cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
Try the Materials Informatics Toolkit, which is designed to easily handle SMILES data. This and other helpful add-ins are available in the JMP® Marketplace
Choose Language Hide Translation Bar
LarsBirger
Level III

How to compensate for hospital volume of cholecystectomies?

I very much appreciate working with JMP and am presently using JMP Pro version 16.0.0 for my research. I know that there now are more recent versions available, but I have not presently dared to upgrade in case something happens. I am using it on a Mac Book Pro (Apple M1 Max with 64 GB memory) and JMP works perfectly with this laptop.

 

I have a statistical question for the community. I presently do research about quality register data and usually use the Analyze/Fit Model for multivariable analyses. In the example below the Y-variable is Intra-and/or postop complications at cholecystectomies (dichotomized to yes or no) and the added x-variables are Sex, ASA-classification, and age and they are also all dichotomized. All of these variables seem to make a difference in the Nominal Logistic equation under Fit Model, se enclosed example.

 

However, the question I have is that in the register all cholecystectomies are performed at 78 hospitals and although they probably perform cholecystectomies in the same way at all hospitals there can be differences in the way patients are treated, the distance to the hospital etc. Furthermore, the number of cholecystectomies in the database varies from 984 at one hospital (A) to 20 cholecystectomies (B). The question I have is there a way to compensate for the different cholecystectomy volumes in JMP and if so how do I do that? My colleague, who is using SPSS, is talking about mixed modules.

 

Would very much appreciate an answer.

 

Sincerely yours

 

Lars

Prof. Emeritus of Surgery.

Umeå University, Sweden.

1 REPLY 1
statman
Super User

Re: How to compensate for hospital volume of cholecystectomies?

I may not understand your query, but I would normalize# the number of cholecystectomies performed (take the ratio of # of cholecystectomies/number performed per hospital). I'm not sure what your data table looks like (can you post an anonymized version or partial version?).  If each cholecystectomy is a row, then every row within hospital would have the same ratio.  The ratio would vary hospital-to-hospital.

I'm not sure what your colleague is recommending regarding mixed modeling?  Mixed models are when you have both fixed and random effects in the same model.

 

https://www.jmp.com/support/help/en/18.0/?os=mac&source=application#page/jmp/mixed-models-and-random...

 

"All models are wrong, some are useful" G.E.P. Box