2016 Public Health Grand Rounds 04/15

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

- Uh oh.
Trying to push this little button.
But it's not doing what it's supposed to.
Oh there we go, yep.
Does this work, do you hear me?
- Yes. - Yes.
- Too loud, too soft?
Okay, good.
What I'd like to do today is
talk a bit about a research study that's been occupying me
for about the past,
eight or nine years from
conception to production of some results.
And also just to think together about
a really set of difficult problems in communication
about prognosis and treatment choices
with patients who have advanced cancer,
and I think this will generalize to other situations.
I believe this is actually my most important slide that
I have no conflicts of interest to report.
(stammering)
So the
goal I have for today is to share with you
some understanding about the impact of pro-communication
and misunderstandings in cancer,
in patients who have advanced cancer,
and I'm defining that as people whose cancer is
generally not regarded as curable.
And I think the motivation for doing this project
was really to see if we could enhance quality of life,
and reduce unnecessary morbidity
that's attributable to patient misunderstandings
and poor communication.
I'll explain why that's an important issue.
So start out by imagining that,
let's say that you've had stage four lung cancer
for the past eight months.
You've had radiation to which you in quotes responded,
then had palliative chemo therapy.
I'm putting quotes around these words
because these are words that are very frequently
misunderstood by patients.
Palliative chemo therapy that offered a good chance
for improvement with a possibility of some mild to moderate
side effects of nausea and nerve damage,
then discovered the chemo wasn't working,
so now your doctor asks you what you want to do.
To start more chemo to see if you respond?
Or take a rest from chemo?
You're having some pain in your hip
and you don't know if it's gonna get better.
Now just kinda plant these questions in your mind.
How are you gonna make this decision?
What do you know about your illness?
What are you confident that you know about your illness?
What do you need to know,
and what questions should you be asking?
So the situation current is there is about
a half a million patients in the US right now
living with advanced cancer,
and that these patients report a heavy burden of physical
symptoms, side effects, and emotional distress.
They receive treatments that they often later think
that they shouldn't have.
They tend to use palliative care and hospice care
late in the course of their illness.
And they experience a high degree of medical bankruptcy.
Most patients with stage four cancer including here
at the U of R, continue to believe that
chemotherapy might cure their cancer.
So chemotherapy for stage four cancer is always
given with the intention of palliation and not cure.
Those who haven't had conversations
tend to get more burden some aggressive treatments
like being in the ICU on ventilators,
dialysis, which reduces quality of life
with little or no effect on survival,
and discussions of these end of life wishes
tend to happen during a terminal hospitalization.
Not before that when people can actually make decisions
regarding their care.
Several large national organizations,
National Cancer Institute,
The Society For Clinical Oncology,
Institute For Medicine, National Priorities Partnership,
have all called for improved communication between
patients and physicians about choices, prognosis,
and their disease status,
and they site that their might be effects
of those communications on quality of care,
quality of life, and also that it's an ethical mandate
for patients to know and understand what's going on with
their illness.
So about 90% of patients when you survey them say
that they want to be actively involved in their care
and that they value frank and sensitive conversations.
But patients typically express ambivalence if you say,
well would you like to talk about this today,
they say oh no, today's not the day.
So there's this paradox, but also on the other side
clinicians when surveyed frequently say
that they'd like to have, they'd be willing to have
these discussions now with their patient today,
but then don't.
So you'll survey them right before a visit,
then audio record the visit, the conversation,
the intention does not match by action.
They say that they wait for patients to signal interest,
and offer either intentionally or unintentionally
optimistically bias or indirect information.
So the words that we use before like respond,
is kind of an indirect word.
Respond means that you don't see the tumor growing
on a scanner or that it might regress,
but it doesn't really mean that you're cured,
or necessarily feel better.
So we started out with a hypothesis
that health communication might
directly reduce anxiety, clarify patients needs,
promote decisions, foster therapeutic relationships,
reduce power asymmetries, and enable patients
to manage their disease better,
and indirectly promote care that's matched to patients
needs and wishes,
improve their symptoms, lower their financial burden,
et cetera, et cetera.
So this is kind of the broad hypothesis of
not this particular study but really of a career of work
trying to improve the quality of communication.
We know that good conversations are associated with
no increase in anxiety.
That is if you talk about prognosis
and then survey patients afterwards,
they may have a little blip of anxiety
after the conversation,
but a week later they're basically back to baseline
or maybe even a little bit better.
