Dear Ben,
You will get the answer to your computer question at:
http://www.top500.org/lists/2010/11/press-release
I don't know whether you have read recently about IBM's Computer 'Watson'
beating the best players on the 'Jeopardy' TV programme in the USA:
http://www.pcworld.com/article/220056/watson_beats_jeopardys_best_whats_next.html
Apparently, this machine costs 3 million dollars today, but by 2020, we will be
able to get them for a few thousand dollars each.
Regards,
Ash
Dr Ash Paul
Medical Director
NHS Bedfordshire
21 Kimbolton Road
Bedford
MK40 2AW
Tel no: 01234897224
Email: [log in to unmask]
________________________________
From: "Djulbegovic, Benjamin" <[log in to unmask]>
To: [log in to unmask]
Sent: Thu, 24 March, 2011 23:49:55
Subject: Re: Clinical Decision Making and Diagnostic Error
perhaps we identify with the conscious but not with the non-conscious because we
consider technological and other advances as an integral part of our logical,
conscious system...but I don't believe that with all inventions at our disposal
we are close to processing 10,000,000 bits of information per second (and
without overheating!)
BTW, what is the fastest modern computer in terms of processing power?
ben
________________________________________
From: Jim Walker [[log in to unmask]]
Sent: Thursday, March 24, 2011 4:13 PM
To: Djulbegovic, Benjamin; [log in to unmask]
Subject: Re: Clinical Decision Making and Diagnostic Error
The non-conscious system is massively parallel and processes something on the
order of 10,000,000 bits of information per second, the conscious between 16 and
32 bits.
So the conscious is something like the tip of a fully integrated iceberg,
dependent on the pre-processing of disparate inputs by the non-conscious system
for the highly (and necessarily) filtered information which it manipulates.
So while the conscious can direct the attention of some aspects of the
non-conscious system to some extent, there is no theoretical or empiric reason
to think that the conscious system can "train" the non-conscious--or that we
would benefit if it could. (Interesting how we identify with the conscious but
not with the non-conscious.)
Jim
James M. Walker, MD, FACP
Chief Health Information Officer
Geisinger Health System
>>> "Djulbegovic, Benjamin" <[log in to unmask]> 3/24/2011 3:31 PM >>>
Hi Amy,
I am at the moment reading some papers by Read Montague (from Baylor)...
He has written some interesting stuff...
Best
ben
From: Dr. Amy Price [mailto:[log in to unmask]]
Sent: Thursday, March 24, 2011 3:13 PM
To: Djulbegovic, Benjamin; [log in to unmask]
Subject: RE: Clinical Decision Making and Diagnostic Error
Dear Ben ,Neal and all,
I am interested in the fMRI data identifying discrete brain areas for systems 1
and 2. I am particularly intrigued by how/when they link and if the unconscious
bias might be trained . Do you have authors names, links or papers you could
share. I am looking at this area in regards to addiction and collaborating with
others who are identifying genomics. Presently exploring QEEG to see if this
uncovers anything of interest...
Best regards,
Amy
From: Evidence based health (EBH) [mailto:[log in to unmask]]
On Behalf Of Djulbegovic, Benjamin
Sent: 24 March 2011 02:55 PM
To: [log in to unmask]
Subject: Re: Clinical Decision Making and Diagnostic Error
Neal,
Thanks for these insights - it is a really interesting thought about the fMRI
data identifying discrete brain areas as location of system 1 and system 2 (at
the moment, I am too reading on this fascinating stuff, and after you pointed
this out, I wonder what would Hammond say about the empirical data that seems to
be falsyfying his thesis?)
Regarding which reasoning system gets its more right than wrong, I think we are
saying the same thing, except (that in my reading of Hammond he seems to be
saying) that when we get it wrong, the consequences of relying on intuition vs.
logic are of the magnitude order less significant? (We are, of course, talking
about decisions related to human affairs and social policies, not to the stuff
in the realm of the quantum physics!)
