DOC News October 1, 2007
Volume 4 Number 10 p. 5
© 2007 American Diabetes Association
RESPONSE FROM DR. HIRSCH:
Dr. Staley, you bring up an important point. The study team led by
Cleveland Clinic cardiologist Dr. Steven Nissen used a now-common technique
called a meta-analysis to reach its conclusions. This is just one strategy for
combining the findings from independent studies.
The final result of this statistical tool is to provide a precise estimate
of treatment effect, giving due weight to the size of the different studies
included. So individual events have to be considered based on the number of
individuals in each trial, not as a single event, which would be calculated in
one large single study.
As we are now learning, the type of meta-analysis is also important. As
noted by Dr. George A. Diamond et al., the statistical tool used in the
analysis, in addition to inclusion criteria, can completely change the results
of any given
meta-analysis.1
Footnotes
Watch for an upcoming guest editorial on meta-analysis particulars and
approaches by Dr. M. Sue Kirkman, American Diabetes Association vice president
for clinical affairs, in the November issue.
References
1. Diamond GA, Bax L, Kaul S: Uncertain effects of rosiglitazone on
the risk for myocardial infarction and cardiovascular death. Ann Intern
Med [Epub ahead of print August 16] 2007. Available online at
http://www.annals.org/cgi/content/full/0000605-200710160-00182v1.
Accessed August 29, 2007.
CiteULike 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
