Peer review is a joke. Journals barely glance at shit the reviewers all engage in favoritism and favor trading.
Fuck peer review. How about peer escrow?
Submit your paper, a full and complete set of instructions for replicating the experiment, and the complete and raw data you collected during your experiment.
The journal reads it and gives if it passes a basic bullshit test, puts the experiment into a queue.
Then the journal pulls different experiments from the queue and presents them to you (only the experiment, not the data or paper). You have to replicate an experiment before your paper can be published. (Different weights can be added for cost and time, and things can be categorized so you don't get biologists doing physics.)
Once you replicate an experiment and submit the complete and raw data, the journal compares both data sets and determines whether or not to publish the original paper associated with the experiment you reviewed. Likewise, once someone replicates your experiment your paper may be published. (Depending on the type of experiment, they can require more than one replication, have different levels of review and different tolerances for discrepancy.)
If a paper is published, everything is published. The paper, the experiment description, and all data from the original submitter and all the replicators.
If you're not doing repeatable experimentation, you're not doing science, so fuck off. If your data contains sensitive info, such as HIPAA, anonymize it. It's not hard, though people claim it's hard when they want to prevent you from seeing their bogus data.