Question

Suppose that a clinical trial will be conducted to investigate the effectiveness of an experimental cancer drug as a treatment for patients with advanced lung cancer. Suppose that the plan is to have one hundred subjects in the trial.  

1. Suppose that the lengths of each one of three metastatic tumors in each subject will be measured approximately every four weeks. How do you suppose that those tumor-size results could be displayed in a way that shows the results for each one of the 100 subjects (i.e., that shows what happened to each individual subject)? (What sort of a chart might you use to do so?) Explain.   

2. Suppose that you want to display the survival results for the 100 subjects, and that you want to do so in a way that makes clear what the results were for each subject. What sort of chart might you create to do so? [Note: As we discussed in class, usually many subjects in a clinical trial have not died by the trial's data cut-off date.]

Answer 1

1) A box plot with scatter points, where each scatter point represents one subject, can show the distribution of the tumor sizes for each case, but at the same time, it provides the mean and SD for the group, along with outliers, to provide information at a single glance. When the measurements for every 4 weeks are provided side-by-side, it enables easy interpretation of the general trend, mean and SD, while also showing points corresponding to the number of subjects for whom the treatment may not be effective.

2) A Kaplan–Meier survival plot depicts the survival of each individual of the group at each time interval, by plotting the number of remaining individuals at each time interval. When plotted between the drug and placebo controls, this type of plot can enable easy and clear comparison of survival trends without discarding individual mortality information.