:27We did conduct this study
• 19:28during COVID, so anyone who's
• 19:31conducted a study during COVID
• 19:32has stories to to tell
• 19:33about how it didn't turn
• 19:34out the way it was
• 19:35designed.
• 19:37But this was,
• 19:38first study for women who,
• 19:40were in a committed heterosexual
• 19:42relationship within the last three
• 19:44months. First study I have
• 19:45ever been able to get
• 19:46funded with an inclusion criteria
• 19:48of psychological abuse. Nobody likes
• 19:50that as an inclusion criteria.
• 19:52Everybody wants physical abuse to
• 19:54be the inclusion criteria,
• 19:55not really understanding that psych
• 19:57abuse actually can have a
• 19:58stronger impact on health and
• 20:00well-being than physical.
• 20:03They used a smartphone and
• 20:04downloaded an app to report,
• 20:06and we have sixteen thousand
• 20:07one hundred and four responses
• 20:09provided by two hundred and
• 20:10forty three women.
• 20:12These are the strategies that
• 20:13women use to stop, prevent,
• 20:16or escape the conflict. We
• 20:17ask these on a daily
• 20:18basis. I'm not gonna leave
• 20:20this up for you to
• 20:20look at because I will
• 20:21show the types of strategies
• 20:23on each slide.
• 20:24But big picture, again, to
• 20:26kind of what can you
• 20:26get descriptively from this data.
• 20:28Sometimes I think the most
• 20:30exciting information we get from
• 20:31studies is the really basic,
• 20:33boring,
• 20:34descriptive information.
• 20:37Seventy nine percent of the
• 20:38women in this study, again,
• 20:40some were recruited for psych
• 20:42abuse alone. Seventy nine percent
• 20:43used a safety strategy at
• 20:45some point over those ninety
• 20:46days.
• 20:47Thirty eight percent of all
• 20:49daily surveys were characterized by
• 20:51the use of a safety
• 20:53strategy. And the three most
• 20:55commonly used were placating,
• 20:58prayer,
• 21:00and preparing to leave the
• 21:01relationship, and you can see
• 21:02those statistics by participants and
• 21:04by surveys. Right? So sixty
• 21:06seven percent of participants or
• 21:08sixty eight percent,
• 21:09use placating strategies, and those
• 21:11happened on twenty,
• 21:13around twenty three percent,
• 21:15of days or surveys.
• 21:18And so to look at
• 21:20the day level placating strategies,
• 21:22as I just defined from
• 21:23the table, we had almost
• 21:24sixty eight percent who used
• 21:26it in twenty three percent
• 21:27of surveys where they tried
• 21:28to keep quiet, stay calm,
• 21:30pretend that nothing was wrong,
• 21:32to stop, prevent, or escape
• 21:34the conflict.
• 21:35Next was prayer.
• 21:37Some people might be wondering
• 21:38how is prayer a safety
• 21:39strategy. If you wanna ask
• 21:40questions about that at the
• 21:41end, we can get to
• 21:42that, but used by fifty
• 21:44eight
• 21:45percent of participants on twenty
• 21:47five percent of surveys.
• 21:49Preparing to leave, so packing
• 21:51a bag, hiding money, having
• 21:53a code with a neighbor,
• 21:55was used by almost forty
• 21:56percent of participants on seventeen
• 21:59percent of days.
• 22:00Creating distance, which actually can
• 22:02be the most,
• 22:04the riskiest time for participants,
• 22:07ending or trying to end
• 22:08the relationship, staying with family
• 22:10or friends, changing locks, used
• 22:12by almost forty percent, only
• 22:14in five percent of surveys.
• 22:17Weapons related strategies,
• 22:20in part what the study
• 22:21was about. This is where
• 22:22women either
• 22:24hid a knife or a
• 22:25gun or had another, weapon
• 22:27available to protect themselves
• 22:29or separately, or were they
• 22:31removed or hid guns or
• 22:33weapons,
• 22:34among almost
• 22:35twenty percent of participants
• 22:38on almost six percent of
• 22:39days.
• 22:40Domestic violence services used by
• 22:43twenty one percent of women
• 22:44across four percent of of
• 22:46days. So going to support
• 22:47group, having a safety plan
• 22:49in place, not even just
• 22:51creating one that day, but
• 22:52having it in place,
• 22:53and using the criminal justice
• 22:55system. So calling the police
• 22:57or on a day having
• 22:58a protective or restraining order
• 23:00in place, only nine percent
• 23:02and in less than one
• 23:03percent of daily surveys.
• 23:05So what does this tell
• 23:06us in summary?
• 23:08The most used strategies
• 23:10are placating prayer and preparing
• 23:12to leave, important for us
• 23:14to understand as providers.
• 23:16The least used criminal justice
• 23:18system and domestic violence service
• 23:19providers, which is what we
• 23:20all think we're gonna do
• 23:21if we have someone who's
• 23:23experiencing IPV. Let's we'll refer
• 23:25them to the local agents
• 23:26here. We'll have them call
• 23:27the police. Women do not
• 23:29want to do that for
• 23:30lots of reasons that we
• 23:31can talk about at the
• 23:32end if you'd like.
