Many articles are peer reviewed without actually looking at the data or recreating all the results.

The whole idea is that a paper should be reproducible so that people can come along later and reproduce it. If they fail to do that then that is what letters are for. Every paper that has been peer reviewed isn’t faultless.