Unless you can pinpoint all the causal factors of a relationship, there is no other option I have heard of other than the scientific process, i.e. testing, calculating, re-testing, calculating, retesting...etc that could provide us with knowledge about the relationship between cause and effect. And obviously, it will be wrong sometimes, hence the need to re-test, but there is no better option.
If anything, we need more actual scientific process, meaning using our newfound computing capabilities to open up data so that all experiments can be vetted for accuracy and differing populations tested, etc.
That's why data should be free, in order to let this testing happen. Of course we have limited resources to do testing, so there must be some triage involved somehow, but that has nothing to do with the scientific process.
https://en.wikipedia.org/wiki/Kefauver_Harris_Amendment
This is after the FDA blocked approval of thalidomide, pointing to insufficient testing:
In refusing to approve thalidomide earlier, many thousands of patients died horribly agonizing deaths. Thalidomide was actually tested quite intensively however the company that did it "lost" the records that showed harm to pregnant women.
The whole story is far more complicated, and involves chiral mixtures, pregnancy tests, and a lot more subtleties that belie the simple narrative.