2016 Public Health Grand Rounds 02/05

PUBLIC HEALTH GRAND ROUNDS Linking Research to Community Health Improvement Jointly sponsored by the Department of Public Health Sciences & URMC Center for Community Health

(speaks off microphone)
- All right everybody, I think we're going
to go ahead and get started.
Welcome everyone to
Public Health Grand Rounds 2016, I guess.
This is the first iteration for the Spring series.
We meet on Fridays at noon every first and third Friday
for February, March and April.
So, welcome.
The schedule for Public Health Grand Rounds
is on the Public Health Sciences website
as well as the Center for Community Health's website.
And all the recordings of prior lectures,
or many of them, are up there as well.
So enjoy. Thank you for coming and please continue.
Public Health Grand Rounds is a collaborative effort
between the Public Health Sciences Department
and the Center for Community Health
to present opportunities of community
working with research to improve the health
of Rochester and communities.
So hopefully you will learn a lot today
and continue coming.
We're thrilled to have Doctor Bob Gramling
here with us today,
especially because his days are numbered, sadly,
here at the University
where he's moving on to bigger and better things.
So we're really excited to squeeze you into our series.
I will read you Doctor Gramling's bio very seriously.
(laughing)
Doctor Robert Gramling is a palliative care
physician scientist at the University of Rochester,
where he is Associate Professor of Family Medicine,
Public Health Sciences and Nursing.
Doctor Gramling's research focuses on
decision making conversations, which particular
attention to communication about prognosis,
trade off and uncertainty.
He has authored more than 70 publications
and received research funding from the
National Institutes of Health,
the National Palliative Care Research Center,
the Greenwall Bioethics Foundation
and the American Cancer Society.
He is currently principal investigator
of the Palliative Care Communication Research Initiative,
a multi-site cohort study funded by
the American Cancer Society to understand
the features of palliative care consultation,
that promote patient centered outcomes.
Please join me in welcoming Doctor Robert Gramling.
- [Robert] Thanks, Teresa.
(audience applauds)
Thank you very much.
This is great.
Can people hear me with the, is this mic
at the right level?
Okay, good, good, good, good.
So I have not been in this room before.
This is beautiful.
I'm glad I'm here now, and as Teresa mentioned,
I am going to be leaving in May,
but I'll be around for the next few months.
So if anything I'm talking about today
sparks your interest and you want to
talk more about it, please do reach out,
because I will be here on campus.
So I have the luxury today of talking about
a study that's near and and dear to my heart,
both as a physician, as a scientist,
and as a patient and family member.
But this is really the crew who should be
up here with me, and if we could figure out
a way to do this, I would.
We have a fabulous research team.
This type of work is challenging.
It's rewarding.
And it takes a group of people,
as most endeavors of value do.
And so I'm not going to read everybody's name,
but I wanted to put them up there.
And those of you on my team who didn't send me a picture,
next time I'm just going to go pull one off the web,
so it's up to you to send me one that you want.
I have no financial or ethical conflicts to report.
I have plenty of conflicts I'm happy
to talk with you about if you'd like otherwise,
but nothing relevant to this.
So for the next 45 minutes or so,
I wanted to share with you some of the rationale,
particularly the public health rationale,
for why study communication in serious illness.
Describe some of the relevant methods from
our team's current cohort study.
Share particularly, because I'd like to promote,
any of you who are interested in doing this,
I'd like to offer that it is fun, doable and important.
So I'd like to share some of our
recruitment and enrollment strategies,
some of the data collection strategies
for the hospital setting,
and some of our thinking about how we analyze
such things as conversations.
And also I'd like to share some of
our preliminary findings.
Enrollment is still going on,
so these are, indeed, preliminary.
So, this is a good news or a bad news graph,
depending on if there's...
Raise your hand if you're a scientist
in here of some kind.
Any scientists?
It's Public Health Grand Rounds.
Anyone want to admit to being a scientist?
Anyone know a scientist?
Okay, fantastic.
So if we extrapolate, pretty soon we'll be living forever.
But at least the good news is we're living longer.
We're not living more simply.
Does anyone know what this is?
Anyone want to take a gander at what this picture is?
And also, I have really bad vision,
so it's okay for you to go like this in the back
or just to jump in.
That's perfectly fine.
So I'm not seeing any takers.
These are crayons.
So in the beginning of the century,
we'd open up our crayon box, and we'd have 8 choices.
And we could decide how to draw our picture
with some relative contrast.
As our century has progressed,
we now open our crayon box,
we've got more than 500 colors that we choose from,
many of which are almost indistinguishable.
And so this is very...
The reason I use this picture is this is
a reasonable metaphor or analogy
to what's happening in health care decision making
with respect to treatment choices that people face
as the years go on
and as medical technology advances.
