I completely agree with Brian and Grant that unplanned subgroup analyses
with small sample sizes would have low sensitivity and specificity in
providing 'true' answers to clinical questions. But imagine a database that
had an infinite number of patients. Would this not resolve the issue of
power? Having said that, until we do have these databases, then the results
will be fishing for answering, and I disagree with this concept because the
probability of a type II error is high.
Also let's take a real-life situation that occurs everyday. When a patient
is in need of an organ transplant the process for deciding who is a 'match'
is very exhaustive to increase the probability of a successful outcome.
These decision support systems help to guide decisions, but don't make
decisions on their own; just as guidelines are just that... guides to proper
decision making.
In the end, as we have seen with many new technologies, if there is any
truth behind the assumptions then we as clinicians and researchers will be
forced to accept the concepts or be left behind. I expect that this won't
happen for many, many years.
Sincerely,
Ahmed
_______________________________________
Ahmed M. Abou-Setta, MD, PhD
Independent researcher/ systematic reviewer
Principal Evidence-based Medicine Consultant, PharmArchitecture Limited (UK)
Consulting Senior Systematic Reviewer, Evidence-based Practice Center (EPC)
Ottawa Health Research Institute (OHRI), University of Ottawa
Member, Editorial Advisory Board, The Open Medical Devices Journal
Member, Geneva Foundation for Medical Education & Research
Member, Menstrual Disorders & Subfertility Subgroup, Cochrane Collaboration
5, El-Ashwal Street, from Pyramids Road
Maashal Station, Pyramids
Giza, Egypt
Mobile: (+20-11) 852-0067
E-mail: [log in to unmask]
--------------------------------------------------
From: "William Grant" <[log in to unmask]>
Sent: Monday, March 23, 2009 2:52 PM
To: <[log in to unmask]>
Subject: Re: A recent article for your comments
> I agree with Brian that this is an attempt to develop a subgroup analysis
>
> The caution that is appropriate for any investigation which reports an
> unplanned subgroup analysis
> is that it is likely that the original design was not powered on the
> subgroup level. That is, original
> sample size, effect size and power were predicated on the overall sample
> not on individual subgroups.
> And, as I hope we all know, post-hoc power analysis is not very reliable.
>
> It is relatively easy to predict behavior of large groups given prior
> information on their behavior.
> Even so, it is very difficult to predict the behavior of a single
> individual within the group even given prior information on the group
> behavior.
>
> This seems to be an interested mathematical curiosity with little clinical
> implication for individual patients.
>
> William D. Grant, Ed.D.
> Executive Director, Center for Emergency Preparedness
> Professor, Emergency Medicine
> SUNY Upstate Medical University
> 750 E. Adams St. EmStat
> Syracuse, NY 13210
> 315-464-4365 (p) 315-464-6220 (f)
>
> STATEMENT OF CONFIDENTIALITY
> The information contained in this electronic message and any attachments
> to
> the message are intended for the exclusive use of the addressee(s) and may
> contain confidential or privileged information that is protected by law.
> If
> you are not the intended recipient, please notify the sender at (315)
> 464-6997 or at [log in to unmask], and destroy all copies of this message
> and any attachments. This communication may not be forwarded without the
> express written permission of the original author.
>
>
>>>> Brian Alper MD <[log in to unmask]> 3/22/2009 7:36 AM >>>
> On the first read it sounds like a very sophisticated method for
> subgroup analyses.
>
>
>
> To some degree subgroup analyses help suggest better fits between the
> data and subsets of patient populations.
>
>
>
> Taken to the extreme (finding the single person in the data that is most
> like your patient) would be like using a case report.
>
>
>
> Sounds like there could be some discovery from the methods described but
> it is hard to picture how it would replace our current methods for
> accepting scientific research and applying it to patients.
>
>
>
> --------------------------------------
>
> Brian S. Alper, MD, MSPH
>
> Editor-in-Chief, DynaMed (www.DynamicMedical.com
> <http://www.dynamicmedical.com/> )
>
> Medical Director, EBSCO Publishing
>
> 10 Estes St.
>
> Ipswich, MA 01938
>
> office (978) 356-6500 extension 2749
> cell (978) 804-8719
> fax (978) 356-6565
> home (978) 356-3266
>
> "It only takes a pebble to start an avalanche."
>
>
>
> This e-mail and any attached files transmitted are confidential and
> solely for the use of the intended recipient. It may contain information
> which is covered by professional or other privilege. If you are neither
> the intended recipient of this e-mail nor the person responsible for
> delivering it to the intended recipient, be advised that you have
> received this e-mail in error and that any use of it is strictly
> prohibited. Please notify the sender immediately by reply e-mail and
> then delete from your system. EBSCO Industries, Inc., its subsidiaries
> and divisions, accept no responsibility for any loss or damage suffered
> by any person arising from the use of this e-mail.
>
> ________________________________
>
> From: Evidence based health (EBH)
> [mailto:[log in to unmask]] On Behalf Of Rakesh Biswas
> Sent: Sunday, March 22, 2009 6:05 AM
> To: [log in to unmask]
> Subject: A recent article for your comments
>
>
>
>
>
>
>
> http://www.casesjournal.com/casesjournal/article/view/5051/103
>
>
>
> would be grateful for your comments.
>
>
>
> rakesh
>
>
>
|