syntrial

synthesize clinical trial data

From an clinical trial new synthetic data are generated. Synthetic patients are created by

• randomly selecting a fixed number, default 3, of real source patients
• assign random weights to each source patient
  such that no source patients gets more weight than a certain limit, default 2/3
• take weighted means for continuous variables
• select master source patient for categorical variables,
  optionally adding noise per observation.

SDTM

Clinical trial data are expected to be in SDTM-like format without RELREC tables. Adaptation to ADAM datasets should be straightforward.