hello
I have question related to how to use mixed procedure in meta-analysis contexte.
let suppose that I have this kind of data (case-control summarized data i.e not raw data):
example: summarized data for 6 study
study_ID number_of_patient score_mean_cases std_case score_mean_control std_control
1 30 36.2 8.3 35.1 4.5
2 70 45.2 4.2 39.0 2.8
3 20 .......................................................................................ect
4 102 .......................................................................................ect
5 48 .......................................................................................ect
6 20 .......................................................................................ect
to fit fixed- or random-effect models for metaanalysis,
I find this SAS example but I dont have the effsize variable, only what I have is in the example that I mentionned i.e number of patient, mean for case, mean for control, and thier respective standard deviation.
by advance thanks
DATA elderly;
INPUT study effsize cogtest $;
CARDS;
1 1.75 Trail
2 1.94 Trail
3 1.34 MOT
4 1.30 MOT
;
PROC MIXED DATA=elderly;
CLASS study cogtest;
MODEL effsize = cogtest / P SOLUTION;
RANDOM study / SOLUTION;
REPEATED / GROUP = study;
PARMS (.08)
(0.1209) (0.07) (.048) (0.0757)
/ EQCONS=2 to 5;
RUN;
