This website uses Cookies. Click Accept to agree to our website's cookie use as described in our Privacy Policy. Click Preferences to customize your cookie settings.
Accept
Reject
Preferences
cancel
Turn on suggestions
Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.
Showing results for 
Show  only  | Search instead for 
Did you mean: 
Have your say in shaping JMP's future by participating in the new JMP Wish List Prioritization Survey
Choose Language Hide Translation Bar
Sandeep123
Sandeep123
Level III

controlling for site clustering

Created: Mar 9, 2021 02:00 AM  |  Last Modified: Jun 8, 2023 08:30 PM (417 views)

A large data set of >200,000 patients is from 125 hospitals. I am analyzing difference in outcomes ( mortality ( nominal) and Length of stay ( continuous) based on admission source. The other predictor variables include age, diagnostic categories, severity of illness score etc.

 

I need to account or site clustering. I can do this by adding site ID in the Fit model...I can make it into random effect also as I read on one of the discussion posts.

 

I wanted to ensure this is appropriate method to account for the clustering. I will be doing linear regression by standard least square and nominal logistic regression.

 

thank you for looking into it

0 Kudos
1 REPLY 1
Mark_Bailey
Staff Mark_Bailey
Staff

Re: controlling for site clustering

Mar 9, 2021 05:31 AM (397 views)  |  Posted in reply to message from Sandeep123 03-09-2021

Yes, it sounds like a logistic regression model with covariates and site in the linear predictor is appropriate. You might also include cross terms for potential interaction effects such as length of stay at a particular site.

0 Kudos