• 23:34Another example of what you
• 23:35can do with these data
• 23:36at the day level, and
• 23:37it's just one example,
• 23:39that I'm gonna share findings
• 23:40instead of doing it across
• 23:42all of them. You can
• 23:43also look at same
• 23:45day associations.
• 23:47So for example, you can
• 23:48say, on days when intimate
• 23:50partner violence was experienced, was
• 23:52there greater odds of using
• 23:53placating strategies?
• 23:55So if psych
• 23:57IPV alone so if a
• 23:58day was characterized by psych
• 24:00IPV alone, they were four
• 24:01point seven times more likely
• 24:03to use a placating strategy.
• 24:05If there was physical IPV
• 24:07that may or may not
• 24:08have included psychological, There were
• 24:10five point two greater odds
• 24:11of a placating strategy.
• 24:13And sexual IPV days, which
• 24:16could also have include physical
• 24:18and psych it's kind of
• 24:18a hierarchy where sexual is
• 24:20the the worst is how
• 24:21it was conceptualized.
• 24:23The odds increased two point
• 24:25two percent of placating.
• 24:27So it's just one example
• 24:28of other analysis you can
• 24:30do at the day level
• 24:31with those data.
• 24:33We have another study that
• 24:34looked at day level experiences
• 24:37among women who are living
• 24:38with HIV.
• 24:40When so I do, CME
• 24:42talks across the state for
• 24:44providers.
• 24:45And when I talk about
• 24:46women living with HIV who
• 24:48experience partner violence, I actually
• 24:50encourage people to think about
• 24:51how this might generalize to
• 24:52other things, other chronic health
• 24:55conditions, physical health or behavioral
• 24:56health, because what I'm gonna
• 24:57share with you may,
• 24:59be related to those conditions
• 25:01as well.
• 25:02I first want to tell
• 25:03you about pilot work we
• 25:04did to inform this day
• 25:05level study. We actually met
• 25:07with clients from, at the
• 25:08time they were called AIDS
• 25:09service organizations across the state,
• 25:11and asked women who were
• 25:12living with intimate partner violence,
• 25:13how does intimate partner violence
• 25:15influence your HIV care? And
• 25:17four themes emerged.
• 25:19The first was that their
• 25:20partner actively
• 25:22interfere in their HIV care.
• 25:23So he used to make
• 25:24fight with me and make
• 25:25me not take my medicines.
• 25:27Or Jamie Meyer, a colleague
• 25:28here who's an infectious disease
• 25:30doc,
• 25:31worked with women in the
• 25:32prison. A quote
• 25:34from one of her clients
• 25:35is perfect.
• 25:36Unfortunately, she said my partner
• 25:38would take my meds and
• 25:39flush them down the toilet
• 25:41and say which now you're
• 25:42gonna die faster.
• 25:44It doesn't get more direct
• 25:45than that. But we can
• 25:46think about what we prescribe
• 25:48I don't prescribe. What other
• 25:50people prescribe meds for,
• 25:52and how that could be
• 25:53done with any type of
• 25:54medication for any condition.
• 25:58Also, a theme was that
• 25:58partners passively interfere. Well, he
• 26:00says, you know, he forgets
• 26:02to take me to my
• 26:02doctor's appointments.
• 26:04Right? This kind of passive,
• 26:06interference.
• 26:07The third is that self
• 26:08worth was so affected
• 26:10that women didn't feel that
• 26:11self care was important. They
• 26:13didn't care enough about themselves
• 26:14to engage in their HIV
• 26:17care. This was actually the
• 26:18most endorsed theme among all
• 26:21of the women who participated.
• 26:22And again, likely links back
• 26:24to the psychological
• 26:25abuse piece. So I wanted
• 26:27to stop myself from taking
• 26:28my HIV medication so that
• 26:30like I could
• 26:31die. That's how bad my
• 26:33nerves were, what I was
• 26:34going through. I just wanted
• 26:35to like die slowly.
• 26:38And the fourth was that
• 26:39physical harm impacted relationships with
• 26:41service providers. So sometimes it
• 26:43was always
• 26:44I was bruised, and then
• 26:45like I said, I didn't
• 26:46want anyone to see that,
• 26:47and I refused to call
• 26:48the police. So I guess
• 26:50I was dodging the doctor,
• 26:51the police, and everything. So
• 26:54those themes
• 26:55informed
• 26:56a study while we proposed
• 26:57to NIMH,
• 26:58with Jamie Meyer, who I
• 26:59invited to
• 27:01MPI with me. It's a
• 27:02twice a day study for
• 27:04thirty two days, understanding the
• 27:05ways in which partner violence
• 27:07influences engagement in HIV care.
• 27:09And so at first, we're
• 27:10trying to understand at the
• 27:12day level
• 27:13what influences
• 27:14medication
• 27:15adherence, people taking their HIV
• 27:17medications, and then second, because
• 27:19obviously I do not believe
• 27:21in a one size fits
• 27:22all model,
• 27:23how are there subgroups of
• 27:24women with shared experiences
• 27:27at that day level, shared
• 27:28experiences of different factors
• 27:30that predict medication adherence, that
• 27:33ultimately influence how they engage
• 27:35in appointments with their doctors,
• 27:37and then ultimately their viral
• 27:39suppression.
