We introduced the 100% method at a time when purely statistical, random sampling was common. Although random sampling is still used by some practitioners today, it is far less accurate than our 100-percent methodology and typically leads to a lengthy, tedious, on-site validation process.
Claims administrators favored the random-sample methodology because few payment errors can be identified that way, and a plan is only reimbursed for errors actually found. At the same time, for reimbursement purposes, extrapolation beyond the very minimal and restricted claim sample derived is not allowed. Finally, there was no alternative.