| In order to estimate the total overpayment amount made to a suspicious health care provider over a period of time, the statistician selects a probability sample from the payment population for investigation. From the adjudged sample overpayment amounts, a 90% lower confidence bound for the total population overpayment is constructed and used as a recoupment demand. This thesis is a continuation of work by Don Edwards and colleagues on overpayment investigations in Medicare. In previous papers, he discussed a new method that yielded consistently good results in simple random samples for certain kinds of payment populations. This new method, called the Minimum Sum method, is generalized here for stratified sampling and tested and compared to other strategies for different populations and samples. |