They tend to enroll in hospice earlier,
they have lower rates of aggressive care,
enjoy better quality of life, lower costs,
either have some improvement or no effect on survival,
and lower risk of morbidity in the caregivers
and those taking care of them.
So this is in observational studies.
These are not intervention studies.
So in terms of interventions,
there isn't a lot out there in advanced cancer.
There are,
is a study of intensive residential longitudinal
workshops for trainees
either residents or fellows.
James Tulsky did a brief in office intervention
where he video recorded conversations
then embedded bits of physicians own videos
back into a CD that became kind of an educational medium
for those physicians to review.
In patients with oncology settings
who weren't necessarily stage three or four,
there was an intervention to help patients talk about
their pain which was effective.
And in palliative care settings using question prompt lists
which I'll describe in a minute,
increased question asking but only when physicians
also encourage patients to ask questions and
endorse the use of the booklet.
Why is this a problem?
No one seems to want to talk about prognosis,
and in fact there's some indirect evidence
that patients are less satisfied
after visits when prognosis discussions occur.
So they say they want it,
they tend to avoid it,
and they're less satisfied afterwards,
so this acts as a huge disincentive.
Cognitive and emotional overload,
there's complexity of multiple treatment options,
when you're sicker you tend to differ decisions to others,
and also when you're terrified
you simply can't take in complex information as well.
Oncologists where communications feels inadequate
is it just that they're rushed,
is it they're emotionally overwhelmed?
And we don't know this.
And also from our data we're seeing that
caregivers both facilitate and impede information transfer,
and they impede patient agency.
So they can both help the patient express their needs,
but also cut patients short.
What we wanted to do is to create an intervention
that would reduce power asymmetries,
activate and support patients to take a more active role
in their care.
Help oncologists elicit and respect patients preferences.
Help patients and families collaborate to make more
timely and informed decisions.
Attend to both cognitive and socioemotional aspects
of decision making and also inframing information,
balance, pessimistic, and optimistic ways
of expressing information.
This treatment has a 20% chance of helping you
and an 80% chance of not helping at all.
That would be a balanced way of framing it.
We worked for a long time
trying to figure out how to do this,
and recognizing that if you just train patients
to ask questions,
those questions might be blocked by the oncologist,
and if you just trained on oncologists who
kind of open a space for questions,
patients might not know what questions to ask.
So we thought to intervene on both levels,
with oncologists and patients separately.
Pragmatically speaking oncologists are very busy,
and so to ask them to come a group training session
of some sort clearly was not going to work.
We have to come to them.
For the oncologist intervention
we wanted to improve some key communication skills
to reinforce patients abilities
to communicate back with them,
so to basically open space for patients
without unduly raising anybody's anxiety,
the oncologists or the patients.
And we identified four sets of skills
based on the literature that we kind of
created one word names for them,
but they're clearly more nuances than that.
Engaging really refers to activating patients
to take a greater role in the consultation.
To ask more questions.
To express their concerns.
And for clinicians to then open the door for questions
and clarify those concerns that patients raise.
Responding in this context means expressing concern,
support, expressing empathy,
dealing with the emotional side of these discussions.
Informing, asking patients what they wanna know,
providing information, and checking understanding.
And also framing either using a best case/worst case
scenario or dual framing as I just mentioned a minute ago.
We trained actors to portray
a common patient situation.
In this case it was a patient with,
with advanced colon cancer.
But prior to the first session,
training session with oncologists,
we had them actually review this case,
and it was written up as a chart note
just like you'd see one any record.
They would meet with a standardized patient instructor.
We created a 15 minute DVD that had both
some kind of basic principles,
but most of the DVD was taken up with
actually providing video examples
of effective communication strategies.
How do you ask an open ended question?
How do you raise the issue of prognosis?
What do you do
when the patient seems emotionally distressed?
How can you present information in a balanced way?
After the DVD,
they basically have a mock office visit.
So the patient comes in with his wife,
and at a particular juncture
where a decision has to be made.
The disease is progressing and the patient,
the standardized patients are prompted to ask questions
like what does this mean in terms of prognosis?
This standardized patient instructor give feed back.
And then the oncologist gets the DVD,
a set of communication guidelines,
and I'll just show you what they actually get.
There's an intersession during which
we recruit two study patients,
audio record them,
the standardized patients
actually listen to the audio recording,
and actually give feed back by email to the oncologists.
And then a month later there's another in person session
where this patient has now failed
another set of chemotherapy
and is looking at end of life care.