It's not beginning similar,
and to some degree the colors are starting to blend.
So the question then becomes is how well do we do this?
And this is a palliative care talk,
and I'm interested in people who are very ill
and for whom death is something that is
a plausible probability.
And so the question I guess I'd have for you
is how well do you think our medical culture,
our institutional systems and our healthcare industry
addresses decision making for people for whom
death may be near?
Anyone want to raise their hand if they think
it's done really well?
Want to raise you hand if you think it's done really poorly?
Want to raise your hand if you don't have an idea?
Okay, raise your hand if you didn't raise your hand yet.
(audience laughter) There we go, excellent.
Good, we got another answer I got to think of next time
before we do it.
So, indeed, our system is not designed
to treat, to help people navigate
these color choices as death approaches.
Our system is designed almost exclusively
to treat and cure disease,
which is a good thing, right?
Another current trend, if has anyone read
Atul Gawande's book is a social commentary
on the emerging, growing evidence and passion
by which we're all recognizing that these systems
were not designed to take care of people
who are dying and who are suffering.
And Atul Gawande says, maybe more strongly,
is that "Our decision making in medicine
has failed so spectacularly that we have reached
the point of actively inflicting harm on patients."
Active isn't malice, right?
We're not going around saying, gosh we gotta
try and hurt people.
Point is, we're not very good at navigating that.
So what do we do about it?
So the Institute of Medicine has released
a treatise on dying in America,
honoring patients preferences.
And trying to summarize what we know,
what we don't know and what are expert opinions
might tell us about, so what do we do about this?
And one of the key findings from the
Institute of Medicine's report is that
frequent clinician-patient conversations
about end of life care, values, goals and preferences
are necessary to avoid unwanted treatment.
However most people, particularly younger,
poorer, minority and less educated individuals,
do not have these conversations.
Clinicians need to initiate conversations
about end of life care choices
and work to ensure that patient and family
decision making is based on adequate
information and understanding.
So that's a little bit of what to do about it.
And the challenge that the Institute of Medicine,
National Priorities Partnerships and others
place on the next 100 years of medicine
is to figure out how to do this.
But this really, this is Public Health Grand Rounds, right?
This is really a medical issue?
Why are we talking about this from
a public health point of view?
Well, two and half million people die every year
in this country and that number's going up.
This is actually 2013.
Does anybody in here know anybody's who's died?
And I don't need to, I mean, this can be a really
heartfelt experience of knowing somebody or
experiencing this process of end life.
Is anybody here human?
Anybody here going to die?
Okay, we don't have to think about it all the time.
In fact, it's probably very productive
not to spend our time thinking about it all the time.
But this is the one event we are all going to experience.
There's no other event that we are all going to experience.
This is going to happen.
The good news, for the public health interventionists
here about thinking how we're going to address this,
is that there's plenty of opportunity.
If we're trying to promote clinician-patient
family conversations that are meaningful and valuable,
well, in the last 6 months of life, according
to the Dartmouth Atlas of Medicare Beneficiaries,
there are at least 10 hospital days on average
in the last 6 months that 42% of patients see
more than 10 physicians in the last 6 months of their life.
And 25% die in the hospital.
So regardless of whether this is good news
about people say these are the things I would want or not,
there are a lot of contexts with the medical system
during this really crucial time in people's lives.
So where might we think about trying to learn
what it means to have a valuable conversation
because although we may recommend this
and we say logically this makes a lot of sense,
we know very little scientifically
about what good looks like.
What is valuable?
When is it valuable?
How much of it is valuable?
And when is it invaluable?
The Dying in America recommends as one of the key points
is that palliative care is associated with
higher quality of life, including better
understanding in communication.
I'm not going to read the rest of this.
But a lot of this message is that
the approach that underlies a lot of
palliative care consultation in the specialty
context right now seems to be having some benefit
navigating these last, very complicated
months of people's lives.
So I'm going to read this because I imagine
there's some people in here who don't know
what palliative care is.
And we as a discipline have had a hard time
getting the sound byte about what it is we do
in a way that really captures the breadth
and the focus for different audiences.
So I'm going to read a relatively long definition.
This is from the National Palliative Care Research Center.
"Palliative care focuses on relieving suffering
and achieving the best possible quality of life
for patients and their family caregivers.
It involves symptom assessment and treatment,
aid with decision making and establishing
goals of care.
It is offered simultaneously with life prolonging
and curative therapies for persons living with
serious complex and eventually terminal illness."
Palliative care is a new discipline in medicine.
It's not new philosophically, but it's a new medical
sub-specialty or specialty as of, I believe, 2008 or 2010.
Tim Robb, was it 2010, 2008?
Somewhere in that ballpark.
And because of that,
we saw a big jump in palliative care
programs around the country.
