Single group comparison of incidence pre and post

[Melk]

I have a single group with incidence pre and post some intervention. The populations are different pre and post but I am not sure how to account for this in my analysis.

[ballardw]

What type of analysis are you actually attempting - research question to answer?

And some example data might help, replace anything sensitive with different values like xxxx, yyyy or similar.

[Melk]

Research question is: is the intervention effective in reducing the number of injuries?

I have data from 5 different counties, with the populations pre-intervention and post-intervention, along with the incidence of injuries pre and post intervention as well.

[ballardw]

Do you want to know about each county or the data overall?

Is the population in the same data set as the incidence rate?

Do you have a variable that indicates whether a value is pre or post?

If you had posted some example data these questions would not be needed.

McNemar's test is sort of designed or a 2x2 table. is your data in form conducive to that? That would tend to look like

When  value

Pre      count yes

pre       count no

Post    Count yes

Post    count no

Also there are number of procedures that will do that test, or similar? Which one have you mind?

[Melk]

Sample data:

ID       Time      Count     Total

1          pre         3           1200

1          post        5           1300

2          pre         0           1250

2          post        3           1310

..

I want to do by both county and overall. The event is very rare, I wondered if an exact poisson would be appropriate but we cannot use the repeated statement with an exact poisson. Any advice would be helpful.

[rsanchez87]

Hi, I am looking to understanding the change pre-post for a single group as well. 

Here is my data: 

SRC	COUNT	AGE0_10	MEMCOUNT	PERCENT_SRC
POST	18794	0	17166	91.3377
PRE	6956	0	6832	98.2174
POST	18794	1	1628	8.6623
PRE	6956	1	124	1.7826

These are dependent samples. The category being evaluated is AGE0_10 (age between 0 and 10) in the pre-period and post-period. 

As you can see there is a 7% increase from pre-perod (2%) to the post-period (8%)

This is the code i wrote: 

ODS HTML CLOSE; ODS HTML;
PROC FREQ DATA = AGE0_10_SUM ORDER = DATA; 
WEIGHT MEMCOUNT;
TABLES AGE0_10*SRC / AGREE; 
TITLE "AGE0_10"; 
RUN;
;QUIT;

Which is giving me a p-vale <0.0001 ---> I believe this is wrong. The change may be significant, but not at this magnitude. 

Can you advise how to set up the summary table, I have subject level details. 

Thank you.