And so again they have a mock office visit
and some feedback following that.
And so,
this is what the DVD contains,
an introduction, challenges that oncologists face,
what the research says, how the program can help,
and then demonstration of all these key skills.
And this is the role,
an accountant with colon cancer, medistatic delivered,
diagnosed six months ago, who wants to be informed
and involved, but doesn't really know a lot about
his prognosis.
Standardized patients are programmed
so that if the oncologist uses
effective communication strategies,
they then respond by giving more information,
expressing emotions.
They would offer less information
if the patient interrupts them, uses closed ended questions,
ignores their emotions, et cetera.
More information if they listen,
ask, prioritize, answer, respond, discuss prognosis,
and involve the caregiver appropriately.
So these are actually very skilled
and experienced standardized patients.
And the feedback for us just really includes
an item by item review in terms of the skills
that were first presented on the DVD.
And a focus more on reflection and discovery
not on evaluation.
And then the,
the oncologist actually gets these criteria.
About establishing focus,
encouraging questions,
answering the difficult questions,
exploring and validating emotions,
and engaging family and caregivers.
So they get,
the same information in visual form from a DVD,
an experiential form from the standardized patients,
and in written form on a card.
We want to get every possible sensory loadality,
to facilitate learning.
Now patients meanwhile,
when they were rooted for the study,
agreed to have a coaching session with the coach.
And at that coaching session they received a booklet
that looks like this called My Cancer Care,
What Now, What Next, What I Prefer?
And the booklet contains suggested questions
that patients might ask
that were derived from a series of focus groups
that we conducted and also a group in Australia conducted.
And the coach reviews the booklet with the patient
to ask the patient to identify two or three questions
to prioritize for their next visit with their oncologist.
And they would actually practice with a coach,
asking some of those questions.
You know is this question hard for you to ask?
How would you ask it?
Tell me in your own words how you would phrase this?
What do you think might be difficult
when the doctor seems rushed?
How do you deal with a rushed doctor?
So it's really coaching on about
how to communicate more effectively.
And they would have three follow up phone calls,
with the coach,
usually before subsequent office visits
with a physician.
So this is what the interior of the booklet looks like.
It has a few questions on each page.
And this is just the last two pages.
The last page actually contains statements
rather than questions.
cause we found that some patients
just simply wanted to express a value.
So I wanna know the pros and cons of the treatment,
like I worry I'm gonna suffer.
I would rather not discuss how much longer I have left.
So these are both in the positive and negative.
Questions like, if cancer treatment doesn't work,
should I consider stopping anticancer treatments?
Who will be my doctor?
How will I know when I might need to consider hospice?
It's really a wide array of questions.
In order to get enough patients and physicians
we had to go outside of Rochester.
We involved every major cancer group here.
Three groups in Buffalo.
And also a group in Sacramento California.
And we wanted patients whose cancer
was not going to be curable, and we're likely to
have an average survival of our a year or less.
So we asked oncologists, their nurses, their staff
to identify patients with a stage three and four disease.
If they had stage four disease
we were fairly confident that there would be
a reasonablyhood of them not surviving a year or more.
With stage three it's a little bit more iffy
so we added what's called the Surprise Question,
which is asking the oncologist would you be surprised
if the patient died within the next 12 months?
Not asking to predict the future but just...
And that seems to be fairly predictive.
And then we also recruited caregivers.
So in order to get a baseline measure
of physicians communication skills,
we actually recruited some patients
who just simply had an audio recording
of their office visit and completed surveys,
and we call that phase one.
Then we randomize the oncologists to a training intervention
or a control.
And patients of that oncologist
all were intervention patients.
So this is why it's called a cluster randomized trial.
So if your doctor was getting the intervention,
you would get the intervention as well.
And then the primary outcome measure
was an audio recorded office visit.
That is one effect of this training have.
We also followed patients for up to three years
after this, giving surveys every three months
about quality of life,
what was discussed in previous visits,
and several other things that I'll talk about in a bit.
Finally for those patients who died, we audited their charts
to see what interventions they received
in the last 30 days of life.
So phase one and phase two.
Oncologists practicing in this region as well as Sacramento
patients age 21 and older,
and caregivers could be anybody the patient
identified as such as long as they were 21 or older.
The audio recordings were with,
they put an audio recorder in the room,
turned it on, left the room,
and collected it afterwords.
They were later transcribed.