This is a graph of hospitals of greater than 50 beds
of the prevalence of palliative care programs.
And what we've seen over the last 15 years or so
is a rapid increase in palliative care, to the point now
where about 7 out of 10 hospitals of greater than 50 beds
have a palliative care specialty consult service.
So the goals of this study, in this context,
is that we wanted to understand and want to understand
communication as it naturally occurs
in palliative care consultations,
to begin to identify patterns among the features
of those conversations that predict good outcomes.
We'll call them patient centered outcomes.
I'll be back to that in a minute.
The purpose of doing this is ultimately to begin to inform
what are best practices, what are the active ingredients,
how might we consider scaling that up, reaching other
hard to reach populations, informing our specialist
and generalist colleagues in palliative care practices.
This is all next steps for our discipline.
When I say patient centered, this gets used a lot,
so I am going to define what I mean,
at least for the purposes of this lecture
and for this project.
Patient centered, what I mean are the clinical processes
that promote persons feeling heard and understood
and valued, that lead to treatment decisions that
honor the person's values, preferences
and clinical realities, and that address suffering.
Sounds squishy, huh?
Well some of this is going to be squishy,
but the measurement is where we're going to talk about
how we do this, and how we think about what concepts
such as patient centered means in context.
Let me also offer, I mentioned that
I had trouble seeing people in the back row,
but I also want to welcome you, if people have questions
as I'm going here, please do ask.
I'd rather that you don't get derailed early and not be able
to follow what the rest of the talk is about.
So please do interrupt me, I can get back on track.
Okay.
This study is a cohort study, it involves patients
who have been diagnosed with an advanced solid cancer,
advance being operationalized by being metastatic
beyond local spread, that they have been referred
for palliative care consultation, that they are hospitalized
and that they speak English and they are adults.
That's the eligible population for our study.
On the patient's side.
On the clinician's side,
it's anyone who is providing palliative care services
at our study hospitals, as far as the consult team.
It's observational, we're not intervening,
or we're doing our best not to intervene.
When I say that I mean that any time
when we design our methods we're trying to have
as small a footprint as possible because we want to try
to understand the natural processes.
But we're also measuring things that can be arguably
be some forms of interventions,
and I welcome your thoughts on those things.
It starts with enrollment in the study,
at the time of referral to palliative care in the hospital,
we observe and record up to three palliative care visits
that unfold during that hospitalization.
We interview patients or their health care proxies
and clinicians, both before and after
the first recorded consultation.
Clinicians are actually interviewed at the beginning
of the study as far as their baseline description,
and then after each patient who is enrolled,
have a brief interview following one of
the recorded conversations.
We then follow for medical record and administrative record
for six months or the rest of this persons life,
whichever is longer... shorter, sorry.
For our outcomes they're matched to our definition
of patient centered.
We measure quality of life,
we measure perceived communication quality
as a metric of how well people feel heard and understood
by the communicants in their environment.
Whether they have received medical treatments
that are concordant as best as we can define,
by established preferences that people report
at their baseline.
The cohort data that we collect,
comes in from multiple sources,
we are looking at patient interview data,
and I'm going to let you take a look at some of the topics
that we're asking people.
The clinician interview, the observer field note,
the medical record, and the recording itself
of the conversations.
We have many sources of data for this cohort study.
Those of you in here who have taken part in collecting
the data, you're not going to get off easy,
I'm going to introduce you later.
If you need to run to the bathroom, guys,
this is the time to get out of here.
So enrollment, so how has it been going?
We started this in January of 2014,
our goals were to enroll 40 palliative care clinicians,
physicians, nurses, nurse practitioners, social workers,
chaplains with a primary focus of our
palliative care team definitions.
250 patients or their proxies and 2 to 3 recruitment sites.
Here we are, we're going to finish enrollment
in July of 2016.
For the two to three sites, we have our site here at Strong
and Highland as well,
and at University of California San Francisco
at the Moffitt/Long hospital where they have
an inpatient palliative care team.
So far we have 52 palliative care clinicians enrolled.
The structures of our teams are little bit different here
than in San Francisco.
In San Francisco they have more under specialists,
and chaplains and social workers that actually round
with the team and see the patients who are referred
and participate in conversations.
Most of those participants in the team are coming
from San Francisco.
Among those who are invited to participate, 98% agree.
Patients, we have 230 enrolled as of yesterday.
And 10% by proxy, which means that they didn't have
sufficient capacity to consent to research themselves,
but through the participation of their identified
healthcare proxy, they were able to participate.
The good news for this is that among the people
who are informed about a study going on,
are you interested in learning more?
80% of them said "Yes, I am interested in learning more."
And among the people who were told specifically
about this study, and what it entails, 60% of them
wished to participate.
