Session 2: Survival Analysis (Cunningham)

Learning Objectives: 

  • To understand the basic principles of survival analysis and Kaplan Meier curves, including:
    • The advantages of using survival analysis
    • How to create a survival table and Kaplan-Meier curve
    • How to critically analyze studies that use survival analysis


Reading:

 

Questions:

  1. In what kind of studies do we use survival analyses or Kaplan-Meier analyses?
  2. What are some of the advantages to analyzing data using survival analysis?
  3. What is censoring? When does it occur? Why is censoring important in survival analysis?

 In the article by Lee, et al.:

  1. What study outcome is analyzed using a survival analysis method and Kaplan-Meier curve?
  2. How was “death” or “failure” defined?
  3. What are other ways that this study outcome could have been analyzed (e.g. other than using survival analysis)?
    1. How would another type of analysis be better than survival analysis?
    2. How would another type of analysis be worse than survival analysis?
  4. Why is there such a steep decline in the Kaplan-Meier curve (Figure 1) at 1 week?

In the article by Alford, et al.:

  1. In 1-2 sentences, describe the study findings demonstrated in Figure 1.
  2. What study outcome is analyzed using a survival analysis method and Kaplan Meier curve?
  3. How was “death” or “failure” defined?
  4. In what situation(s) were data censored? Do you agree with why/how the authors censored data?
  5. In the results section on p. 75, the authors state that the mean duration of treatment retention for both the homeless and housed groups was 9 months (180 days). How/why does the Kaplan-Meier curve in Figure 1 extend out to 350 days?
  6. From the Kaplan-Meier curve in Figure 1, what is the 3-month (90-day) treatment survival rate for both groups? What is the 12-month (350-day) treatment survival rate for both groups? How does this compare to what the authors state in the results section in terms of remaining in treatment at 3 and 12 months? Why is there a discrepancy?