Hi
Annabel
Thanks
for some thought-provoking questions. I will respond with some background before
commenting on your blogs and answering your question.
My
interest in the limits of common sense arises from a difficulty we often have at
CKS. We have to justify all our recommendations, and always struggle to find
sciencey sounding words to explain common sense recommendations. For example, we
recommend that people with Raynaud’s syndrome wear cheap gloves to prevent their
hands getting cold rather than waiting until their hands get cold and then using
expensive electric hand warmers. We have been reluctant to say this is common
sense, because (a) it doesn’t sound very professional, and (b) we know that
common sense can sometimes be disastrously wrong.
I
think that I may have worked out when you can safely use common sense (but am
still looking for the sciencey sounding words to say this in our “Basis for
recommendation” sections).
To
explain the limits of common sense I need first to explain (superficially) how
decision-making works.
The
theorists have a simple algorithm for making the optimum decision for
treatments:
1.
List
all the treatment options — don’t forget the option to do
nothing
2. For each option, list all the outcomes that would influence
the decision, and work out each option’s expected value (the probability of the
outcome times its value (positive for benefits; negative for
harms)).
3. For each option, work out its expected value (the sum of
the expected values for each outcome).
4. Your best bet is to go with the option with the largest
expected value — every decision is a gamble.
Unfortunately,
life is a little more complicated than this.
First complication: We (almost) never have all the
information.
Second complication: Our heads don’t work like this when making a decision. They
automatically, and without thinking, find exactly 1 good reason to prefer one
option, and ignore the all the other reasons for and against. Our heads just
don’t like to consciously work out all the different values and probabilities of
all the pros and cons. Thankfully this is seldom a problem, because most of the
time our heads are very good at selecting what is a good reason to do something
(see Gerd Gigerenzer’s book Gut Feeling).
When
we use common sense we use the 1 obviously good reason to make a decision,
without critical appraisal of the evidence and reasoning for the decision.
This saves us the trouble of checking that what we imagined to be a good
reason isn’t actually a bad reason: maybe the treatment doesn’t work, or doesn’t
work as well as another option, or maybe there could be unexpected serious
adverse effects.
We
rely on common sense all the time; decision-making would be paralyzed if we had
to check everything every time. What is amazing is that common sense doesn’t
lead us into trouble more often.
Because
we have lots of evidence that common sense does sometimes lead to disaster, we
should be asking ourselves more often when we can and when we shouldn’t rely on
common sense. (I know of no work that is being done in this area apart from Gerd
Gigerenzer and his colleagues.)
If
we follow the algorithm for optimising decision-making when using common sense,
we should assess its expected value. The expected value of a common
sense-based treatment decision is:
The value of the benefit times its probability
less
The costs of any adverse effects times their probabilities
Because
we don’t have all the information needed to calculate the expected value of
common sense, we have to use our common sense.
(Sorry,
I can’t resist using recursion whenever there is an
opportunity!!)
To
come to your examples:
You
are right and safe to say that it is common sense not to recommend e-cigarettes.
We have lots of experience of nicotine, so we know its effects, good and bad.
Common sense tells us that the device is likely to be about as safe as other
forms of nicotine delivery — I can’t imagine it bursting into flames — and there
is little reason to think it is more effective. It would be against common sense
to invest in the company that makes the device, even if they have a very
creative marketing department. This is because (a) they did cruel experiments on
poor little mice for no good scientific reason (they probably had to do the
experiment to get a marketing licence). And, more seriously, (b) they have a job
to do to show that it works in practice. You can just imagine someone,
desperately craving for a cigarette, reaching for their e-cigarette and finding
that the bloody battery is flat. They would relapse straight
away.
You
are right and safe to suggest trying the warm version of the cold remedy. The
warm version probably tastes better, is unlikely to be more risky, and the
additional contribution to your carbon footprint from putting the drink in the
microwave is probably less than that generated by the last search you did on
Google.
So,
to answer your last question, common sense can be both useful and dangerous —
but (and this is common sense) not at the same time. After all, it is common
sense that you should trust your bank to look after your money. It is common
sense that you should trust the review written by the experts. It is common
sense that you should trust your doctor.
Michael
From: BENTLEY,
Annabel [mailto:[log in to unmask]]
Sent: 14 January 2009
16:11
To: Michael Power;
[log in to unmask]
Subject: Example - electronic
cigarettes? RE: Evidence that common sense can be dangerous
Hi
Michael
I like your idea that common sense can be dangerous.
But, is
it more complicated than that?
- eg: on my blog I recently argued that
clinical decisions about "electronic cigarettes" should not be based on common
sense (ie agreeing with you; http://tinyurl.com/5gzv4p , you may need to
login to see comments).
- but, on the other hand, I've also recently argued
that common sense can be useful http://tinyurl.com/8zybr7 [from
Eccles and Sanu, Rhinology 2008 http://www.rhinologyjournal.com/article/view_abstract.php?mgzn_id=148&rtcl_id=811 ]
So,
I've either undermined my previous argument, or common sense can be both
useful and dangerous. What do you think? Should I change my
thinking?
best
wishes
Annabel
------------------------------------------------------------------------------------------------
Dr
Annabel Bentley | Assistant Medical Director | Group
Medical
Bupa,
15-19 Bloomsbury Way, London WC1A 2BA
T: 020 7656 2069 | F: 020
7656 2708
www.linkedin.com/in/annabelbentley
Please
consider the environment before printing this email.
-----Original Message-----
From:
Evidence based health (EBH) [mailto:[log in to unmask]]
On Behalf Of Michael Power
Sent: 14 January 2009 12:40
To:
[log in to unmask]
Subject: Evidence that common sense can
be dangerous
A little while ago, Richard Lehman said in his weekly
review, and in his inimitable way: "Not very long ago, I could confidently
assert that the best way to manage atrial fibrillation was to stop the atrium
fibrillating. Obvious, isn’t it? And, like so many obvious ideas in medicine,
wrong."
I would like to build a list of obvious ideas in medicine that
have been dangerous; for example, advice to put babies to sleep on their
tummies.
Suggestions (preferably with references!!) for my list would be
much appreciated.
Thanks very much.
Michael Power
Clinical
Knowledge Author, Guideline Developer and Informatician Clinical Knowledge
Summaries www.cks.library.nhs.uk
Feel better
with Bupa
To find out
what Bupa can do for you, visit http://www.bupa.com
Bupa House
15-19 Bloomsbury Way London WC1A 2BA
Internet
communications are not secure and therefore Bupa does not accept legal
responsibility for the contents of this message. Any views or opinions presented
are solely those of the author and do not necessarily represent those of Bupa.
Bupa
Insurance Limited, Bupa Health Assurance Limited, Bupa Insurance Services
Limited and Goldsborough Estates Limited are authorised and regulated by the
Financial Services Authority.
For a list of
Bupa's main UK trading companies visit
www.bupa.co.uk/html/statements/trading_addresses.html