• 27:41So this is gonna characterize
• 27:43day level experiences
• 27:47and ultimately
• 27:48latent classes with latent class
• 27:50analysis through different and, different
• 27:52health trajectories.
• 27:53So this is the study
• 27:54design. Essentially, they came in,
• 27:57to a clinic visit with
• 27:58their HIV care provider.
• 28:00We did a baseline I
• 28:01didn't do the anniversary. They
• 28:02did a baseline,
• 28:03in person interview,
• 28:05twice daily,
• 28:07data collection for thirty two
• 28:08days. And then eventually, they
• 28:10would go back in for
• 28:11a clinic visit at some
• 28:12point if they weren't out
• 28:13of care, and they would
• 28:14come for a follow-up interview.
• 28:17In total,
• 28:18a hundred and fifty nine
• 28:19women participated.
• 28:21We had fewer who did
• 28:22the follow-up interviews, but the
• 28:24beauty of something like this
• 28:26is viral suppression we grab
• 28:28from medical records. So it's
• 28:30unfortunate that people didn't come
• 28:31in for follow-up interviews, but
• 28:32we have a primary outcome,
• 28:34at that level anyway.
• 28:36In terms of the antiretroviral
• 28:38regimens, I'll share this because,
• 28:40for folks who do prescribe,
• 28:41you're thinking about kind of
• 28:43the number of times a
• 28:43day someone, takes medication
• 28:46is important to understanding adherence.
• 28:48And so we had four
• 28:49percent that were not on
• 28:51any HIV medication,
• 28:53term I learned from my
• 28:54colleague. They're what are called
• 28:55elite controllers.
• 28:57We had seventy seven people
• 28:59who are on one pill
• 29:00a day.
• 29:01We had three percent that
• 29:02were on injectables. They actually
• 29:04were not included in the
• 29:06daily analysis because their providers
• 29:07are injecting, so there's not
• 29:09much to be adherent to.
• 29:10We had sixteen,
• 29:11percent that were on more
• 29:12than one pill a day.
• 29:15This is another demographic slide
• 29:16just to help you understand,
• 29:18the the population a little
• 29:19bit better.
• 29:21This is a
• 29:23it's very diverse sample. It
• 29:24is a largely
• 29:26insured
• 29:27sample,
• 29:28which is different than in
• 29:29other states. That's because we're
• 29:30a Medicaid expansion state and
• 29:32that influences the care available
• 29:34to folks,
• 29:35including,
• 29:36another benefit is that there's
• 29:38Ryan White funded,
• 29:40drug resistant programs that provide
• 29:42antiretrovirals
• 29:44and medications for all other
• 29:45chronic
• 29:46conditions free of charge to
• 29:47anyone with HIV.
• 29:49Ninety eight percent had a
• 29:50primary care provider.
• 29:52In this study, they did
• 29:53not have to be in
• 29:55a current relationship. They had
• 29:56to have experienced intimate partner
• 29:57violence across their lifetime
• 30:00because what we'll be looking
• 30:01at is whether it matters
• 30:03if the partner violence is
• 30:05current or happened in the
• 30:06past.
• 30:07Right? We know that the
• 30:08effects of intimate partner violence
• 30:09are long lasting,
• 30:11and so we wanna understand
• 30:12that relationship as well.
• 30:14In this study, fifty eight
• 30:15percent were currently in a
• 30:18relationship
• 30:18at baseline.
• 30:19This is the study that
• 30:20bumped that average number of
• 30:22years in a relationship up
• 30:23to ten, almost eleven.
• 30:26Experiences of partner violence, we
• 30:28have lifetime and recents.
• 30:30The lifetime experiences,
• 30:32they had to have experienced
• 30:33IPV to be in the
• 30:34study, so you can see
• 30:35those totals are high.
• 30:38The percent of time that
• 30:40folks spend in an abusive
• 30:42relationship is really depressing. Thirty
• 30:44one percent. And that's really
• 30:46an underestimate.
• 30:47Age was the denominator. Obviously,
• 30:49you're not you don't have
• 30:50the ability to be in
• 30:51an intimate relationship at six
• 30:52years old or nine years
• 30:53old.
• 30:55I guess we could have
• 30:55calculated it across adult life,
• 30:57but nonetheless,
• 30:58spending a fair amount of
• 30:59time in abusive relationships.
• 31:01And of those in,
• 31:03current relationships,
• 31:05you can see that eighty
• 31:06four percent of those relationships
• 31:08are
• 31:09abusive, and many women have
• 31:11experienced partner violence in the
• 31:12past thirty days, psychological, physical,
• 31:15and sexual.
• 31:16So in terms of medication
• 31:18adherence,
• 31:20we were very fortunate. Our
• 31:22methods, our support, the way
• 31:24we train staff, the way
• 31:26that folks
• 31:28collaborate in the interviews with
• 31:29women,
• 31:30and the incentive structure really,
• 31:33contributes to people staying engaged
• 31:35in our studies. And so
• 31:36we had eighty four percent
• 31:38of surveys completed, which is,
• 31:40again, really high.
