[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
RE: st: Interpreting conditional logistic regression equations using 2 similar types of matching.
Thank you for your reply. Let me elaborate a bit.
This is not a longitudinal study. These children come to emergency after
having had a seizure and part of the workup is measurement of hemoglobin
(plus other stuff). We want to see if anemia is related to seizures, so we
are pairing the child with a seizure with a similarly aged child without a
seizure who has come to emergency for another reason. Thus this is a paired
case/control design with the match being made on age. The problem is that
if we match loosely on age (within 3 months) as opposed to a tight match (1
month) we get different results and I am unsure which result is 'best'.
[mailto:email@example.com] On Behalf Of Ronán Conroy
Sent: February 7, 2007 3:28 AM
Subject: Re: st: Interpreting conditional logistic regression equations
using 2 similar types of matching.
On 6 Feabh 2007, at 23:40, Donald Spady wrote:
> This may be a second posting. The first one did not seem to go
> through and
> on reading the FAQ I realized that I might have sent non-plain
> text. (I
> copied the output from Stata using the copy table command rather
> than the
> copy text command. Hopefully this one will work.)
> I am looking at the effect of iron-deficiency anemia on the
> likelihood of a
> child having a seizure.
I'm not sure I understand this design. Children can have more than
one seizure, so is your case variable whether the child ever had a
seizure or never had a seizure, or are you doing a nested case-
control and matching a child in whom a seizure occurred with a child
of the same age in whom one did not occur?
> I am using clogit but have a problem in that I have
> matched my case/control "group" variable (age) as either very tightly
> matched (same month: "tight") or more loosely matched (up to 3 months
> difference: "loose").
Age is a measure of exposure to risk. This, and the fact that cases
are defined by the occurrence of an event, suggest survival models
here rather than conditional logistic regression.
> The whole age range of difference is 30 months, so 3
> months is still fairly closely matched. When I do the 2 clogit
> equations I
> get different Odds ratios (which seems OK), but don't know which
> one is the
> 'right' one to use. Any help would be appreciated.
You are losing cases (output omitted) with both of your current
designs, which is going to lower your power and, I'm guessing, bias
the estimates in ways you can't predict but have certainly demonstrated.
I'd opt for a strategy that didn't lose information, starting maybe
with a survival model approach.
Royal College of Surgeons in Ireland
+353 (0) 1 402 2431
+353 (0) 87 799 97 95
* For searches and help try:
* For searches and help try: