Hi all,
I am stuck at the following problem: I need to determine if there is a significant difference in Length of stay (LOS) in two groups: decendents and survivors, in different age groups. Can somebody help me with this? In a similar article I found they did use kruskal wallis to calculate this. Somebody some ideas to solve this in stata?
Thanks a lot!
Hi Daniel,
One of the way is to create a dummy variable = 1 for decedents and 0 for survivors, or vice versa. Then you can use ttest LOS if age == "18-40", by (dummy). And so, for all age groups.
Daniel van den Kroonenberg, you can create an interaction term between dec/surv and age, include this interaction term in the model together with dec/surv and age. The coefficients on the interaction terms will tell you the differences between decendents and survivors depending on age.
Hello Daniel van den Kroonenberg. It is quite common to use survival analysis to analyze length of stay (LOS), with the discharge from hospital (or wherever the stay is happening) as the event: 1 = discharge, 0 = no discharge. Here is an example I found just now:
- https://www.medrxiv.org/content/10.1101/2020.04.07.20057299v1
Those authors used R, but Stata has good survival analysis capabilities too. Here are a couple of links to get you started.
- https://www.stata.com/features/survival-analysis/
- https://stats.idre.ucla.edu/stata/seminars/stata-survival/
HTH.
Hello Daniel,
The two independent variables (age group: 3 levels; and type of case: 2 levels) could be evaluated in a two-way anova design. If there is any concern about conformance to normality and/or homogeneity assumptions, then exact/resampling/bootstrap estimation may be used to derive the significance tests. These are increasingly available within or as adjuncts to many statistical software packages and libraries (e.g., R, SPSS, SAS).
Here's a couple of links that you may find helpful in this regard:
https://statisticsbyjim.com/hypothesis-testing/bootstrapping/
https://moderndive.com/8-confidence-intervals.html
https://statweb.stanford.edu/~tibs/stat315a/Supplements/bootstrap.pdf
Good luck with your work.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics