2011 (3): Clarence Braddock III--Supporting Shared Decisions When Evidence Is Low

So, it's always difficult as these meetings go on and you're
listening to really smart people who have thought about things in
more depth than you have to sort of figure out whether your slides
and what you have planned to say is that still going to work,
but I'm going to try it anyway.
When I was first invited to give this and I was, of course,
very grateful and honored to be asked.
I wrote back to Rick [Street] when he asked this question,
"Is shared decision making relevant or how is it relevant in
low evidence situations" to write back and say, "Well, yes.
Is there anything more to say about that?"
And so, what I'm going to try to do is sort of explore that
question because he basically said, "No, no, Clarence,
that's not enough.
We have to go into this a little bit more in depth."
By way of other just disclaimer, I'm coming at this not as an
expert in decision theory or cognitive psychology,
but as a physician ethicist and I say that purposely because I got
interested in this as a physician, being one of those kind of mucking
it up as we've been talking about and I still struggle with this as
I'm sure all the clinicians around the table.
Even though we talked about this and we're all talking the same
language, it's a challenge, but as an ethicist,
thinking about the ethical foundations of shared decision
making and what that has to say for this question.
The other thing I want to say just as a housekeeping thing is being a
Dean, I'm not "the" Dean, I'm a Dean, you're never really away.
So, I have to actually step out at 2:00 to make a call.
So, if you see me step out it's not because I'm not interested.
So, let me just start in then.
So, by way of overview, I wanted to say a little bit about the
ethical foundations of shared decision making,
which you all know, but I think is important to go back to
particularly around this question of why would we do this and why
would we do it in low evidence situations.
I want to contrast the question around evidence with a question
around values and preferences and suggest at least a framework for
the question of what might this look like in a low evidence
situation.
Particularly, I'll talk about four concepts which none of them are
mine because you've heard all these already,
but I think are out there as ways of really thinking about what
shared decision making, the recommendation might look like in
a low evidence situation - the notion of a shared mind around
uncertainty, the notion of closing the knowledge gap,
the idea of promoting patient empowerment and the idea of
promoting trust.
So, you get extra credit points if you know who that's a picture of.
I know you know.
You probably have a better picture of him, maybe more flattering.
So, that's Immanuel Kant who's, you know,
a well known moral philosopher, known as really sort of the
shapers of a branch of normative ethics that's typically called
deontology, the notion being that you determine the rightness of
actions by looking at moral rules.
Consequences are okay, but moral rules are the fundamental precept.
And of course, he became known because for the categorical
imperative, a litmus test by which you could think about whether you
had the right moral rule or had the moral rule right.
But with regard to shared decision making,
it's important to think about what Kant would say.
Kant would say to the invitation from Rick [Street] and Bob [Volk]
and others to come and talk about this, he'd say, "Well,
what are you talking about?
You know, we're supporting patient autonomy through informed choice.
What more is there to say about it,
whether the evidence is high or evidence is low?"
So, I think it's good to reflect on that.
So, the other thing I think is interesting for me is all the
different voices in shared decision making.
Decision theorists have been at this a long time.
People in cognition, metacognition have been at this a long time.
Actually, you know, bioethics was a relative late comer to shared
decision making, but an important one.
So, again, most of you know this story,
but the President's Commission in about 1981-1982 published a report
on shared decision making in which they said the ethical foundation
was connected to the notion of personal well being and self
determination, a.k.a., you know, autonomy and beneficence,
that there's these two ethical principles that underscore shared
decision making.
I also just realized the timings might be on,
so I might have to fiddle around with this.
It's doing its own thing.
The purpose of this report was to contrast consent with shared
decision making; "consent" being, having grown up sort of in the
legal world, as a unidirectional disclosure of information to the
patient.
For the purpose of satisfying a litmus test that disclosure was
adequate and then a patient signed a form.
Those two things constituted ethically valid consent by legal
definition.
