The work flow currently is Drug A comes to market with expense X, it is competing with older Drug B with expense Y. There are no head to head trials, but an expert (usually a pharmacist) can reach solid benefit/harm conclusions. That pharmacist writes up a nice summary with a recommendation and present to a committee made up of other experts (usually doctors w/ academic background). They discuss and come to a formulary decision for the whole hospital.
In insurance companies, some background financial shady stuff can take place (kickback schemes) that influence this process, but some form of honest data analysis takes place anyway.
I used to do this for a while, it can work. Just have to convince the right people that your software would be better than the pharmacist. The market incentives doesn't really encourage that kind of automation/savings, so it will be an uphill battle that need experienced sales folks and good VC money.