LECTURE 24 2011

LECTURE 24 2011 - Interim Analyses of Clinical Trials A...

This preview shows pages 1–12. Sign up to view the full content.

Interim Analyses of Clinical Trials A Requirement

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Outline Background and how DSMBs function Group sequential methods Examples
Suggested Readings Ellenberg S, Fleming TR, DeMets DL. Data Monitoring Committees in Clinical Trials. A Practical Perspective Jennison C and Turnbull BW. Group Sequential Methods with Applications to Clinical Trials Chapman & Hall/CRC 2000

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
DSMB Decision Making Can Be Complex Internal consistency Benefit/Risk External consistency Current versus future patients Clinical and public health impact Statistical issues – monitoring guidelines
Overall Probability of Achieving a Result with Given Nominal Significance of 0.05 After N Repeated Tests Under H o 1 .05 2 .083 3 .107 4 .126 5 .142 10 .193 25 .266 No. of Tests (N) Probability Ref: McPherson, NEJM , 1974.

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Value of Nominal Significance Level Necessary to Achieve a True Level of 0.05 After N Repeated Tests 1 .05 2 .0296 3 .0221 4 .0183 5 .0159 10 .0107 No. of Tests (N) Significance Level Which Should be Used Ref: McPherson, NEJM , 1974.
Simulated Trial (T. Fleming Example) Patients enter trial over a 3-year period 1 year minimum follow-up 60 on A; 60 on B Survival distributions are equal Log-rank test used for analysis 5 situations 1 Log-rank test - at 4 years 2 Log-rank tests - every 2 years 4 Log-rank tests - every year 8 Log-rank tests - every 6 months 16 Log-rank tests - every 3 months 100 simulations of the study

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Results The log-rank p-value was less than 0.05 at the final test (i.e., at 4 years) in 5 of 100 studies either the 2- or 4-year test in 10 of 100 studies at least 1 of 4 yearly tests in 17 of 100 studies at least 1 of 8 semi-annual tests in 21 of 100 studies at least 1 of 16 3-month tests in 25 of 100 studies
Conclusions 1. If one monitors one’s data closely and considers a p .05 result obtained at any time to be reflective of a true difference in survival distributions, one will obtain a grossly excessive number of false positives. This excessiveness will be further magnified if one follows multiple endpoints if one uses multiple test statistics 2. 25 of 100 studies have a log-rank p-value < 0.05 at at least one of the 16 3-month tests, whereas only 5 of 100 studies do at the last test. Thus, if there are no differences between true survival curves, very many of the closely monitored studies will have “statistically significant” early differences which will disappear later in time.

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Early Work Acceptance sampling Wald (1947) sequential probability ratio test Manufacturing problems, continuous monitoring of the data, no upper bound on sample size
Useful References Ellenberg SS, Fleming TR, DeMets DL, Data Monitoring Committees in Clinical Trials, Wiley, 2002. DeMets DL, Furberg CD, Friedman LM.

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
This is the end of the preview. Sign up to access the rest of the document.
• Spring '07
• ph7420
• Statistical hypothesis testing, critical values, Group Sequential Design, 2-sided Group Sequential, • Pocock

{[ snackBarMessage ]}

Page1 / 47

LECTURE 24 2011 - Interim Analyses of Clinical Trials A...

This preview shows document pages 1 - 12. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online