And we try to code the audio recordings
using the exactly same skills
that we were training the patients and physicians to,
in terms of the communication training.
So there's a coding system that was developed by
Rick Street in Texas that basically codes
patient assertive behaviors like question asking,
interrupting, clarifying, providing opinions.
There's a coding system developed in Virona Italy
that Bob Gramling revised.
To code health physicians responded to patients
emotion statements.
Whether they kind of opened up space for discussion,
or whether they went on to a different topic
and seemed to ignore it entirely.
Cleve Shields who used to be here,
who's now at Purdue developed a scale
that encompassed both informing and framing
that is which topics were discussed,
who brought them up,
and also how were those, if there were decisions to be made,
how those decisions would be framed.
Because we were using different measures,
we had to find some way of aggregating them,
and this is where,
we had a lot of discussions with our statisticians
because we wanted to create one outcome measure.
That is the did communication improve or didn't it?
And so each measure was then Z scored,
so it was scored on a minus one to plus one scale.
And then,
(stammers)
each of those Z scored scales was aggregated
into one summary scale for all four of those measures.
We did the same thing in measuring
the physician/patient relationship.
So we had three different scales,
each of which captured a different aspect
of doctor/patient interactions.
And then we also did the same,
and so we aggregated those as well.
We also assessed shared understanding by asking doctors
and patients, both before and after the intervent...
I'm sorry, patients before and after,
but the doctors after that audio recorded visit,
questions about prognosis.
How likely do you think that you will be alive
two years from now?
How likely, and I mean that's the same thing
as the physician,
how likely do you think this patient will be alive,
and we actually compare those responses.
It's a pretty loaded question.
And I know in at least one other study,
a high percentage of patients actually refused to answer
questions about their own prognosis.
So we did pretty intensive training
of our research assistants to try to avoid that problem,
and also kind of plan to that question,
in the middle of the questionnaire
so after they had developed some rapport,
and had kinda ramped up the intensity
of the kinds of questions being asked.
And we actually had very, very few refusals.
Some patients said, I don't know.
But there were very few patients who said,
I wanna stop this interview right now,
I don't think there's anyone who did.
Or who actually refused to answer the question outright.
So again, kind of a methodological challenge.
Even things like a scale like this.
How we came up with the seven point scale.
Other people have used a little, you know,
very likely, very unlikely kind of words,
or who have actually graded it as
by every 10%, like 0-10%,
11-20%,
21-30% et cetera.
And then we ran into issues of numeracy.
Do people really understand percentages.
And would words of percentages work better?
Again a method/logic challenge,
I don't know if the results would have been different
had we asked the questions differently.
But we defined discordance as two or more categories
different so if your doctor felt
you had a 10% chance,
and you said that you had a 50% chance,
then that was considered to be discorded.
Quality of life surveys.
There are no surveys,
or at that plan there were no quality of life surveys
explicitly designed for patients with advanced cancer.
But there are lots of surveys designed for patients
with cancer and patients in palliative care.
So we actually did an aggregate of those scales.
The beguile was developed for palliative care patients,
and their psychological essential well being scale
seemed to speak to the needs of this population,
where as the fact,
we took the physical and social functioning subscales,
their psychological and existential scales,
really did not see a bit more off the mark.
So again we kind of created and aggregated measure using
a composite Z score.
Similarly for utilization there's no standard way
of calculating utilization of healthcare at the end of life.
So we had to go to the literature
and basically create our own.
And think about what indicators are the most important.
So this is in the last 30 days of life.
whether you receive chemo or not seems important,
'cause chemo can increase burden of suffering.
Things like, CPR, mechanical ventilation,
dialysis, even insertion of a feeding tube,
we added to that next because those are all indicators
of aggressive care,
and going to the hospital nobody likes.
Hospice utilization we measured yes or no,
and also the number of days in hospice,
and we had a number of other metrics that we used
for secondary analysis.
We also noticed that oncologists are different.
They sub-sub-specialize so oncologist who take care
of just breast cancer patients,
have a patient population that's probably 90% women.
And to compare that oncologist with another oncologist
who just takes care of lung cancer patients
or head and neck cancer patients
who tend to be predominantly men,
would introduce some statistical bias.
So we created nine strata,
one for each oncology setting,
and then also strata for whether they are predominantly
breast cancer versus not.
We didn't have GYN oncologists in the study.
And then we allocated them into intervention or control.