Best
ben
From: Evidence based health (EBH) [mailto:[log in to unmask]]
On Behalf Of Maskrey Neal
Sent: Thursday, March 24, 2011 2:04 PM
To: [log in to unmask]
Subject: Re: Clinical Decision Making and Diagnostic Error
Ben
Good to hear from you. Spring's arrived in England and it's lovely, but it sure
ain't Florida!
My reading is slightly different. From that, neither system 1 nor system 2 are
superior, but system 1 is our "default". If we are talking about errors of
commission, then then Pat Coskerry shows these occur predominately in system 1
due to the common cognitive and affective biases. My own personal dramatic
memories are fortunately few in number, but when I got it wrong the usual
contributing factors were availability bias and affective - notably work
overload and sleep deprivation. And I'm sure Hammond isn't right about us never
being spectacularly wrong in system 1. In system 2 the common problem is
omission - if a common medical emergency presents it isn't optimal to have to
spend lots of time working out the diagnosis and treatment plan. Broadly
speaking, as Pat Croskerry says, we're safer in system 2 - but we're a lot
slower.
The MRI emission data identifies particular and quite different areas of the
brain being used live with system 1 and system 2 decision making and that's
harder to explain as a continuum - happy to debate. We've obviously "cartooned"
some of this necessarily in these brief, awareness-raising group discussions and
one of the facets we've not mentioned is the frequent live toggling in decision
making between system 1 and system 2.
The key trick is to train people to step back and re-examine the issue...both
from logical (system 2) and affective side (system 1)...a tough to do when
decisions have to be made in a short time-frame...
Absolutely. The next step is to prove teaching these approaches changes
behaviour, and then that change in behaviour improves the processes used in
decision making if that's possible. Showing improved outcomes or less errors or
both would be ideal, but if we're honest the definitive literature showing
"teaching EBM" or for that matter teaching anatomy or teaching genetics improves
outcomes isn't massive either. The quick "could this be anything else" or "are
there any other / better treatment options" or "did I check for
contraindications / interactions" calibration at the end of the consultation is
simply plain old fashioned good medical practice, so if this is well presented
its (a) an interesting subject to teach and learn, and (b) some of the content
has excellent face validity for students. Likewise the more leisurely reflection
- but then maybe the St Pete police would want drivers to be concentrating on
the road!
I keep hearing about on going research which might move all of this forward, and
there's certainly lots of interest. Fabulous.
Best
Neal
Neal Maskrey
National Prescribing Centre
Liverpool UK
________________________________
From: Evidence based health (EBH) [mailto:[log in to unmask]]
On Behalf Of Djulbegovic, Benjamin
Sent: 24 March 2011 11:02
To: [log in to unmask]
Subject: Re: Clinical Decision Making and Diagnostic Error
Neal, a challenge is, of course, to know when to use our intuitive (system 1)
vs. logical (system 2) inferential process...Kenneth Hammond, who has promoted
idea of a continuum between system 1 and system 2 (instead two sharply
demarcated systems), has made a point that if we rely on the system 1, we will
often be more wrong than right, but never spectacularly wrong. On other hand,
reasoning based on system 2 (this would include EBM) would be more often right,
but when it is wrong it can be phenomenally wrong (as our models of the world
can be hugely mistaken)... A practical/educational corrolary from this is "rely
on EBM, but if it does not agree with your intuition, don't go there
go/re-examine your decision..." The key trick is to train people to step back
and re-examine the issue...both from logical (system 2) and affective side
(system 1)...a tough to do when decisions have to be made in a short
time-frame...My own approach has been to ask students/residents/fellows is a) to
review mentally all patient'stories ("cases") at the end of the day and focus on
any possible oversights/errors that may occurred that day (e.g. while driving
back home), b) ask yourself " if I were the patient, would I like to be taken
care by thy guy/girl like me?"
I have been doing this for years- wish I can report that I found the
"solution"...unfortunately, I still find myself much deficient as I try to
answer these questions...but resoluted to do better next day...
ben djulbegovic
