RobertNiles.com
About Robert
Reporting Help
Finding Data on the Internet
Questions for Candidates
Stats Lessons
Mean
Median
Percent
Per capita
Standard Deviation
Margin of Error
Data Analysis
Sample Sizes
Stats Tests
Help Board
Bookstore


Statistics for the Utterly Confused

This book offers a super-accessible approach to the much-misunderstood subject of statistics.
More information
-->

Home > Statistics Every Writer Should Know > The Stats Board > Discusssion

Survival analysis
Message posted by Jim Atienza on October 27, 1999 at 12:00 AM (ET)

I'm doing a study on the prognostic value of two diagnostic tests. According to the literature, one ensures a benign outcome in all but less than 1% of patients. Now I would like to know if a second test predicts at least a 15% higher incidence of adverse events. I have gathered two independent samples totaling 89. So far no one's had any adverse events yet. This means that all my data would be censored in a Kaplan-Meier analysis. Is there any other way to analyze my data?


READERS RESPOND:
(In chronological order. Most recent at the bottom.)

Re: Survival analysis
Message posted by JG on October 27, 1999 at 12:00 AM (ET)

If the adverse reaction is 1% then the probability is approximately 10% that in 89 patients there will be no adverse reaction. You need a bigger sample. Once adverse reactions start to appear, you should plot a cumulative estimate of the probability of an adverse reaction.


Re: Survival analysis
Message posted by Jim Atienza on October 27, 1999 at 12:00 AM (ET)

Thanks for the response. The problem is I don't have any more time to get more subjects. My sample size was based on how many subjects it would take to detect a 15% difference (16% - 1%) with a single-sided test with a level of significance of 0.05 and a power of 0.8.
Isn't there anything I can do with the data I already have?


Re: Survival analysis
Message posted by JG on October 28, 1999 at 12:00 AM (ET)

By doing some basic computations you can see that if you wish to distinguish between p=.01 and p=.16 , with alpha .05 and beta .2 then the sample size is 1. That is, if you have more than 1 observation and there are no occurances then accept p=.01 otherwise accept p=.16 .



Your $5 contribution helps cover part the $500 annual cost of keeping this site online.

Niles Online Sites:RobertNiles.comTheme Park InsiderViolinist.com

RobertNiles.com™, the site, content and services 咀opyright 1996-2002, Robert Niles.
All rights reserved. Questions? Comments? Read my Privacy Policy, or E-mail me!