Now if you remember we audio recorded some patients
in this phase one of the study,
that weren't actually followed for the full three years,
that we used as base lines,
so in calculating the effect of the intervention,
we took into account the physicians baseline
communication skills as well as the communication skills
demonstrated in the audio recorded visit
for the phase two patients.
And we adjusted them for the usual things.
We did actually better than expected recruiting physicians.
We initially expected to get 30.
We got 38, actually 43
but then five dropped out very early on.
And that's actually a good thing
because the more clusters you have,
the more robust your analysis are.
They identified 453 patients,
316 of whom ultimately consented.
281 were found to be illegible.
And we ultimately got down to our target number of 265,
of 135 in the control, and the 130 in the intervention.
That was our target.
Not surprisingly these are cancer patients
with advance cancer.
There were 64 years of age on average,
but the range was quite large, from 20s to the 80s.
Largely white.
28% had less than a high school education,
or high school or less.
We divided the cancers into those that we called
aggressive like lung cancer, pancreas cancer,
versus those that are more indolent like
breast or colon cancer,
to see if the intervention would have different effects.
About 20% were very low incomes.
And you can read this.
Nearly 3/4 had an enrolled caregiver.
Let's see.
So what did we find?
This composite communication outcome basically,
did the intervention doctors communicate better,
taking into account all four of those domains,
and the control physicians?
They did.
That is that we saw a meaningful difference.
The effects saw was point three,
and so I think we can confidently say
the intervention actually had the desired effect.
We broke It down into the four component measures,
and three of the four were going in the right direction.
And one of them was actually significant independently
the engaging one.
Responding to emotions again showed
a trend in the right direction,
with a fairly large effect size.
The informing also a large effect size.
The one that we saw no effect whatsoever was framing.
That is the 10% versus the 90%.
And we only found eight cases among all of these 265,
actually including phase one nearly 400 audio recordings
where a physician actually presented a decision
or said, well this treatment has a 20% chance of helping you
and an 80% of not.
So this behavior despite the intervention
just doesn't occur very often.
This is the discordance in terms of survival prognosis.
59% in intervention and 62% in the control.
There was no effect of the intervention on this.
But this just speaks to the fact
that even now in 2016,
patients have a wildly inflated sense
of how long they're gonna live with advanced cancer.
In fact, there were quite a number of patients,
I think about 10%, who said that their likelihood
of being alive two years from now was a 100%.
And there was zero chance that they would not...
I mean any of us can get hit by a meteorite,
so it was really quite striking.
Curability clearly was better, but the fact,
and this was all just discordance in favor of optimism.
The fact that nearly 40% of
patients here in Rochester believed that
their stage three or four cancer can be cured
is also striking.
So we didn't make a dent in that.
Previous studies have actually showed that,
frank communications about prognosis
and treatment choices actually can reduce
patient satisfaction.
So here we are actually not looking for an increase.
There are large ceiling effects here.
Almost all of these scales
were near the top of the rating scale.
And this is always problematic in any physician.
So these surveys, where you are surveying
very vulnerable patients who dependent on their physician,
it kinda creates this cognitive dissidence
that you wouldn't want to rate anyone
to whom you're entrusting your life as anything but perfect.
So we were afraid that there would actually be some decrease
and there was none.
So that's what we were looking for here.
Now this is,
(stammering)
I'll take a step back here.
Quality of life in cancer, even advanced cancer,
tends to be stable for a very, very long time.
If you ask people to report their quality of life,
it tends not to change
even when they're receiving chemotherapy.
It's really quite extraordinary 'cause looking at patients
receiving chemotherapy you think,
oh they must feel awful or their life must feel meaningless,
but in fact it doesn't.
And this just may be due to adaptation or
just feeling thankful that they're alive.
We don't quite know but this is what we observed
until a few months before death.
And starting about two, three, four months prior to death
there's what's been termed a terminal decline
in quality of life.
So we did our analysis on quality of life
in this groups of patients,
by taking that terminal decline into account.
We didn't invent this method, it was invented by
Marie Bakitas's group.
In which we look at the slope of quality of life,
and here you can see that it's really flat.
And then for that,
period before the terminal decline
and then do a separate analysis
for that terminal decline period,
and empirically we define that as six to nine months.
Now this graph shows not quality of life,
but the difference between quality of life and as reported
by intervention patients versus control patients.
And so you'll see that at baseline,
intervention patients were
reported a slightly lower quality of life than control.
