The challenge is getting a software to read data, forest plots and charts etc. and then to integrate that to existing material.
I think personally this is doable but it starts with how the Cochrane (CR) is written in the first place. Maybe the CR has to be redone in a way that could accomodate an update in the future, in tandem with Revman. For example, the CR can have its routine sections yet have the option to produce an update in the future. The key is to allow the updating person to access the prior Revman file. I would do it this way (but only some basic thoughts):
1.) Allow a CR update interface...Cochrane can do this
2.) in this updating interface that is linked to the prior SR, the updating folk can enter an addition or rational for the update (not a complete background) e.g. "Since the last CR published in 2010 by xxxxxxxx, new RCT evidence has emerged that may impact the overall summary findings".....
3.) the update can then enter a section, maybe a para on the methods used e.g. search period, strategy etc. I mean both the prior and new search etc. will be printed in the updated SR for continuity
4.) the updating folk can then have access to the prior RevMan file and enter the new studies...the output should be the prior FP, the new studies FP, and a combined FP of all studies so the reader can see the change in summary estimates, heterogeneity, funnel plots, and even sub group or sensitivity analyses etc.
5.) And finally, the updating folk can look at the new results and write a para in an updated discussion section on how the finding change. Again, the prior discussion and conclusion sections should be in place, but have an added discussion and conclusion section, labelled as such.
Just some thoughts...
Best,
Paul E. Alexander
--------------------------------------------
On Thu, 7/18/13, healingjia Price <[log in to unmask]> wrote:
Subject: Re: Using NLP for critical appraisal
To: [log in to unmask]
Received: Thursday, July 18, 2013, 2:31 PM
Hi
Teresa,
I am updating one as we speak and sifting through
articles to decide what to pull. Which software are
you thinking of as I could do au natural and put the
same papers through automated language systems and
report back on how it works. I am interested in elements of
automating for doing participatory research on questions
patients public are interested in but are not likely to get
funding for. I think conceptually it is a great
idea!
If anyone has software preference for this or if
there is one that is low cost and free that would be
awesome!
Thanks
BestAmy
Amy Price Empower 2 Go Building
Brain PotentialHttp://empower2go.orgSent
from my iPad
On 18 Jul 2013, at 08:47 AM, "Benson, Teresa"
<[log in to unmask]>
wrote:
I’m wondering if anyone
out there has toyed with using natural language
recognition/programming software to “read” journal
articles and apply elements of a critical appraisal
checklist. While it wouldn’t replace humans, it
might be interesting
to see if it’s helpful for screening large numbers of
articles when doing systematic reviews. Thanks,
Teresa Benson, MA, LP
Clinical Lead,
Evidence-Based Medicine
McKesson Health Solutions
18211 Yorkshire Ave
Prior Lake, MN
55372
[log in to unmask]
Phone:
1-952-226-4033
Confidentiality Notice:
This e-mail message, including any attachments, is for the
sole use of the intended recipient(s) and may contain
confidential and
privileged information. Any unauthorized review, use,
disclosure or distribution is prohibited. If you are not the
intended recipient, please contact the sender by reply
e-mail and destroy all copies of the original
message.
|