I don't know if anybody has done hospital based
types of research before but these are good numbers.
So... who did we enroll?
The snapshot of our enrolled participant
looks like the referral base from which they come.
Which was our primary focus of recruitment,
was to try not to over select for just those people
who felt like they had lots of energy and vigor
to be able to participate in the study,
but rather we wanted to understand the experience
of everybody who is referred to palliative care
within some possible bounds.
So women, about half our participants are women,
about 15% self identify as black or African American,
which is 2 or 3% above our endemic referral rate,
financially insecure, about 60% of people say they don't
always have enough money to pay for the basics,
housing, repairs to the home, food, and medicines.
As we'll learn, as the science is unfolding,
this financial insecurity is an important player in
your end of life type decisions.
We're not going to get at that today,
but it is in parallel emerging in the literature,
something that we need to attend better to.
About 30% have completed a four year college degree.
About 20% are older than age 75, and about 5 or 6%
are younger than age 50.
The bulk of the patients who we see, and who are enrolled
in this study, are between the ages of 50 and 75.
Primary cancer type... I say primary,
many people have multiple cancers,
so when we say the primary cancer type,
that's the one that's really the overwhelming focus
of their treatment or the one that is likely
to be the cause of their death.
G.I. cancers, primarily being esophagus, stomach, pancreas,
gallbladder, liver and some colorectal cancers as well.
Lung cancer, breast and prostate, unknown,
meaning that we know there is cancer but we don't know
exactly where it came from, it's been biopsy proven,
but the source is not known.
Kidney, cervix, ovary and other...
Other's typically sarcoma, melanoma, head and neck cancers,
and a variety of other relative and thus
common types of cancer.
In the baseline survey we ask participants
the global measure from The McGill Quality of Life
Inventory, considering all parts of your life,
physical, emotional, social, spiritual, and financial.
Over the past few days, how would you rate
the quality of your life?
And it's one a scale from very bad to excellent, 0-10.
And what we see from our participants,
is a range on this scale,
they're using the full scale to respond,
you're seeing a larger proportion answering 5
as you might expect, for anybody who does survey based
research, when you give a middle answer,
you're going to get more of that.
The striking thing for us is that there really is a
breadth of experiences people are rating
their quality of life.
The participants in our study,
this is a survival curve for people
who have enrolled in our project,
and from what you'll see of this,
within about six weeks or so,
or actually less than six weeks,
half the people who are enrolled
in the study will have died.
And only about 10% will be alive after six months.
So this is a precious and very confusing
and often times terrifying times of peoples lives,
when they are seeing palliative care.
It's also joyous, there's plenty of joy
that happens as well.
But it's intense.
Intense humanity happening and this is what we see.
The question that we are often times asked,
or I'm asked frequently is...
this is fan... you know, people are really excited.
Wow this is really good you can do this,
this is really important, we want to understand this.
And it didn't occur to me that we were going to be able
to enroll people and study this process at
such a precious time of peoples lives
in such a chaotic environment as a hospital,
so a lot of the questions I get is... how do you do it?
My first answer is, you hire the right people.
Can I please, my friends on my team, just raise your hand
if you were involved in enrollment, data collection,
thinking about it, anything from the research
side of things, please raise your hand.
Look at that.
This is an amazing group of people,
who have the ability to be present,
to help advise our clinical colleagues about
how to participate and how not to get in the way
as best as humanly possible, this is an amazing skill set.
And much of it is from the school of nursing
that people on our... I have the luxury of being able
to have school of nursing graduate students
who are returning for their PhD after many years
of experience in hospitals and in health care,
so they are extraordinarily experienced people,
whom I am extraordinarily under paying for how good
they are at this (audience laughs)
and thankfully the Dean of Nursing covers me
because that's a set fee, but either way
it's wonderful people.
The other thing, and those of you, if you're a participant,
you do not need to say that for confidentiality reasons,
but if there's anybody here on the palliative care team
who wants to acknowledge, raise your hand as being
part of the team, rather than...
raise your hand for a second.
This cannot be done without an extraordinary
group of clinicians who are confident in their abilities,
who are open to being studied and are passionate
about wanting to know how to do this.
This could not be done without a team of people
who are welcoming us into the sandbox to do this.
I don't have patients and families here
to say the same thing, but they as well are
providing us entry into this time.
I cannot say much about that but I will stop at this point.
From a structural point of view,
given that you have the right people,
you have clinicians who are invested
and wanting to understand,
patients and families who are interested
and wanting to understand this.
Now become some kind of structure of how you design
the data collection and enrollment.
This is the snapshot in time that is
the most intense in our study.
It is often times a one to two day period of time
in which is 80 to 90% of our data is collected.
