Beginners question. Study design. Statistical methods

#1
Hi everyone. I´m a beginner and looking for some help regarding statistical methods and study design. I´m working with a epidemiological retrospective study where I investigate patients in a 10 years period who are developing a certain complication following a certain operation. Every patient are registered with different attributes (independent variables) which should be analysed as risk factors and weighted and presented as potiental risk for complication.

I need help with which statistical method I should use? Multivariate analysis?
Power calculation?
Presentation of data?
 

Karabiner

TS Contributor
#2
I need help with which statistical method I should use? Multivariate analysis?
You have only one end point (dependent variable), as it seems. You did not exactely describe what it is, though.
So it's rather some multiple regression, I suppose (Cox, or linear, or logistic regression maybe).
Power calculation?
What for?
Presentation of data?
Probably.

With kind regards

Karabiner
 
#3
You have only one end point (dependent variable), as it seems. You did not exactely describe what it is, though.
So it's rather some multiple regression, I suppose (Cox, or linear, or logistic regression maybe).

What for?

Probably.

With kind regards

Karabiner
Thanks for your reply. More details:

Primary outcome (end point) is the event of a certain operation following the primary (index) operation. Time span is 10 years. This will be presented on a graph with time on x-axis and risk on y-axis.

The next step is to analyze the independent (risk) variables (about 20-30 both numerical and catergorical) in the affected patients, to see if there are any significant risk factors (for re-operation). This will be presented in a table or (odds ratio?). Here i´m a bit worried that the ratio between independent variables (20-30) and number of patients (200) will need som kind of power calculation?

Very thankful for more tips, and litterature or examples that relates to the above.