Case match study

[TraumaticFun]

I'm using Stata. I have 2 groups, one received new pain medication and the other didn't because they refused. I monitored their pain and other stuff.

My hypothesis was that those who received medication felt less pain.

Because there was no randomisation, to avoid bias I decided to do case-match statistics and perform McNemar test and conditional logistic regression.

But I never did it in Stata, or any other statistical program. So can someone give me some tips and explanation how to do that stuff? (I have general info like gender and age).

Thanx

 

[hlsmith]

So you have or have not matched patients yet? Unsure of your exact question.

 

[TraumaticFun]

Didn't matched yet. But don't really know to do it properly :/

 

[hlsmith]

I guess some basic questions may be, how big is your sample size and how many potential covariates do you need to control for?

 

[TraumaticFun]

I guess some basic questions may be, how big is your sample size and how many potential covariates do you need to control for?

I'm a bit new to this matching thing, sorry for dumm questions.

My sample size is 102. 42 got treatment and 60 didn't. Selection was not randomised. So I found out that match case statistic is an opinion.

I want to found out if pain and delirium scores was different in not treated group.

If i understand your question correctly, my covariates could be age, gender, and pain score befor treatment.