Patients are referred to palliative care,
they're interviewed, conversations recorded,
the clinicians are interviewed right afterwards,
the following morning the patient to health care proxy
is interviewed and the next conversation is recorded.
This is intense.
So how do we think about doing this
in a way that is sustainable?
Two ways.
One is that on the interviews,
what you're asking of patients and health care proxies,
is that they're not answering a lot of questions.
They're answering only 20 questions total.
What we worked really hard to do, and painfully sometimes,
is to identify... what can we measure
from the administrator or medical record
that is a reasonable proxy for patient experience,
and what can we not?
And if we couldn't answer it using medical record data,
we put it on our list of what needed to be asked directly,
and then we had to really struggle with what do we let go
and what do we keep, what was essential for
our study questions, what would be lovely if we could have
and we got it down to 20.
The questions that we ask are also designed in the way
that the Dartmouth Coop charts have designed,
measures for use in the clinical setting.
Good size font, relatively simple questions.
You'll see this is is not a,
this is multiple barreled question,
we can describe a little bit of why that is,
but with outcome measures that are...
this is actually what they use in the Coop charts.
So we design them off of other metrics
that are used for the clinical setting.
We also in the information packet,
when people are first approached,
are you interested in hearing about a study, yes,
we give them some information about it
and we make a brochure,
and in the brochure we're very clear about
what we're asking of them.
And I think.. where's Sally.
My friend Sally was key in structuring this,
letting us know, what does this mean for you?
People can really quickly look at it,
what am I committing here beyond philosophically of this,
and that 20 questions,
man we could not break that 20 barrier.
When we got 21 or 22, participation rates in our pilot study
would start to go down so we found that that 20 questions
has some symbolic meaning that we had to really honor.
So we stuck at 20 questions.
Whether it's the child game we play or whether
it ended in a 0, I don't know.
But this was a real threshold.
How much time it takes, what they're going to do,
then that we would call them, not always,
but up to two times in the next six months
if we could not find out where people
had been administrated to the hospital and other things.
The other thing, having communication
makes a big difference.
Patients and families have experienced poor communication
and really value anyone who seems to be study,
what does good communication look like?
The other thing with the clinicians,
so we're asking them after they've just come out of one room
and are running to the next three,
something about that conversation.
We ask only three questions.
And three that take some judgement.
Because if we're going to bother to ask clinicians
about something, it has to be something that we couldn't
otherwise get a proxy measure for reliably.
And so we ask, one of the hardest ones, is prognosis.
A lot of clinicians don't like answering questions
about prognosis.
Either for superstitious reasons, if you say,
if I speak to death then it might happen,
or the perceived challenges of being able
to estimate prognosis, which are really no different
than any other population based science
when you're trying to interpret for an individual,
but it feels a heck of a lot more onerous
to try to estimate prognosis with people.
Often times this is asked and would you be surprised
if this person died within the next year?
It's a little bit easier thing to commit to
or not to commit to, and we tried that with palliative care
team and the palliative care team
wasn't surprised by a heck of a lot.
(audience laughs)
We didn't necessarily find that was a very valuable measure,
and we did find that our palliative care colleagues are
willing to take a guess at what we think
the best estimate of prognosis is,
within structured time periods,
these are time periods that we often think about
in palliative care because they have implications
for what kind of resources might be available to people
if we think that's how long they're going to live.
In this process and in the very beginning here,
before you get that first interview, this all has to happen.
A call comes to the palliative care team,
the palliative care team has to identify...
let our palliative care research team know
that a potentially eligible participant is on the radar.
Those of you who have participated in that process,
please raise your hand.
There's only one I've seen here, two,
these guys are fabulous because carrying around that pager
all day today, all day any time, can be...
can be lots of fun.
And it can be crazy making.
So when you get pages and you think, gosh,
maybe I just won't tell the palliative care research team
about this one because this day is really crazy
and I don't want to... I'm sure that happens, I'm sure
it would happen to me, but it doesn't happen very often.
You guys really go out of your way to know when
potentially eligible people are there
and what has to happen then is our team has to respond.
We have to get to the floor where the patient is,
we have to have someone on the clinical team
make first contact.
Clinical team can be one of the nurses or physicians
or staff caring for the person on the referring side,
or on the palliative care side.
If we are just making that first approach,
saying there's a study going on,
are you interested in hearing more?
Because the University of Rochester,
every institution has to define this
a little bit differently, is a private space
is a patient room, and a research team has no reason
to be in there to be able to invite participation.
Where something like a waiting room is a semi-private place,
so you can semi set up shop in a waiting room
and invite people to participate,
but not in a hospital room.
Which is by the way is different in San Francisco.
Their team can go into the hospital room,
because San Francisco interprets it differently.
They say, if you have a clinician going in,
are you interested in hearing more about a study?
