Life Tables Exercise: Comparing surgical treatments for malignant melanoma

Unlike other examples used in this class, the data from this assignment is fictitious. There are two reasons for this. First, it is unethical for us to release private medical information such as diagnosis, date of diagnosis, and prognosis (we will talk about ethics later in the semester). In order for you to complete this exercise, we must provide such information. Second, even if we did release 'real' information, a real study would potentially include several hundred people making calculations complex and time-consuming

In this exercise, a clinical trial was conducted to compare two surgical treatments for malignant melanoma: simple excision (Treatment A) versus radical surgical excision (Treatment B). The clinical trial was stopped on April 1, 2006, after 14 patients were entered (7 in each treatment group). Using the data below, calculate a life table based on your Gordis textbook. For simplicity, use one month categories, assume all months have 30 days, and use the randomization date as your start date.

Patient

Treatment

Clinical Course

1

A

Diagnosed Dec 10, 2003; randomized Jan 1, 2004; died Jan 4, 2004, before receiving treatment

2

A

Diagnosed Jan 1, 2004; randomized Feb 1, 2004; switched from Treatment A to Treatment B on Feb 10, 2004; died May 1, 2004.

3

A

Diagnosed Jan 1, 2005; randomized Feb 1, 2005'; died Sept 1, 2005.

4

A

Diagnosed Dec 1, 2004; randomized Jan 1, 2005; left the country and lost to follow-up Nov 1, 2005.

5

B

Diagnosed Feb 1, 2004; randomized Feb 15, 2004; died March 15, 2004.

6

B

Diagnosed March 1, 2004; randomized March 15, 2004; died April 15, 2004.

7

B

Diagnosed May 10, 2004; randomized June 1, 2004; died Dec 20, 2004.

8

B

Diagnosed April 1, 2005; randomized May 1, 2005; died Aug 1, 2005.

9

B

Diagnosed Jan 1, 2005; randomized Feb 1, 2005; diagnosis revised to nonmalignant tumor, Feb 1, 2006; alive at stopping date, April 1, 2006.

10

B

Diagnosed April 1, 2004; randomized May 1, 2004; died June 10, 2005.

11

A

Diagnosed March 1, 2004; randomized March 15, 2004; lost to follow-up July 1, 2004.

12

B

Diagnosed April 15, 2005; randomized April 25, 2005; alive at study termination on April 1, 2006.

13

A

Diagnosed May 1, 2004; randomized June 1, 2004; died Feb 20, 2005.

14

A

Diagnosed Sept 1, 2005; randomized Sept 6, 2005; died Feb 1, 2006.

Questions (Be sure to include your life table with the answers you turn in!):

1.Why would you use the randomization date as your start date, rather than the diagnosis date? What are the pros and cons for each?

2.What are the crude death rates based on the two treatments?

3.How do the cumulative probabilities of survival compare, and what conclusions can you draw about the relative effectiveness of the two treatments?

4.In which treatment group was patient #2 included, and why?

5.Did you exclude patient #9? Why or why not?

6.What are the underlying assumptions of survival analyses?