And for the period of time they were in the study
until about nine months prior to the time that they died,
they experienced a relatively stable quality of life
and there was no difference, no change
in the relationship to intervention and control.
However starting about nine months before death
those patients who,
who died
we saw that the difference in the quality of life
between intervention and control actually changed.
That difference wasn't statistically significant.
These error bars cross the zero.
But we found this change compelling,
that in fact there was a delayed effect,
or there maybe a delayed effect to interventions like this.
It's impossible to tell what would have happened
if we had a larger sample size.
But this was,
intriguing.
Utilization of, there was no difference in chemo,
aggressive treatments, hospital days, and hospital visits.
What we did notice though was that,
what happened with survival was not quite what we expected.
First of all patients in the study overall lived longer.
The mean survival was,
was 16 months overall,
and then we saw there was actually a difference
between the intervention group which you see in red,
and the control group which you see in blue.
The survival in the intervention group was 19 months
compared to 14 months in control.
Now this was not statistically significant either.
We'd need a much larger study to show something like that.
And I would have discarded the result
except for the fact that two other studies
that involved in very different ways,
changing communication patterns
between patients and physicians,
also showed a survival advantage in patients
who were in the intervention.
And the one thing I'm finding,
I'm wondering about is whether if we aggregated
those studies if we would then see some kind of overall
survival effect.
Why this happens, no one really knows?
And also it's really true, we still don't know.
What we've shown in the study is that it's possible
to actually create an evidence,
an evidence based intervention,
that is based upon the literature that we have
that's feasible, it's acceptable,
the oncologists liked it, the patients liked it.
They said they learned a lot.
That at least in the short run was able to improve
pretty well established communication patterns.
Most of these oncologists have been in practice
for quite some time.
And so to see a difference in communication
even on a single observation,
is significant because
we all get into habits of communication over time.
It improved communication in those key areas
relevant to end of life care,
except for framing,
and that's something we still haven't gotten to.
And the patients were more active and engaged.
They asked more questions, expressed more concerns.
In analysis we are currently doing
those questions were much more likely
to be about the future, about prognosis, about treatment.
And not about, you know, can I get my chemo on a Thursday
or on a Friday?
Which is what the control patients tended to ask.
They tended to ask more logistical questions.
And there were no declines in ratings
of the patient/doctor relationship.
The other news or the other areas in which we have to
do further work, there really was not
increase in shared understanding and prognosis.
Increase of 5% discussing prognosis in their visit,
to 16% discussing prognosis during their visit,
clearly was not enough.
Again this was just one visit,
but still they were prompted to do so.
(stammering)
And this has not changed in 20 years.
There have been studies of prognosis discordance
dating back to the 1990s.
This has not changed.
So what do we need?
Do we need more training sessions for patients
and oncologists, booster sessions?
Or is this more of a societal issue?
Is this,
because Americans, for Americans death is optional?
(stammering)
There's really cultural norms that differ between here
and for example the UK,
where I think people are somewhat more comfortable.
The quality of life effect I talked about a little bit.
We were possibly a little bit hampered
by doing the majority of the study in Rochester.
More than half the patients came from the Rochester area.
The norms in terms of palliative care and hospice care,
are very different than the rest of the country.
More physicians are likely to have had
communication training.
(stammering)
We may have shot ourselves in the foot,
in terms of doing a study in a place where the quality
of communication at least has received more attention
than other places.
We only recorded one office visit.
Patients lived longer than expected.
So you'd expect any behavioral treatment to attenuate
over time.
There isn't a lot out there.
I mean this is the largest trial to date.
And I'm hoping that there's more,
'cause this is just kind of a first baby step
into a realm that can make a big difference
in terms of patients' lives.
We developed robust communication measures
that I think others will be able to use in the future.
There again was no guide posts in terms of what to do.
The clinical sites we used weren't diverse.
They were places like Straw,
and places like Pluda, and Interlakes, and Rochester General
but also we had patients from Roswell
and Community Practice in Buffalo.
It's about as diverse as you can get it
in terms of practice sites.
Our patient sample was diverse in terms of
gender, cancer type, income, age.
We could have used more minorities,
and although we have some data on dispairities,
they're not strong enough.
And the methodology I think was something
I'm hoping other people could take advantage of.
So to be conservative and thinking about,
you could take one view and say,
well we actually changed one conversation.
This is a large intervention to change one conversation.
And so the question is,
we've showed that it's feasible,
but is it really worth it?
Will it really make changes in terms of things
that patients and clinicians value downstream.
