Extensions and applications of event charts

Event charts, which are used to track and plot timed events at the individual level, are highly effective for presenting multiple time-to-event data, can serve as valuable data-monitoring tools, and can be used to supplement conventional survival plots. Depending on the information chosen and the format to be plotted, commonly used event charts are classified as calendar, interval, or Goldman event charts. A very flexible S-Plus routine was developed for drawing standard event charts and for providing many extensions, such as sorting, subsetting, alignment, and multiple point and line formats, to facilitate the graphical data presentation. If used appropriately, these extensions are useful for revealing the correlation between covariates and multiple events. For example, they can be set up so that a continuous time-invariant covariate such as age of diagnosis is plotted on the y axis, and multiple timed events are plotted on the x axis. They can also be set up so that discrete time-invariant covariates such as the treatment arm are sorted and characterized using different line formats (line type, width, and color). In addition, discrete time-varying covariates can be shown by changing the line formats between line segments. The strengths and weaknesses of various event charts and their applications are shown using examples ranging from historical cohort studies to randomized clinical trials.