San Francisco is more concerned about the applied coercion
of having a clinician make that offer,
rather than someone who the patient knows
has nothing to do with their clinical care.
This one decision of how do you get in the room,
and do it appropriately, is a big one for you guys.
You have to identify first approach,
and people who are interested, you have to
inform them well about what this study
means and doesn't mean
and if they're interested, interview them.
This can all happen in very minutes to a few hours maybe,
sometimes even quicker, but if it's quicker,
we have a hard time making it happen,
given the amount of things that have to
physically take place.
The way you do this is gotta be with a small footprint.
If we came in there and said you know what Dr. Gramling,
you're a good guy and everything but just sit over there
and just wait, we're not done yet,
you say that enough to me and I'm going to call up
and beg Laura not to tell the team next time
that we have a potential eligible consult because I want to
get home for dinner sometime.
What I do need though, is someone to say, if I say,
hey I have someone who is participating, I'm interested,
what do I need to do that can make this happen
and can be hopefully the least amount of pressure on me.
I need someone to advise me.
Liz does that the most.
Raise your hand Liz.
There you go, right on Liz Deity Koontz,
our study coordinator, and often times on the floor
answering these questions.
You need the guidance, there is some footprint,
but it's got to be gentle,
or else it's not going to be sustainable.
There's plenty of studies of physician behavior
in particular, you get people ramped up and do 20 to 30
of something, and you just exhaust the team so much
that you're done.
That wasn't what we wanted to do.
We want to try to do this simply and sustainably.
What are finding so far?
Let me ask before I go on to some of the findings,
are there questions that people have
about any of the methods,
or people that do this type of thing that were wondering
something we had been... a decision we had to make?
(speaking off camera)
Mhm...
(speaking off camera)
Yes, there was one original criteria that we abandoned,
and that was the reason for palliative care consultation
was not for symptom management only,
because there are some people who are referred,
we used to have a system where there was a checklist
that people would put in their documentation
of why we were consulted.
That went away with EMR, so we lost a little bit of that.
There are some people who are referred to for whom, yes,
death is always possible because people are sick,
but probably pretty far off and it's not really active
decision making that they're making at the moment,
that's maybe talking 5-10 years down the road,
yes, those people are not the ones who are typically
referred to us who have an advance cancer.
Those are relatively rare.
If they have a cancer that's metastatic
and the team thinks that they have a long survive prognosis,
we're not typically seeing them as much
in (mumbles) setting,
it kind of naturally takes care of itself.
Most of people by the cancer criteria itself,
is putting it in the range of...
of a months to year kind of expectation.
Or I guess days to year, depending on the spread.
(speaking off microphone)
I was, in fact I had to do that through
The American Cancer Society, through IRB here,
as I felt both so I could understand the processes
that my colleagues were going through,
I should be doing this so I have
an intimate understating of it.
There's a couple of places that might make me
exclude my data when I present it, but most probably won't.
In fact I think I might have been
the first person that did it,
or second or something like that, I don't know.
Go ahead.
- [Woman] Are you... was there a question up there about
the physicians (mumbling)
Are you also testing the physicians
(mumbling)
- Oh, in the conversation itself what are we measuring?
- [Woman] Yeah, what's your primary focus (mumbling)
- Good, I'll jump that.
Hold that, if I don't get to it in five minutes,
jump up and stop me.
Theresa and I worked together,
so she knows she can jump up and stop me.
What we're seeing for some of the outcomes
that we're measuring, are including right after
the conversation asking the clinicians,
what is this persons current status after this conversation,
is this person, their goals of care,
the treatment choices that we should be thinking about,
focus primarily or exclusively on comfort oriented approach,
with no limitation to how aggressive the treatments
we might be thinking about for disease management
or somewhere in between?
Right after the first conversation,
almost half of patients are saying I don't want
any limitation to the type of treatments
I'm being offered for my disease.
30-40% in the middle and about 15% I'm interested
in a pure comfort approach to my treatment.
That's right after the conversation.
About two weeks after the first conversation,
about 20% of documented comfort measures
only approach as a medical order in their chart,
about 71% have enrolled in hospice
at some point before they have died,
I don't have the data about how many days
prior to death have they enrolled.
And 11% of people have had a procedure
in the last two weeks of life that the Dartmouth Atlas
and other health services leaders
would identify as having no conceptual benefit
to the survival and substantial impact on quality of life.
One of the findings that we're,
actually it's just coming out this month,
is our measure of being heard and understood,
this idea of how can you measure
what good communication looks like.
This concept of being and feeling heard and understood
is an important one, but we didn't have a clinical measure
for that that was going to be useful in this environment.
In the process we created ours and this is what people
are responding to at the beginning,
before the consultation, it was over the past two days,
how much have you felt heard and understood by the doctors
and nurses and hospital staff.
