Abstract

Studies on health care practices, financing, and organization increasingly rely on Medicare and other expanded data sets. These studies are of critical importance for public policy and for the development of strategies to contain escalating health care costs, but they often use data that have been corrupted by fraud and abuse. Mistaken conclusions, as to the effectiveness of policy and procedures, are likely being reached in studies that have used corrupted data. Researchers need to consider the suspect nature of results obtained from the corrupted data, and determine methods for making the data more valid