• 31:42The mean medication adherence was
• 31:44seventy seven percent,
• 31:46and that's calculated out of
• 31:48thirty one days. And, of
• 31:49course, people miss surveys sometimes.
• 31:51If they missed a survey,
• 31:52we included that, as a
• 31:54day that they were nonadherent
• 31:56to be conservative.
• 31:57So it's likely that adherence
• 31:58is even higher.
• 32:00Perfect adherence is at fourteen
• 32:01percent.
• 32:03So this is a study
• 32:04on IPV, and we wanted
• 32:05to look at the difference
• 32:06of IPV on adherent days
• 32:08and not adherent days. This
• 32:10is not sophisticated. There's we
• 32:12didn't even do statistical comparisons.
• 32:14We would have to nest
• 32:15we the statistician would have
• 32:16to nest data. I don't
• 32:17do the statistics.
• 32:19And, we didn't do that
• 32:21here. We're just visually depicting
• 32:22what you can see on,
• 32:24nonadherent days versus adherent days.
• 32:26It's the same in the
• 32:28chart as it is in
• 32:29the table. Some people are
• 32:30visual. And, essentially, you can
• 32:32see on nonadherent days, it's
• 32:34more likely that IPV
• 32:36occurred. Again, these are not
• 32:38statistical comparisons.
• 32:40But you can see
• 32:41that nonadherent days had more
• 32:44minor physical IPV, severe physical
• 32:46IPV,
• 32:47sexual IPV, and psychological
• 32:50IPV
• 32:51than those nonadherent
• 32:53days. When you move to,
• 32:56when you move to look
• 32:58at
• 32:59substance use, because we know
• 33:00there's an intersection of HIV,
• 33:02IPV, and substance use. When
• 33:04you look at substance use,
• 33:05you actually see something very
• 33:07different.
• 33:08Right? You can see that
• 33:09drinking days,
• 33:10marijuana days, misusing prescription drugs,
• 33:13those look about the same.
• 33:15And
• 33:17when for the non adherent
• 33:18days.
• 33:19When you get to illicit
• 33:21drug use, though, it looks
• 33:23the same for drinking
• 33:25marijuana prescriptions,
• 33:27but illicit drug use,
• 33:29again, not statistical comparisons, but
• 33:32we see what looks like
• 33:33it could be a significant
• 33:35difference.
• 33:36Eight, fifteen percent of nonadherent
• 33:39days were characterized by illicit
• 33:40use versus three percent of
• 33:42adherent days.
• 33:43Really important for substance use
• 33:45providers to understand that information.
• 33:49What can you do next
• 33:50with these data? That's day
• 33:51level data.
• 33:52I'm guessing that some of
• 33:54you are wondering, well, what
• 33:55about people who
• 33:57yes. What about people who
• 33:58we can understand person level
• 34:00factors as moderators
• 34:02of those within
• 34:03day
• 34:05relationships. So we can understand
• 34:06how trust in physician
• 34:08might impact some of those
• 34:09day level relationships.
• 34:11Someone's resilience, someone if has
• 34:13a higher level of resilience
• 34:14compared to someone with lower,
• 34:15does that impact medication adherence?
• 34:18Mental health. We know that
• 34:20among women who experience intimate
• 34:22partner violence, about a third
• 34:23typically meet criteria for post
• 34:25traumatic stress disorder,
• 34:27but regardless of if they
• 34:28meet diagnostic criteria,
• 34:30about eighty percent
• 34:32experience functional impairment on a
• 34:35daily basis because of their
• 34:36PTSD symptoms. And we can
• 34:38think about that with other
• 34:39mental health issues.
• 34:41Problematic drinking,
• 34:43you might be thinking you
• 34:44just showed drinking. I showed
• 34:45you whether or not someone
• 34:47drank on a given day.
• 34:48It could be that people
• 34:49who drink problematically,
• 34:52are less adherent at the
• 34:53day level,
• 34:55and that may or may
• 34:55not be related to actual
• 34:57use on a given day,
• 34:58or we could look at
• 34:59participants who have current abusive
• 35:01partners versus those who don't.
• 35:02So lots of ways to
• 35:04explore how personal level factors
• 35:07influence day level associations.
• 35:09So next to talk about,
• 35:12subgroups of women with shared
• 35:13experiences.
• 35:15These data were from a
• 35:16four times, yes, four times
• 35:18a day,
• 35:19daily study for thirty days.
• 35:21I'm only, using data for
• 35:23the first fourteen days. I
• 35:24can explain why at the
• 35:25end if people have questions.
• 35:26But trying to identify subgroups
• 35:28of people based on their
• 35:29shared daily experiences and show
• 35:31you a really cool data
• 35:32visualization
• 35:33program.
• 35:34So the gap is that
• 35:35I showed you day level,
• 35:38information at the beginning, like
• 35:40sixty two percent of no
• 35:41IPV days, then psych abuse,
• 35:43etcetera. But what about at
• 35:44the person level? How do
• 35:45those things group within person?
