Risk-Based Monitoring and Fraud Detection in Clinical Trials Using JMP and SAS

5.3 Hierarchical Clustering of Pre-Dosing Covariates Across Clinical Sites

The analyses of the previous section are relatively straightforward, but rely on the availability of very specific data about the trial participants. Often, birthdates and initials may not be present in the study database over concerns that providing such information violates patient privacy. In the absence of such data, we need alternate methods to identify multiple enrollments. But fear not! There is a wealth of data that can be used to identify similar subjects, though the best data to use may depend on the individuals under study. For example, height is a very useful measurement for adults that should remain relatively constant over time (especially if clinical sites ...