Before trying to use this study to support or refute resource allocation decisions I'd caution about considering the difference between clinical outcomes (outcome of current respiratory symptoms, or symptomatic COPD) and surrogate outcomes (spirometric measurements consistent with COPD but not obtained in the clinical context of patient suffering)
Associations with surrogate outcomes only should be considered part of the story in development, not the "answer" to guide such decisions. Perhaps smoking has a greater impact on clinical ailments and resource use for COPD (assuming you have a forced choice and have to quantify relative impact) but the clinical impact is not necessarily through the spirometric measurements.
Brian S. Alper, MD, MSPH
Medical Director, EBSCO Publishing
[log in to unmask]
DynaMed Weekly Update - Timely information on articles most likely to change clinical practice in a FREE weekly e-newsletter from the DynaMed Editors. Earn FREE CME for reading the update.
From: Evidence based health (EBH) [mailto:[log in to unmask]] On Behalf Of Paul Elias
Sent: Wednesday, June 22, 2011 12:55 PM
To: [log in to unmask]
Subject: Re: Help with appraisal of a cohort study please
was there a reason why the cohort was not followed for the full period and then
the attempt at association made in 2002, looking back at the early life factors?
I do not understand why they would pool the measurements... can you explain a
bit more from the methods....??
Paul E. Alexander
----- Original Message ----
From: Anne Gray <[log in to unmask]>
To: [log in to unmask]
Sent: Wed, June 22, 2011 12:36:14 PM
Subject: Help with appraisal of a cohort study please
Can I ask for your collective thoughts on a paper published in Thorax please?
The study aims to use a cohort study to identify associations between childhood
factors and reduced lung function and therefore risk of COPD. Interviews
identidied a range of early life factors in a multisite cohort (n=13,359)of
participants age 20-45 in 1991-1993, then repeated the measurements 9 years
later in 7738 of the participants.
The two sets of measurements have been added together, giving n= 16,832 to make
associations between childhood factors an dreduced lung function.
I am concerned that they are counting some people twice to make the association.
Is this appropriate?
Interestingly there is no signficant association between these supposed
childhood factors and decline in lung function over time (ie in those measured
THis does have implications - if childhood factors are more important that
current smoking rates we may have to rethink stop smoking services.
I have tried to contact the author but no response so far.
I am happy to be told that this is appropriate and my understanding is wrong. I
would just like to be able to explain the study to my Stop Smoking lead.
The article is:
THorax 2010 v65 p14-20
Early life origins of COPD
Svanes et al
NHS Milton Keynes