• 35:47So this is an exploratory
• 35:49late in class analysis,
• 35:51and we use the interval
• 35:53level data, much more accurate
• 35:55level data to inform
• 35:57these analyses. And we found
• 35:59that three subgroups
• 36:01of women with shared patterns
• 36:02of experiences
• 36:03emerged. The yellow orange,
• 36:06bar at the bottom is
• 36:08the largest class,
• 36:09and that's a hundred and
• 36:10eighteen people.
• 36:12And that's essentially low level,
• 36:14physical abuse. So you can
• 36:16see,
• 36:17you can see psychological IPV,
• 36:20minor physical and severe physical.
• 36:21Severe almost never happens in
• 36:23this group. There's some sexual,
• 36:25very low levels of PTSD
• 36:27related distress, some alcohol and
• 36:29drug use, which was an
• 36:30include
• 36:31inclusion criteria for this study.
• 36:33The blue line,
• 36:35is ninety seven women. It's
• 36:37polyvictimization.
• 36:38Basically, just like lots of
• 36:39latent class analysis,
• 36:41that mirrors the patterns of
• 36:43the lower class. It's just
• 36:44more severe. But, again,
• 36:46minor physical abuse tends to
• 36:47happen more often than severe
• 36:49physical abuse, some PTSD related
• 36:51to stress, and substance use.
• 36:53And then we have the
• 36:54salmon class at the top,
• 36:55which has twenty
• 36:57nine folks. It's the smallest
• 36:59class.
• 37:00However,
• 37:01it's what matches the media
• 37:03depictions the closest.
• 37:05Right? But it's the smallest
• 37:06class,
• 37:07and that's where we see
• 37:09high levels of psychological abuse,
• 37:11minor and severe physical abuse
• 37:13which occurred about
• 37:15the same
• 37:16level.
• 37:17So it is happening frequently
• 37:18and it's happening at about
• 37:19the same level. We can
• 37:21see sexual IPV, PTSD related
• 37:23to stress and substance use.
• 37:25So, again, the media depiction
• 37:27matches the smallest class.
• 37:30This is the program that
• 37:31is super cool. It was
• 37:32introduced to me by Seth
• 37:33Posner who is in our
• 37:35department
• 37:36many, many years ago,
• 37:38was not able to use
• 37:39it at that time, but
• 37:41worked
• 37:44with,
• 37:45a fellow here who's now
• 37:46a colleague who taught herself
• 37:48this program.
• 37:50Super cool. You can put
• 37:51in day level or interval
• 37:53level data,
• 37:55and you can see here
• 37:56this is class one, class
• 37:57two, and class three, and
• 37:58it's showing you any IPV
• 38:00occurrence.
• 38:01This column
• 38:02is the yellow is psycho
• 38:04occurrences of psychological IPV. The
• 38:06blue is minor physical. The
• 38:08salmon is severe physical and
• 38:09then sexual. So you can
• 38:10see in class one, it's
• 38:11mostly psychological
• 38:13abuse.
• 38:14The salmon, you could see
• 38:15some severe physical, a tiny
• 38:17bit of sexual.
• 38:18Class two, you see a
• 38:19little bit more minor. In
• 38:21class three, that smallest class,
• 38:23you see a lot of
• 38:24dark blue. You see a
• 38:25lot of salmon.
• 38:26That's,
• 38:28the minor and severe. And
• 38:29you can see this much
• 38:31better visually,
• 38:32with this visualization program. And
• 38:34this is PTSD,
• 38:36any PTSD related to stress.
• 38:38And if you look down
• 38:39the classes, class one, two,
• 38:41and three, you can see
• 38:43that elevated
• 38:44PTSD
• 38:45related to stress,
• 38:47happens more often in those
• 38:48other classes.
• 38:51So moving on to associations
• 38:53of PTSD symptoms at the
• 38:55interval level and alcohol use.
• 38:57So, again, this is that
• 38:58same four times a day
• 39:00study
• 39:01to identify
• 39:02proximal and contingent
• 39:04relationships.
• 39:05So my program of research
• 39:06that centered on substance use
• 39:08really focuses on the self
• 39:09medication or negative reinforcement model
• 39:11of substance use.
• 39:13And that's been applied to
• 39:14explain the PTSD alcohol relationship
• 39:17at the cross sectional level,
• 39:19traditional longitudinal level, and in
• 39:21some micro longitudinal data in
• 39:22other populations
• 39:23like veterans,
• 39:25but not women who experience
• 39:27IPV.
• 39:28There's also an alternative
• 39:29model, which is mute mutual
• 39:31maintenance,
• 39:33whereby
• 39:34it's believed that substance use
• 39:36also exacerbates
• 39:38PTSD symptoms. And this was
• 39:40supported in data among veterans
• 39:42again, but had not been
• 39:43tested with women who experienced
• 39:44IPD.
• 39:45First of all, what's an
• 39:46interval? Well, for this study,
• 39:48we had four different call
• 39:50in times. They were allowed
• 39:51to call in during this
• 39:53reporting window and then report
• 39:55back
• 39:56to a previous period. When
• 39:57we piece those four previous
• 40:00periods together,
• 40:01we get twenty four hours.