So that's an important area for future research.
I think the trend now is to provide oncologists
with more communication earlier in training.
Things like the Oncotalk program
that Tony Back has developed.
A lot of oncology fellows go through this,
and it's a three or 4 day intensive communication
skills program.
So is that really the time to try to
change these patterns rather than trying to
have an impact on people who are already in practice,
or both?
So that still remains a question.
In the beginning I just showed this small,
small number of the study investigators.
In fact there were 26 investigators,
listed on the paper we just submitted.
And this is the acknowledgment section of that paper.
So our team was over 60 people to do a study like this.
So it really does take a village to do
community based intervention,
behavior intervention research.
And although getting three and a half million dollars
in a grant sounds like a lot of money,
it was very quickly spent.
(chuckles)
(stammers)
Actually the protocol is actually published there.
We have a few minutes for questions,
and I don't really wanna restrict the questions to
this particular study.
I'm happy to answer any of those but I'd really like
your impressions about how to,
kinda move the world forward
in terms of a particularly critical set of issues.
That is promoting these end of life discussions
with patients who have cancer.
So with that I'd like to open the floor.
Yeah?
(muffled chatter)
I think the challenge now in a way is greater than it was
20 years ago for several reasons.
One there is more that we can do.
It used to be if your disease progressed
after first line chemo or second line chemo
that was kind of it.
There just wasn't anything else out there.
You could be part of a phase one clinical trial
at a few places in the country,
but now for lung cancer, there are 10 different
regimens we can use for stage four cancer.
So you never run out of options.
So if you view palliative care
even having a discussion like this
as when there are no further chemotherapeutic options.
Now in Rochester we have the sense not to give patients
with solid cancers in general, chemo therapy
when they're hospitalized and acutely ill,
but earlier this year I visited a health system
in which patients on ventilators are receiving
4th line chemotherapy for lung cancer.
And this happens in some communities.
I've never seen that here.
That's because there's always another set of options.
There are huge economic incentives to continue with
chemotherapeutic treatment.
I mean that's a national issue,
and coupled with the fact that patients
might be more upset if you even bring up these issues
which at least in the short run, might be,
it creates a situation you know kind of a perfect storm
in which it makes it really, really difficult.
So yes, we were kind of trying to kind of,
make little holes in this brick wall,
but sometimes it does feel that way.
Yeah?
(muffled chatter)
Uh huh.
Yeah I've seen ads for Neulasta.
You know what that is?
That's a drug that increases your white blood cell count
after chemotherapy treatments.
That says if you have cancer ask your doctor about Neulasta.
Now of course, most cancer patients do not need Neulasta,
because their white blood count bounces back just fine.
But imagine if you have these 500,000 patients
with advanced cancer who are receiving chemotherapy saying,
I saw this ad for Neulasta on TV.
Now how much did that ad cost to produce
and put on television?
Well let's say it was three million dollars.
Well Neulasta is about what is it?
10 thousand dollars a dose.
(stammering)
Or two thousand dollars a dose.
You don't need very many
doses of Neulasta to...
So yes a cynics view of this
is that this is all driven by money.
However,
in the UK for example,
and I think in Canada also,
oncologists don't personally benefit economically
when patients get more treatment.
Yet we see the same phenomenon.
So it's not all that.
That's certainly something that...
I think part of it is that in the training of,
for most of us regardless of specialty.
If you're cardiologists...
You're trained to fix things.
And when you can't fix something,
you feel frustrated.
And so you wanna continue to try to fix things
because it makes you feel better.
So it's kind of a change of the chip.
And to think of palliative care as fixing something,
because it is.
It's just fixing something different.
(stammering)
So that's why I say,
is more training about what words to say
and how to say them,
and how to respond emphatically,
is that really going to make a difference,
in the absence of broader societal change?
And that societal change is happening.
In the 1950s fewer than 10% of patients were even told
they had cancer.
And now everyone is told they have cancer,
but aren't really told what it means.
We've made some progress,
but I think it's,
this is kind of the edge.
Yeah?
(muffled chatter)
Yes so my vision is that,
what if we had a practice nurse in an oncology office,
who was trained to actually essentially do the same coaching
that our standardized patients were.
I'm sorry, that our coaches were,
for patients.
Patient comes in for the office visit.
How are you doing today, how are you feeling?
What questions are the most important questions
for you right now?
And I show you this little booklet
that has some questions that patients commonly find useful.
