A simple method based on broken-line interpolation for displaying data from long-term clinical trials.

In displaying data from long-term clinical studies, variation across patients in the times at which readings are recorded often presents a problem. A simple method for treating such data is proposed, which avoids arbitrary conventions and methodological difficulties inherent in other approaches. It is based on within-patient broken-line interpolation, and is easy to implement on a computer. The method is tested and compared to others with computer-simulated data whose time-pattern is known. Though the comparison is of limited scope, the method appears to compare favourably with the 'closest-reading' and 'window' approaches.