CHAPTER 21 Summarizing and Graphing Survival Data 315
survival curve in Figure 21-7b has smaller steps than the life-table survival curve
in Figure 21-5b, so it’s more fine-grained. This is because the step curve now
decreases at every time point at which a participant died. You can tell from the
figures where participant #1 died at 0.74 years, #9 died at 2.27 years, #4 died at
2.34 years, and so on.
While the K-M survival curve tends to be smoother than the life table survival
curve, just the opposite is true for the hazard curve. In Figure 21-7a, each partic-
ipant has their own very thin bar, and the resulting chart isn’t easy to interpret.
Heeding a Few Guidelines for Life-Tables
and the Kaplan-Meier Method
Most of the larger statistical packages (see Chapter 4) can perform life-table and
Kaplan-Meier calculations for you and directly generate survival curves. You have
to identify two variables for the software: one with the survival time for each par-
ticipant, and a binary variable coded 1 if the survival time represents time to death
or the event, and 0 if it represents censored time. It sounds simple, but it’s sur-
prisingly easy to mess up. Here are some pointers for setting up your data and
interpreting the results properly.
Recording survival times correctly
It is important to draw a distinction between data collection and data analysis.
When recording the raw data, it’s best to collect all the relevant dates for the
study. Before the study starts, the dates of interest for data collection should be
specified, which could include date of diagnosis, start of therapy, end of therapy,
FIGURE 21-7:
Kaplan-Meier
estimates of the
hazard (a) and
survival (b)
functions.
© John Wiley & Sons, Inc.