The volume of prosecution that had occurred or was slated to occur was laughable compared to the amount of fraud known or reasonably believed to have occurred. When it is done at scale, prosecution is inefficient and much less effective than reforming processes so as to preempt fraud, which is not something that happened, as evidenced by the continuing fraud after the initial round of prosecution.

FTA:

> So-called “pay-and-chase”, where we put the burden on the government to disallow payments for violations retrospectively, has been enormously expensive and ineffective. Civil liability bounces off of exists-only-to-defraud LLC. Criminal prosecutions, among the most expensive kinds of intervention the government is capable of doing short of kinetic war, result in only a ~20% reduction in fraudulent behavior. Rearchitecting the process to require prior authorization resulted in an “immediate and permanent” 68% reduction. (I commend to you this research on Medicare fraud regarding dialysis transport. And yes, the team did some interesting work to distinguish fraudulent from legitimate usage of the program. Non-emergency transport for dialysis specifically had exploded in reimbursements—see Figure 1— not because American kidneys suddenly got worse but because fraudsters adversarially targeted an identified weakness in Medicare.)