• 40:02So it covers an entire
• 40:03day. It's called coverage based
• 40:05sampling.
• 40:07The data analysis, the data
• 40:09were collected at the interval
• 40:10level. They're nested in the
• 40:11day level and then at
• 40:12the person level.
• 40:15And this, again, is that
• 40:17hypothetical
• 40:18graph at the beginning, which
• 40:20we actually did end up
• 40:21testing with real data.
• 40:24The purpose was to examine
• 40:25the associations among PTSD symptoms
• 40:27severity and drinking. And so
• 40:29we did that through looking
• 40:30at concurrent models. What's a
• 40:31concurrent model?
• 40:34So basically, we're trying to
• 40:35understand our within day fluctuations
• 40:38from the day's overall mean
• 40:40levels of PTSD symptom severity
• 40:42related to average level of
• 40:44drinking during an interval. Basically,
• 40:46when you're looking within an
• 40:47interval, you're doing cross sectional
• 40:49data analysis at a really
• 40:51precise period of
• 40:53time. And so that's this
• 40:54depiction here. Meaning, when PTSD
• 40:56symptoms are elevated during the
• 40:58late afternoon, for example, is
• 41:00drinking also elevated during that
• 41:02time point?
• 41:04We also examine time lagged
• 41:06models,
• 41:07are within day relative levels
• 41:08of PTSD symptom severity related
• 41:11to changes
• 41:12in drinking in a subsequent
• 41:14interval. So, again, if you
• 41:16are looking at PTSD symptom
• 41:17severity in interval three, is
• 41:19that associated with drinking in
• 41:20interval four?
• 41:22The short answer is yes
• 41:24and yes.
• 41:26There without the statistics embedded,
• 41:28there were both concurrent
• 41:30and time lagged relationships where
• 41:32fluctuations in PTSD symptoms
• 41:35influenced
• 41:36alcohol use. We also tested
• 41:38differences by race,
• 41:39and ethnicity.
• 41:41As moderators, there were no
• 41:42differences. And the papers noted
• 41:44here because we did look
• 41:45at other moderating effects,
• 41:46like by whether or not
• 41:47someone met criteria for PTSD
• 41:49and there were actually moderating
• 41:51effects.
• 41:52So in terms of assessment
• 41:54and methods
• 41:55benefits, so we are we're
• 41:56talking about finding out things
• 41:58about populations. How are things
• 41:59related within an individual?
• 42:02But you can also use
• 42:03this method to better understand
• 42:05what we're doing when we're
• 42:07assessing
• 42:08behaviors more broadly in other
• 42:10studies.
• 42:11So this was a three
• 42:12times a day study for
• 42:13thirty days. This was actually,
• 42:15conducted by a past fellow,
• 42:17now colleague
• 42:19of mine.
• 42:20This was her career development
• 42:21award,
• 42:22and it's focused on,
• 42:25IPV, emotion regulation, PTSD, and
• 42:27substance use.
• 42:28Here, we're talking about how
• 42:30assessment information can inform future
• 42:32work.
• 42:34So this is looking at
• 42:35the concordance
• 42:36of PTSD
• 42:38symptoms assessed
• 42:39via
• 42:40retrospective reports, so at a
• 42:42follow-up interview? And how does
• 42:43that map on to what's
• 42:45collected at the day level?
• 42:48Are they the same? Right?
• 42:49Because a lot of us
• 42:50do traditional longitudinal designs, cross
• 42:52sectional designs. When we ask
• 42:53someone, tell me about your
• 42:55symptoms,
• 42:55does it match what they
• 42:57would tell us at the
• 42:58day level?
• 43:01So there's a way to
• 43:02conceptualize this. It's Schueller's category.
• 43:05The reference is at the
• 43:06bottom for the Schueller category,
• 43:07but, PTSD symptom patterns can
• 43:09be conceptualized
• 43:10in six different ways.
• 43:12You can sum across all
• 43:13of those,
• 43:15interval surveys to get a
• 43:16mean of all PTSD symptoms.
• 43:18You can look at peak
• 43:20symptoms like what is the
• 43:21worst day of the one
• 43:23report among all of the
• 43:25completed surveys where the sum
• 43:26of PTSD symptoms were the
• 43:28highest.
• 43:29What's the mean of PTSD
• 43:31symptoms that would be average
• 43:32daily peak. There's the worst
• 43:34day of symptoms. There's,
• 43:35primacy, which is the first
• 43:37week they reported symptoms and
• 43:38the last week they reported
• 43:39symptoms.
• 43:40Right? What did we find?
• 43:41Are they concordant?
• 43:43The short answer is that
• 43:44they're not.
• 43:45They are not concordant. Retrospective
• 43:47reports differ
• 43:49and they most closely
• 43:51reflect that peak survey over
• 43:54thirty days. That one where
• 43:55this that one survey where
• 43:57the symptoms were the highest.
• 43:59So when we gather information
• 44:00retrospectively
• 44:01for PTSD symptoms, at least
• 44:03in this population, what we're
• 44:05seeing is it reflects that
• 44:06worst survey of all of
• 44:08the past thirty days.