The President's Commission framed in an ethical context this notion
of a bidirectional exchange, being the litmus test,
for shared decision making and I think this goes back to some
things we've already talked about about the intrinsic value of that
exchange, the intrinsic value of the fours that Paul [Haidet]
talked about.
Patient centered care, more of a contemporary concept, includes,
among its precepts, this notion of finding common ground as one of
the core elements of at least one model of patient centered care;
"finding common ground" being, again,
a process oriented relational construct that has less to do with
evidence, less to do with the impact of that,
but more to do with its intrinsic value as part of the interaction.
Even the Institute of Medicine, the great deontologic institute
that it is, included shared decision making in its definition
of quality - good communication, shared decision making,
etc. Okay, so, maybe only one person got this one right.
Who is that person?
And no, it's not Benjamin Franklin although it does look a lot
like him. That's Jeremy Bentham.
Jeremy Bentham, of course, of the branch of normative ethics of
consequentialism, of teleology, basically the notion that the
rightness of action are determined by the consequences;
another way of saying that all of us who are outcomes researchers,
who are looking for shared decision making to influence some
outcome are descendants of Jeremy Bentham.
That is, we are rejecting the notion that the rightness is in
accord with the concurrence with moral rules.
We're saying those aren't enough. We need to see consequences.
There's a danger in this and of course,
the danger is that if you don't find the good consequences,
if you don't find that it positively influences some kind of
outcomes, even if they're really good outcomes,
that you may be left thinking, "Do we really need to do
shared decision making?"
And this is where I got interested in it because as you talk to
physicians in practice and say, "Oh, it's a good thing to do.
It's the right thing to do," "Show me the evidence" because
physicians being reductionists, sort of people that they are,
are looking for evidence.
Now obviously, there is evidence and I think what Hilary
[Llewellyn-Thomas] said right before launch was spot on,
that we sit now at a time where with things like the Cochrane
Review, with the evidence that's growing in literature about
decision quality, by lots of measures we are at a much
different place than we were when the President's Commission put
forth this ethical notion. But at the same time,
the evidence may not be where we want to put all of our chips.
Even the evidence around the impact on cost,
and this is a paper that got a lot of attention last year.
David Wennberg using a foundation decision aid and telephone
coaching was able to show improvements in both decision
quality, but also reductions in cost at a health plan level.
It's this kind of work and this kind of thinking in positive sense
has driven all the interest and shared decision making at a policy
level - why it's in the Accountable Act,
why organizations like Palo Alto Medical Foundation and group
health are investing in shared decision making.
Not because it's the right thing to do;
because they think it's going to save them money.
Again, the problem is if it doesn't wind up saving money,
does it mean we shouldn't do it?
So, a lot of this rests on the important notion of preference
sensitive care, which all of you are familiar with.
And I put this up here just to remind ourselves that in
articulating this notion around practice variation,
Jack Wennberg and others who are in that field talked about three
sort of categories of causation for practice variation - those who
were driven by effective care, those who were driven by supply
sensitive issue; so if you build more, you know,
radiology units you do more x-rays, more CT scans,
and preference sensitive.
And following that reason, that notion was that shared decision
making became important for preference sensitive conditions
and that's what's fueled the rise of decision aids,
but what about the other areas?
Does that mean that shared decision making is less important,
and I think this is really the question that we're trying to
deal with. So, what if the evidence is low?
So, with that background, I'm going to play around with you with
this diagram, which is intended to show sort of on this axis here
sort of - I don't know - building strength of the evidence argument
and along this axis the building strength of the sort of preference
argument for any given decision.
And I would start by saying that for preference sensitive
conditions and sort of restating that argument would be to say that
for back surgery, for a herniated disk,
there are any number of things you could do and that we'd rather have
it be driven by the preferences of the back surgeon to be driven by
patient preference. I've just sort of restated the preference
sensitive decision argument.
And shared decision making, that's sort of viewed right now as a
sweet spot for shared decision making,
the area where you can make a difference in terms of reducing
unwarranted variation, enhancing decision quality,
etc., etc. and that's where most of the work is going.
