For the studies described below:
• Describe the event of interest, beginning of time, time metric, and time-to-event random variable.
• State which type(s) of censoring (i. e., left, right, or interval) may be present in each study, and briefly explain your answers.
a. Survival/ sacrifice experiments are designed to determine whether a suspected agent accelerates the time until tumor onset in experimental animals. For such studies, each animal is assigned to a prespecified dose of a suspected carcinogen and then examined at sacrifice or death for the presence or absence of a tumor. Since a lung tumor is occult (detectable only by microscopic examination or chemical analysis), the time until tumor onset is not directly observable. Instead, we observe only a time of sacrifice or death.
b. Beadle and colleagues report a study carried out to compare the cosmetic effects of radiotherapy alone versus radiotherapy and chemotherapy on women with early breast cancer. 18 To com-pare the two treatment regimes, a retrospective study of 46 radiation only and 48 radiation plus chemotherapy patients was made. After treatment, patients were observed initially every 4–6 months. As their recovery progressed, the interval between visits lengthened. At each visit, the clinician recorded a measure of breast retraction on a 3-point scale (none, moderate, severe). Researchers were particulary interested in moderate or severe breast retraction.
c. A study was conducted to determine the age at which marijuana was first used among high school boys in California.19 Researchers asked 191 high school boys the question “When did you first use marijuana?” Possible answers were exact age, “I never used it,” and “I have used it but cannot recall just when the first time was.”