• 44:09Next, it's followed by the
• 44:11worst day symptoms,
• 44:12but it substantially
• 44:14differed from average
• 44:15symptoms across,
• 44:17which is kind of what
• 44:18some of us would have
• 44:19expected. We would have expect,
• 44:20oh, people just give us
• 44:21an average when they retrospectively
• 44:23report. Apparently, they do not,
• 44:25and this method helps us
• 44:26figure that out.
• 44:28There's another assessment
• 44:31benefit. Right now so, doctor
• 44:33O'Malley mentioned that I'm starting
• 44:34to do intervention research, and
• 44:36we developed,
• 44:37this really
• 44:39cool single session technology
• 44:42facilitated intervention that aims to
• 44:44promote hope among women who
• 44:45experience intimate partner violence. And
• 44:48instead of doing traditional pre
• 44:50post,
• 44:52we did once a day
• 44:53for fourteen days pre and
• 44:55post to better represent experiences.
• 44:58So there's a baseline interview.
• 45:00They do fourteen days of
• 45:02daily data collection.
• 45:04They complete the single session
• 45:06intervention. We do a process
• 45:07evaluation
• 45:08survey right after, and then
• 45:09they do fourteen days of
• 45:11daily reports of hope after.
• 45:13In a traditional study, we
• 45:15probably would do baseline,
• 45:17end of some longer period,
• 45:19and maybe even an immediate.
• 45:22But instead here,
• 45:24we're
• 45:24looking at the daily reports,
• 45:26and that still gives us
• 45:27the opportunity to aggregate them.
• 45:29We can still aggregate over
• 45:30those fourteen days, perhaps just
• 45:31getting more,
• 45:32reliable information,
• 45:34reduces recall bias and improves
• 45:36ecological validity because they're reporting
• 45:38on a daily basis.
• 45:40But it probably misses variability.
• 45:43And we can look at
• 45:44more precisely
• 45:45at the day level. And
• 45:47to sell myself out, to
• 45:48be honest, we know the
• 45:49day level is actually not
• 45:50specific enough anyway. But when
• 45:52you are limited by resources
• 45:53and a funder,
• 45:55once a day was the
• 45:56best that we could do.
• 45:57We know that we need
• 45:58to look at this more
• 45:59than once a day. So
• 46:02what do we see if
• 46:03we look at the person
• 46:04level, if we aggregate to
• 46:06those fourteen days on one
• 46:07of the daily questions that
• 46:08gets at Hope, one or
• 46:10two. So yesterday, I had
• 46:12inner strength. That's the aggregate
• 46:14pre and the aggregate post.
• 46:16Yesterday, I believe today, would
• 46:18be better. Aggregate pre and
• 46:19aggregate post. So I was
• 46:21super sad when I saw
• 46:23this, and I was like,
• 46:24this can't be. Like, there's
• 46:25nothing there.
• 46:27And this is not testing
• 46:29for difference, but is there
• 46:30really even a need?
• 46:32They look nearly identical. You
• 46:34can see that there's more
• 46:36variability post SSI, and there
• 46:37could be something to that
• 46:39as I've been introduced to
• 46:40the broaden and build theory.
• 46:43But nonetheless
• 46:44was disappointing
• 46:45and was actually quite stressful
• 46:47because I had just been,
• 46:48funded by the Office on
• 46:50Violence Against Women
• 46:52for another micro longitudinal study
• 46:54identifying broadly
• 46:56what factors predict hope at
• 46:58the day level. So if
• 46:59there's no variability in hope,
• 47:00I cannot ethically take money
• 47:02to do a day level
• 47:03study of hope.
• 47:05So I
• 47:06I was a little nervous,
• 47:08and,
• 47:09asked if we could take
• 47:10a look at that day
• 47:11level variability. And I'm not
• 47:13showing you this for any
• 47:15other reason than to say
• 47:16there's day level variability in
• 47:17hope. We know there is.
• 47:19It's there.
• 47:20But if we take this
• 47:21one step further, what does
• 47:23daily data allow us to
• 47:24do? I don't actually have
• 47:26any cool findings to show
• 47:27you. This study is an
• 47:28active data collection,
• 47:30right now, and it's just
• 47:31a pilot study. But what
• 47:33can you understand at the
• 47:34day level when you plot
• 47:35this information out?
• 47:37So this is person four
• 47:38zero zero seven o. These
• 47:40are crude,
• 47:41figures.
• 47:43This person completed the intervention
• 47:44on day sixteen. So if
• 47:46you imposed a line right
• 47:47here to look at completion,
• 47:49you would see that after
• 47:50the intervention, there's less variability
• 47:52on how much inner strength
• 47:54this person had.
• 47:55And in terms of when
• 47:57they talked,
• 47:58down to themselves,
• 48:01they don't really seem
• 48:03to have a different experience
• 48:04of their inner strength. Their
• 48:05negative self talk doesn't seem
• 48:07seem to be influencing
• 48:09their assessment of their inner
• 48:11strength. This top one, again,
• 48:13it's the same daily assessment
• 48:15of their inner strength. The
• 48:17green here is whether their
• 48:18partner was psychologically
• 48:20abusive, and this person doesn't
• 48:21seem to be that influenced
• 48:23by their partner's psychological abuse.