We were interested not only in how well you heard
one person heard you, but how do you feel in this
environment that you're in?
Because that's really the end game as what we're really
trying to promote.
We did explore possibly trying to measure
how well people felt heard by the palliative care team
and others that was a no go because a lot of people
were very kind and had no idea who the heck anybody was
and couldn't identify it all even if they wanted to.
The day after consultation, ask how much do you feel heard
and understood by doctors, nurses and hospital staff today?
So it was two measures.
The good measures for doing this
in the clinical environment, is we found people who chose
to read this off of the survey and answer
had the same distribution as those who chose to have
our study staff read it to them
and not actually see the piece of paper.
The other good news, for those of you who try to study
particularly physician behavior but clinician behavior
more broadly speaking, there tends to be
a ceiling effect to those.
For a few different reasons.
If you like your doctor,
you tend to put good on everything, right?
Anyone have PRESKANI results that try to distinguish
what you're good at and not good at,
the yield gets pinned to whether they like him or not,
or her.
What we're finding is that by completely giving people
a little bit of chance to say I feel quite a bit
but not completely, and that allowed a little bit
of distribution to our responses
which has been very valuable.
We also found, as expected and consistent with
other research in palliative care
that the predominant direction effect was going up
after the palliative care consultation,
a few went down,
primarily from the completely to the quite a bit.
But the overwhelming direction was improvement.
And so this is also plausibly going to be
an important measure to use either as indicators of quality,
we are more paid and our systems are designed
to promote things we want to promote,
and potentially for an outcome measure in trials
that are occurring in this setting.
(speaking off microphone)
So any percent...
Anyone who had not at all in the beginning,
85% of them got better
and they couldn't get worse which is why there's not at all.
Theresa, your question was what do we do about this?
We have all these recordings of these conversations.
Our approach to conversation analysis
is one of ecological thinking, and that is that
this conversation is a thing.
It's an organism.
It's really complex.
It's got lots of features.
It takes place in some habitat.
And sometimes they're beneficial,
and sometimes they're not, depending on their context.
What that leads us to, what that helps avoid doing
is saying that did this person say good 16 times
and if so that was good.
We're focused a little bit more on thinking
of this more globally, so do that we need to have some
reasonable conceptual model as someone,
one of my mentors would say,
the purpose of a conceptual model,
the one feature that they should all have,
is that they fit on an 8.5 by 11 piece of paper
and they're readable.
Because everything is complex, right?
Everything is complex.
What we wanted people to see is what we think about
as a nonlinear exchange of information,
but rather your relationship that's circular,
that's dynamic, that's relational, and that has
some observable components, these are the blue pieces,
for which the data up until the time of this study
would suggest these are some domains that are important.
How do we balance or manage these uncertainties
that people are facing?
How are we addressing people's emotion,
how are we beginning this conversation?
Whether this is bout prognosis or goals of care or just
who are you, why are you here?
And the exchanging information,
how much am I learning about you, how much are you learning
about my opinions about what's going on
in information sharing.
(speaking off microphone)
Probably not much, but the...
Tim is it okay if we talk a little bit
about where this is headed?
I've heard this is important to you.
That might be something that I've just opened that door,
and one way of getting into a conversation about prognosis.
Another way of engaging the conversation
might have been someone asking me,
hey, what's going on here?
Or making a statement.
God, I'm going to die of this.
Or something like that.
The first foray into how a conversation started
or how the table is set for conversation
is what that engaging in is.
So the metrics we use, what that looks like,
are different than necessarily the quantity or pace
of information sharing, or content of information sharing.
Does that answer your question?
Other questions before...
Any of these things are incomplete.
Those of you who are considering this type of work,
I recommend book chapters are a great place to get
these ideas out there, these will never get them through
peer review and a journal
because everyone will want something else
lopped onto this sucker.
What we do for our conversation analysis,
is to listen to, or I don't, our coders
listen to the predefined content we're looking for,
directly from audio, interact with a computer program
for identifying when is something happening and tag it,
tag it and label it.
And we do that for a couple of reasons.
One is that some of the things we're going after
are going to require us to hear the words and the meaning
and also going to require us to hear the sounds
that people are making.
Whether they are sounds of happiness, sounds of sadness,
sounds of distraction, whatever happens to be
that paralinguistics becomes important
for some of the things we're measuring.
And you lose that in most forms of transcription.
The approach we take to analyzing these things,
is probably inherent from the model I showed you,
but it says that, okay, I just measured a conversation,
I looked at the wingspan, I look at the beak length,
I've looked at the talons and I can describe those,
and one of the dangers might be,
this longer talons is associated with more patient centered
outcomes, fantastic, right?
But then the response that we all make,
is to put talons on everything.
