Comparing before and after count data

[chuie]

Hi 

I have 2015 and 2016 claims activity data ( ed use, hospital use etc). i would like to treat 2015 ED utilization as pre data as  they are enrolled on prgram A which stated in 2016. Basically i want to compare the effectivenness of   program on ED utilzation(eg).

data would look like,

ID  YEAR  PROGRAM  EDCOUNT  

1    2015        1                    2

1    2016        1                     4

2    2015         2                   10

2    2016         2                      0

;

Basically i want to compare the effectivenness of   program on ED utilzation(eg).What is the best test for this kind of analysis.

I really appricate any information

thank you

[Reeza]

Try the methods illustrated here:

http://support.sas.com/documentation/cdl/en/procstat/66703/HTML/default/viewer.htm#procstat_freq_exa...

[chuie]

Thank you for promt reply. I have quick question.

Since the enrolment/clinic use didnot happen at the same time for all subject, do we have to consider a person-time insead of regular denominator while we calculate RR?

[Reeza]

Depends.

Do you have enough data to do that type of analysis? You'll need a lot of data points.

If not, make sure to highlight this as a potential issue in the results. 

The assumption that time of enrollment is not a factor in whatever you're trying to measure.

[chuie]

Thank you.

I just have a 2016 data when patient started useing the clinic. Thier claims data of 2015 is what I  am considering pre data.

Even in 2016 not all participated on clinic( meaning any encounter at that clinic) form jan 2016.

The analysis is very tricky as we do not have solid a year of data. hwta do you remcomend in this senario?

Bottom line is we have x  clinc users who are defined as any encounter in that clinic in 2016. Ans we have case mached contro  who  have no encounter in that clinic. Both case and control are the employee of ABC. And  we have claims data of 2015 and 2016 of those case adn control.

we are trying to figure out if clinc users are less likey to use ED than control  AND pre and post test of their ED ustilization in 2015 and 2016.

Does it make sence? I have basic statistic so I am not really sure how to handel this  type of analsis.

And Thank you for your help.