• 48:25They stay pretty
• 48:27fairly even. And the same
• 48:29is true,
• 48:30in terms of whether or
• 48:31not they have social support.
• 48:33Doesn't really seem to be
• 48:34moving the needle a lot
• 48:35on whether or not they
• 48:36felt they had inner strength.
• 48:38So this is one type
• 48:39of person.
• 48:41This is a different type
• 48:43of person. This person completed
• 48:45the intervention on day fourteen.
• 48:47You could see there's a
• 48:48little bit less variability in
• 48:50inner strength toward the end.
• 48:52There seems to be less
• 48:53of an impact of them
• 48:55talking down to themselves. The
• 48:56green line is them talking
• 48:57down to themselves.
• 49:00This is their partner's psych
• 49:01abuse. They don't have a
• 49:02lot of psych abuse.
• 49:04You can see their social
• 49:05support.
• 49:06This individual has experiences
• 49:09of social support,
• 49:11but you're not you could
• 49:12see something different than that
• 49:13first person who stayed very
• 49:15even and gives us ideas
• 49:16about how to think about
• 49:18the influence of factors on
• 49:19hope. And then here's another
• 49:21person. There's less variability in
• 49:23inner strength,
• 49:24and we can see differences,
• 49:26in how
• 49:28their negative self talk, psychological
• 49:30abuse,
• 49:31and social support might or
• 49:33might not influence.
• 49:34So again, this one was
• 49:35just to show you, like,
• 49:36when you look at the
• 49:37person level, it also gives
• 49:38you ideas about how to
• 49:40better understand,
• 49:41the relationships at the day
• 49:42level.
• 49:44So in brief or in
• 49:45sum,
• 49:46key take home findings. First
• 49:48and foremost, it's important to
• 49:50me to communicate that heterogeneity
• 49:52exists
• 49:53among women who experience intimate
• 49:55partner violence. It's certainly not
• 49:56just the media depiction, not
• 49:58a one size fits all.
• 50:00That psychological
• 50:01IPV deserves greater attention. We
• 50:03know how it influences
• 50:05day level behaviors. We also,
• 50:06I, a long time ago,
• 50:07published a paper that showed
• 50:09it was a stronger predictor
• 50:10of PTSD
• 50:11in this population than physical
• 50:13and sexual abuse. We need
• 50:14to attend to it more
• 50:15than research
• 50:16in research.
• 50:18We need to think about
• 50:19safety strategies beyond calling the
• 50:21police and using domestic violence
• 50:22services because that's not what
• 50:24folks typically wanna do on
• 50:25a daily basis.
• 50:27We do know that IPV
• 50:28at the day level seems
• 50:29to influence medication adherence.
• 50:32We'll test this again when
• 50:33we can get into data
• 50:34analysis.
• 50:35Negative reinforcement
• 50:37models do fit for this
• 50:38population.
• 50:39Mutual maintenance does not. It
• 50:41does not seem that for
• 50:42alcohol at least that that
• 50:43exacerbates PTSD symptoms
• 50:46and that retrospective reports of
• 50:48PTSD symptoms really represent peak.
• 50:51Right?
• 50:53Also, take home that's important
• 50:54to me is that survivor
• 50:56defined practices where clients have
• 50:58voice and choice,
• 51:01empower people and do not
• 51:02seem to increase risk. In
• 51:03other presentations, I've actually talked
• 51:05about our risk protocol for
• 51:06the firearm study. I was
• 51:08not waiting until the IRB
• 51:09told me no thank you
• 51:10after I just got almost
• 51:11three million dollars.
• 51:13So we prepared what we
• 51:15think is an amazing protocol
• 51:16soon, to be published,
• 51:18about risk. Women can make
• 51:20their own choices about what
• 51:21they wanna do, and how
• 51:23to stay safe.
• 51:25And thankfully, that means about
• 51:26when they stay in a
• 51:27relationship and when they leave
• 51:28as well, and we know
• 51:30that from our practical work.
• 51:31Also, that micro longitudinal designs
• 51:33are really cool. It is
• 51:35nearly impossible for me to
• 51:36conceptualize
• 51:37a study where a micro
• 51:39longitudinal design would not be
• 51:41better than a traditional retrospective,
• 51:44design or traditional longitudinal
• 51:46design. The challenge is finding
• 51:47the resources to pay for
• 51:48them because they are more
• 51:50costly.
• 51:51Many,
• 51:52of
• 51:53the micro longitudinal studies that
• 51:55are done really have strong
• 51:56clinical implications. There is a
• 51:58growing body of research
• 51:59on suicidal ideation and behavior,
• 52:01and you can imagine how
• 52:03pinpointing
• 52:04those factors that increase risk
• 52:05are really important to informing
• 52:08interventions,
• 52:09including just in time interventions.
• 52:11So I will wrap up
• 52:13there and say thank you
• 52:14for your time and attention,
• 52:15and I think we have
• 52:16time for questions.