Hi,
I'm currently using Stata 9.2. I have an analytical question I'm hoping someone could help me with:
- I'm currently examining the association between volume of procedures at a center and outcomes (death vs. no death) using a survey sample. The volume variable is divided into low, medium, and high volume centers. I'm using multivariate logistic regression (svy: logistic) to examine if the adjusted mortality (after adjusting for a number of confounders) differes in high, medium, and low volume centers.
- By definition, because high volume centers have higher number of procedures, they contribute more data points to both the numerator and denomintor of mortality rates.
Does this phenomenon affect the analysis and should it be corrected for?
- Alternately, should I calculate the % death at each center and then calculate the average mortality rate per center among each strata?
- Since the dataset contains the invididual patient data as individual records rather than center information, how do I convert the analysis that the center is the unit of analysis rather than each patient? I'm adjusting for clustering of outcomes within each hospital by using the generalized estimating equations. Is that sufficient?
Thank you for your help,
Ashwin
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
