I am Hatem Ali.
I have 2 groups of patients, one received basiliximab induction therapy and the other didnt (no induction)
Both groups are not similar in baseline characterestics (10 character) with P value less than o,o5 when comparing both groups
Example:
Characteristics | Induction(n=2860) | No-induction(n=740) | P value |
Recipient age (years) | 45 | 41.4 | 0.0000 |
Donor age(years) | 46.8 | 43.2 | 0.0000 |
Recipient gender (male) | 1901 | 489 | 0.8423 |
CIT (Hours) | 9.7 | 14 | 0.0000 |
Total HLA mismatch | 2.4 | 1.9 | 0.0000 |
* Treatment assignment model:
logit induction recipientage donorage gender CIT HLA
* IPW weights:
predict double ps, pr
gen double ipw = 1.induction/ps + 0.induction/(1-ps)
* -stset- data:
stset atime [pw=ipw], fail(fail)
* K-M estimate:
sts graph, survival by(induction)
Now I want to assess the similarity of baseline covariates between treated and untreated subjects in the matched sample or in the sample weighted by the inverse probability of treatment.
How do I do that? and how do I know the number of matched patients in each group?
Final question, Is the weighted sample an ATE weights or ATT weights?
Looking forward to hear back from you
0 Response to assessment of similarity of baseline covariates after performing propensity score analysis
Post a Comment