Hi all,
I've been asked to look at analysis that has been carried out on a case control study. Altogether there are 30 cases, each of which has been matched to 1-3 controls and a conditional logistic regression carried out (using the 'clogit' function in R) to determine the influence of particular factors/exposures on an individuals odds of becoming a case. Initial analyses showed exposure to a particular foodstuff (milk) as highly predictive of infection status.
24 of the cases were lab confirmed. When the analysis is repeated on this subgroup and their controls, the software gives a message saying that the iterations have not converged and then the confidence interval for the coefficient for milk that was significant in the original analysis ranges from 0-Inf. Can any conclusion be drawn from this or does it mean that the problem is intractable? Eyeballing the data does not shed light for me as the combinations present within the case control sets are:
control(s) case total with this configuration
no.exposed to milk 1 1 6
0 1 13
0 0 4
1 0 0
2 1 1
2 0 0
So there are 7 sets in which both the case and at least one control have been exposed to milk, 4 sets in which neither case nor control(s) were exposed and 13 in which only the case was exposed.
I'd be interested to hear any thoughts on this as it is not an area I have experience in and I'm not sure what conclusions can be drawn. Obviously, it is a very small dataset and the removal of 6 non lab-confirmed cases and their corresponding controls represents a significant drop in sample size.
Many thanks for your help,
Kate
Kate Hunter
Statistician
HSE - Health Protection Surveillance Centre
25-27 Middle Gardiner St
Dublin 1
IRELAND
We are part of the new Health Service Executive (HSE). As an interim measure all emails we send will have the structure of [log in to unmask]
*************************************************************************************
The information contained in this e-mail transmission may be confidential. It is intended only for the addressee(s) stated above.
If you are not an addressee, any use, dissemination, distribution, publication, or copying of the information contained in this e-mail is strictly prohibited. If you have received this e-mail in error, please immediately notify us by telephone at
353-1-8765300 or e-mail [log in to unmask] and delete the email from your system.
This footnote also confirms that this email message has been scanned for the presence of computer viruses.
************************************************************************************
|