The question becomes what about these other areas?
What about when the evidence is really strong and patients don't
really have much in the way of values and preferences,
or when values and preferences are really strong but evidence is low
or this really tough spot down here where patients don't really
have strong preferences that are informed by anything and we don't
have a lot of evidence.
And parenthetically, by lack of evidence,
I'm thinking both about areas where there's tons of evidence,
but lots of uncertainty to pick up on the theme from earlier,
but also areas where there's really just not much evidence,
a situation where the physician might say, "You know what?
No one really knows what to do in this situation."
So, in some earlier work, we did some focus groups with patients
and we were trying to sort of think about whether all decisions
are created equal and mostly from the standpoint of complexity.
And Ron [Epstein] sort of stole my thunder on this,
but the notion was that through focus groups,
we found that both patients and physicians agree that there are
some kind of decisions that seem very straightforward for which the
impact on the patient might be minimal,
there's a clear consensus that it seemed like a reasonable thing to
do, you know, various straightforward outcomes and it
might be something like screening cholesterol.
There are others on the other end of the spectrum that were very
complex - extensive impact on the patient,
lots of controversy around what's the right thing to do,
outcomes may be uncertain, and, you know,
surgery for a rare condition might be in that situation.
And so, the notion was should shared decision making look the
same in those three? We essentially said no.
We suggested that you might have gradations of shared decision
making where if as measured by elements of a shared decision
making model you might say for a basic decision maybe you only have
to talk about a couple of elements and then all the way up to a
complex decision where you might have to talk about more.
So, again, the underlying premise is that shared decision making is
relevant to every clinical decision. It's just relevant
in a slightly different and decision-specific way.
So, the first answer to the question would be in situations
like cholesterol screening that you might characterize that as
basic shared decision making for which you might imagine sort of a
more streamlined form of shared decision making,
but shared decision making nonetheless.
So, Alex Kon wrote this interesting article about, again,
along a continuum of shared decision making characterizing one
extreme, and someone else brought this up earlier - patient or
agent-driven decisions; that is ones that are really owned by
the patient. We're just sort of, as physicians, innocent
bystanders and the other extreme where physicians
drive decisions. He suggested there's this continuum
of patient and physician responsibility.
In the middle would be sort of the kind of equal partner form of
shared decision making, but he suggested that there might exist
these two ends of a continuum.
So, you'd have situations where maybe this might be where evidence
is low, but preferences are high, patients really care about what
they want to do and want control and physicians don't really have
any reason to care or know which direction to point them and on the
other hand, it might something like the cholesterol screening.
So in this frame, I've renamed this basic or physician driven
shared decision making and suggest that there's other category,
maybe Tony [Back] will talk about this from his own work,
of patient driven shared decision making and the situation of
low evidence.
An example might be navigating an end of life discussion with
patient and family where there are lots of choices to be made.
Sometimes there's evidence, for example,
around tube feeding in people patient with advanced dementia or
life sustaining treatments and CPR in people with advanced metastatic
cancer, but lots of times there's really no evidence.
It's a matter of trying to elicit those preferences and trying to
value them. So, what about this final quadrant?
I would say that the four ideas, again,
were this notion of shared- all around the intrinsic value of the
process not about the outcomes - creation of a shared mind,
closing the knowledge gap, promoting empowerment and trust
through transparency.
So, the shared mind concept we've heard about already,
about how uncertainty has these different components.
And again, I guess the assertion I would make is that shared decision
making in these low evidence situations can still reduce
personal uncertainty.
Obviously, we know from the work that many of you have done about
reduced and ways that may at least influence decisional conflict.
But, this interesting study, which some of you probably saw,
which I thought was fascinating, was looking at personal
uncertainty in over a hundred dyads of patients and physicians
and looking at the role of information in reducing or
increasing personal uncertainty.
