1
$\begingroup$

I am dealing with a dataset which patients with 3 different types of treatment. Since the baseline characteristics are significantly different and biased, I need to match the baselines. However, the number of each group is too small to match using PSM. I tried to use mnps and iptw functions of twang package, but I cannot join the two functions and failed to generate the new table 1.

How can I use iptw function with mnps?

$\endgroup$
1
  • $\begingroup$ i've miss written twang as hwang $\endgroup$
    – DANA
    Jan 5, 2023 at 10:49

1 Answer 1

2
$\begingroup$

Have you tried using WeightIt to generate the weights and cobalt to assess balance? WeightIt supports many methods to estimate weights for multi-category treatments, including generalized boosted models like in twang but also many other methods that you should try as well. cobalt provides balance tables and plots for multi-category (as well as binary and continuous) treatments. There is no reason to stick to twang unless you really like its functionality and know you want GBM.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.