So we have to be careful in understanding
that longer talons, if the rest of the bird looks like that,
are probably valuable.
Longer talons if the bird looks like that are probably not.
So what it makes us think about is conversations as being
different types of conversations.
They're just different.
And we can recognize that by looking at patterns
in how the table is spread out,
in how information is shared,
what types of emotion are displayed
and how are they responded to,
how is uncertainty addressed or not addressed.
In seeing these different types of things,
it begins to allow us to think about, okay,
if this was a type X conversation, or type middle,
does that seem to lead to good outcomes
when people are facing this host of decisions.
Yes or no?
Because each one is going to have, theoretically,
some adaptive capacity depending
on the environment in which it exists.
That's the theory of how we approach
our conversation analyses.
Is that kind of what you were asking?
Or have I avoided it completely?
- [Woman] Just wondering what was the most important there,
but you're right--
- Penguins.
- [Woman] Yeah, okay. (audience laughs)
- Yeah, good.
So this is what you might look like.
When we analyze, it's a form...
For those of you who have done factor analysis types things,
it's a form of exploratory factor analysis.
We call it a latent class analysis.
Where is Sue, Sue Ladwig, there she is.
So Sue and I had a lot of fun in a preliminary study,
of trying to do this justice and not just kind of
make up stuff, right?
Because when you do exploratory factor analysis,
you try to make qualitative judgments
about quantitative data,
it's just inherently what you're doing.
We measured in the pilot study for this,
50 different features of conversations
within those different domains.
Of those, there are about eight that seem to identify
different types of conversation.
We tended to see from most different ways of thinking
about this, about three different types of conversations.
The way that you can see these, and how they look different,
is these lines represent the proportion of conversations
having X type of feature.
When I do a talk like this for people who are geneticists,
they love it.
They can now see why this is valuable, because it's a gel.
That's why I keep using it.
This gives you an idea about how we're going to think about
types of conversations within each type, some of these
different features are going to be important.
It isn't to say that for every type of conversation
out there, there may be some features that are always
good, there are some that may always be bad,
or close to always, right?
So it may be that compassion is going to be something
that's going to be valuable regardless of
the type of conversation depending how you
operationalized compassion and how it's expressed.
(sighs)
Questions on this so far?
So no, we haven't analyzed this part of the data.
Those of you who think I'm leading up to something
extraordinarily pivotal here, it's not...
You're not going to get it. (audience laughs)
I wish I could give that to you but I am really
not able to do that today.
The conclusions for what we have so far is...
If these methods that we've used have been sustainable
for a relatively long enrollment period
in a multi-site study both here and in San Francisco,
that recruitment enrollment is possible.
That self report, so getting self report from people
who are very sick is possible.
Directly observing and recording conversations
in these very intimate and important settings is possible.
Health service legalization follow up is possible.
That our heard and understood item I think is a promising
metric for quality measurement in palliative care and we're
kind of promoting it as such because as of yet we don't
have patient level measures and that's really something
that the discipline is needing.
We expect to complete our enrollment in June
and our conversation analysis will kind of unfold over
the following months.
I want to thank American Cancer Society as one of...
Really just... if it hasn't felt that way to you
from thinking about this context
when we're doing these studies,
just a really deep gratitude to the patients and families
and the palliative care clinicians
who make this meaningful work, make it possible,
and we'll make it possibly by what we produce from this.
This is from a patient who, I think this is from a guy
I hear from every few months or so,
sends us letters on his update and we keep contact
with a lot of these people who have actually seen this
participation as part of a legacy, as part of something
important for them, and send us pictures
and member participation and it's really fabulous
and you're really thankful that you're able to help.
Well, thank you very much. (audience applause)
- [Woman] We definitely have time for questions
and while we're doing questions,
if you could make sure to fill out our evaluation
so that we can get these back
and especially the part about patient topics,
we'd love to have topics here that you're interested in.
You're available for questions, Dr. Gramling?
- I am.
- [Woman] There, in the middle.
- Oh, please, yes.
- [Audience Member] Where are you going
(voice drowned out from distance of microphone)
- To the University of Vermont, started a new division
of palliative medicine there
and I'm going to be their chief.
- [Audience Member] That's a tough one.
- Yeah, I got that one. (chuckles)
(speaking off of microphone)
- [Audience Member] Groups or providers
(voice drowns out) but the ethics teams is usually deprived
of folks who are gratified in those different groups
(mumbling) was there any specific effort to have
some of the ethics team involved in research (mumbling)
- So the short answer is yes, not part of this project
but we did have some early conversation with the ethics team
about, is this something you might be interested in,
a similar type of project.
At least our initial conversations was yes,
that's quite interesting,
now we need a million dollars to make it happen.
I mean this is expensive to do.