Essentially what they found, to sort of paraphrase,
is that more information had an impact on decisional or personal
uncertainty and that there was an interrelationship, a connection,
a correlation between the directionality of personal
uncertainty between the physician and the patient; said again,
that sharing that uncertainty had an impact.
So, at least to me, this is some evidence of the notion of just
going through the process of acknowledging and sharing
uncertainty may contribute to some interconnectedness to this notion
of this dyadic dialogue that needs to take place.
So, what about the closing the knowledge gap?
Well, I don't really have any evidence to show except that it
seems to make sense to me that the notion of closing a gap between
patient and physician, which always starts out large,
except maybe in the rare instance where the patient is a physician
and has just as much knowledge if not more than the real physician,
but there's always a closing of the knowledge gap.
And again, does that influence outcomes?
I don't know, but it's probably of some intrinsic value because it
helps the patient to be more informed.
If that's the goal then it certainly helps.
We know that one component of decision quality is knowledge and
I think there's evidence that decision aid is another effort
that shared decision making can close that knowledge gap.
So, this is the second way in which from a process standpoint,
just the process alone even in low evidence situations.
So, if these two are situations where the evidence is high by
whatever measure you want to use, before shared decision making
and after. There's still a knowledge gap, but it's less.
So, this is just a schematic to show the same thing can occur in a
low evidence situation. What about patient empowerment?
So again, patient empowerment, separate from whether they
actually know or can remember or answer correctly on a test the
knowledge that's been shared or the risk estimates that have been
shared. The fact of receiving that knowledge contributes to a
sense of control and self efficacy, which are core components
of empowerment.
So, even if they don't get the answers right on the test,
having received that knowledge is of relevance.
I don't have time to talk about the notion that some have touched
on already, the notion of gist knowledge,
that whether our emphasis as physicians and our emphasis as
scientists on a particular way, a particular epistemology of
knowledge versus knowledge that may be more at a gist level might
be just as good if not better.
Finally, the notion of promoting trust and I was influenced very
early in my career by a book by Jay Katz,
The Silent World of Doctor and Patient,
and many of you know that he was a psychiatrist and brought into this
area around shared decision making,
some interesting concepts from psychiatry about transference and
counter transference.
The essential notion he brought forth was that there's blind trust
between the patient and physician and earned trust.
Blind trust is the physician says, "You clearly need treatment A.
You know, I've been practicing for 20 years and I've seen hundreds of
patients. I can tell you this is for you."
And Mary was talking about that, and then you say, "Okay.
Okay" and you get home and you think, "You know,
I didn't ask about this or that or the other thing."
And so, that trust is set up to be very frail,
frail because of either unmet expectations or the agency of the
person is lost in that encounter versus earned trust,
which is more of a mature relationship in which the trust is
earned by trustworthy behaviors.
The trustworthy behaviors in this instance being shared decision
making and patient centered communication.
Again, this is an interesting study of looking at the
relationship between different components of patient centeredness
and I'll draw your attention to the fairly high correlation
between trust and shared decision making.
I mean it's not huge, but there is a statistically significant
relationship between, and they were reported behaviors,
shared decision making behaviors and ratings of trust on a
validated instrument and found a correlation.
I'm just throwing you little tidbits here just to make the
point that sharing through transparent sharing of uncertain
information, moving from blind trust to earned trust is another
piece of this intrinsic value of shared decision making.
So, all said in this schematic, in all four quadrants,
there's an argument to be made for shared decision making that's
separate from evidence, that's separate from outcomes,
that rests on acknowledging that there's some intrinsic value
that's met even in low evidence situations.
So, I suggest that there's a strong ethical foundation that
shared decision making, again, it can promote a shared mind through
uncertainty, getting at the same place with regard to uncertainty.
That has intrinsic value.
Closing the knowledge gap has intrinsic value.
Promoting patient empowerment and promoting trust.
I think therein lies some important threats and perhaps a
framework for thinking about this question we've been talking about
earlier, which is how to think about the physician side of
shared decision making. Thank you.