"Drug companies have an incentive to heavily invest in such a program."

Not necessarily. Personalized medicine would lead to the fact that in some cases, where you have lets say a success rate of 40%, you would eventually only treat those 40% after having spend significant money to elucidate why they respond and the others not. So in this example you revenues of the drug would get more than halved - without compensation. But also in a case, where just 10% of patients do not respond to an established drug, do you think the company would spend millions of dollars for diving deep to find out the wherabouts, if overall in 90% all is fine?

Don´t forget that the current challenges in drug development are not the lack of knowledge about DNA profiles, but understanding how to block/trigger certain signalling cascades on the way to cancer or how to find new drug candidates and how to develop them as fast as possible. DNA profiles are not the primary approach for these goals.

The big pharmas who would have the money to stem such a program are unfortunately quite risk avers.

Also not all drugs are affected by the genetic landscape of a patient. Hard to tell the percentage, but I would guess more than 50% of drugs would not improve in performance at all from a better knowledge of each patients DNA.

But overall your concept is very interesting.

For info, I wrote some time ago an article about possible use cases of blockchain in clinical research (german only)
https://steempeak.com/deutsch/@stayoutoftherz/blockchaintechnologie-in-der-klinischen-forschung-wo-stehen-wir-heute