Dear allstaters
Expert opinion will be highly valued on this case scenario. Many thanks in
advance
In a population survey, baseline info was collected on individuals on a
disease severity scale: mild, moderate and severe.
1 year after the intial survey, subjects were contacted to find out whether
they were dead or alive.
Mortality data could not be obtained on 30% of the subjects, and there was
evidence that the proportion of severe disease in this missing group was
higher than in those who did report whether they were dead or alive.
Odds ratio for mortality for the 3 disease categories were calculated on the
sample with non-missing mortality info.
Given that those with missing mortality data had more prevalence of severe
disease, would it be reasonable to think that the odds ratios calcuated with
the non-missing sample were underestimated?
Many thanks
Roman Romero-Ortuno
PhD student, TCD
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|