And while you're waiting for the doctor,
you might wanna take a look at that.
In fact I'll come in a couple minutes before the doctor
comes in and just review with you what those questions are,
and maybe we could write them down.
In fact, I could even enter them into the electronic medical
records so before the doctor comes in,
he or she is actually seen those questions,
and is prepared to answer them as best as he or she can.
That would be my...
(stammering)
Doing this kind of research you kind of have a
interventionist, I'm kind of this interloper,
I'm not an oncologist.
You have these people kind of coming and visiting,
but if it was woven into the fabric of care,
then it would not be a one time thing,
but rather kind of business as usual.
Yeah?
(muffled chatter)
What's that?
I'm sorry, I still don't...
(stammering)
We actually approached the GYN group,
and they weren't interested.
(muffled chatter)
Yeah, no it's ...
It's hard.
Last week on palliative care service,
there was 28 year old I was taking care of
who had testicular cancer.
I mean this is something that 90% of the time
or more is curable, completely curable.
In this particular study, we didn't target or tailor things
by age, but clearly if you're coaching,
a patient and their family
you have to take that into account.
And also there are cohort affects.
I mean older patients were socialized in an era
when you didn't ask doctors questions.
And so they don't.
So it takes a lot more coaching.
We're in an era where the average,
30 or 40 year old,
has already seen the internet before their first visit
to an oncologist.
So how do we actually help them use that
without being obnoxious.
Without bringing in a print out of 400 pages that say,
gee, I'm learning about my cancer.
I really wanna know more about this,
but can you help me with how this really applies to me?
I read on the internet that people with stage three B
colon cancer have a expected survival of two years.
Is that me or am I gonna do better or worse?
So you can really capitalize on what skills they actually
do have.
But the emotional support I think is really
particularly poignant because,
most oncologists fortunately spend most of their days
with patients in their 60s.
And then they get a 25 year old,
and it's just...
Are they prepared for that?
Probably not.
Yeah?
(muffled chatter)
(laughing)
(stammering)
I think you're question's a loaded one.
I don't know if you heard in the back but
I think ultimately was do you always
guide what you say to patients
on the basis on whether they will be satisfied
with what you've said.
I'm in palliative care so,
I'm basically in the business of telling everyone
things that they don't wanna hear,
and so there are ways of doing it,
and there are ways of doing it wrong.
So I think that probably that lowered satisfaction maybe,
a result of blunt or insensitive delivery of information,
and we can't know that.
Cause I didn't do those earlier studies.
And I think it talks about your own professional
satisfaction as a physician.
If you've given someone news that they don't wanna hear,
and they get really, really angry at you,
or their family gets angry at you,
or they throw you out of the room
and never wanna see you again,
and this happens to me.
I wouldn't say every time I do a two week rotation
of palliative care but it's fairly predictable.
From time to time this will actually happen.
So you have to really uncouple your own
sense of professional worth from,
the feedback patients give you in moments of distress.
You can't ignore it but you have to kind of
put it into perspective in a way.
Where it starts becoming nefarious though,
is in for example HCAP scores.
Which are ratings that patients have of physicians,
and they are now being made public.
So whether a patient is satisfied with me as a doctor,
totally depends on their own subjective impression.
They can write whatever they want.
Or Angie's List, they can write you know,
this horrible, insensitive Doctor Ebstein
and it will stay on the internet forever.
So all of these are really potent
issues to consider.
Now at this phase of my career,
it doesn't really affect me that much,
but when I was 30.
This stuff could be really devastating.
Exactly.
So it's the forces,
and in fact now we're being paid based upon HCAP scores,
and Press Ganey scores.
So if you do better in terms of these patient assessments
you make more money.
It's a crazy set of incentives.
You lie to patients and tell them they're going to be fine
and live forever and you get rewarded for it
in lots of ways.
I think it comes down to self awareness
and self knowledge.
Yeah?
(muffled chatter)
And that's a really good question.
With just 38 physicians we can't do that.
We have another study that's going on
that's actually an unannounced standardized patient study.
Where we sent actors in physicians practices.
And there we have a 100 physicians.
So we'll be doing those analysis and it's exactly
the same patient.
They all see the same actor who portrays
exactly the same role.
So we'll be able to answer that question
in about six months.
So hold on to it.
(stammering)
I would hope,
that younger physicians would be more sensitized,
but then there are all these other counter veiling
social forces.
You know.
I think we're out of time.
So thank you.
(applauding)