st: Log-linear modeling for capture-recapture using STATA
Date
Mon, 29 Dec 2003 20:24:12 -1000
I am trying to use STATA to model data from two overlapping data
systems (a reportable disease registry and a clinical data system)
using the log-linear method to calculate the number of cases missed by
both data systems (i.e. capture-recapture). The mathematical basis is
described by Tilling, K, et. al., Capture-recapture Models Including
Covariate Effects, American Journal of Epidemiology, 149(9), 2/15/1999,
pp 392-400.
After matching records, I have a dataset with indicator variables for
data source of origin (dichotomous for source 1 and source 2) as well
as a number of covariates (hospital size, region, disease, etc.) that I
would like to explore for interaction via modeling.
Does anyone know how to set up and interpret the models in STATA? I’m
not new to regression via STATA but I can’t figure out how to derive
the missing cell of data (i.e. the records missed by both surveillance
systems) from the model.
Thank you.
Eric Kasowski, DVM, MD, MPH
Navy Environmental and Preventive Medicine Unit No. 6
