Confidence intervals and statistical power of the 'Validation' ratio for surrogate or intermediate endpoints

Laurence S. Freedman

Research output: Contribution to journalArticlepeer-review

52 Scopus citations

Abstract

There is currently much interest in the use of surrogate endpoints in clinical trials and intermediate endpoints in epidemiology. Freedman et al. [Statist. Med. 11 (1992) 167] proposed the use of a validation ratio for judging the evidence of the validity of a surrogate endpoint. The method involves calculation of a confidence interval for the ratio. In this paper, I compare through computer simulations the performance of Fieller's method with the delta method for this calculation. In typical situations, the numerator and denominator of the ratio are highly correlated. I find that the Fieller method is superior to the delta method in coverage properties and in statistical power of the validation test. In addition, the formula for predicting statistical power seems to be much more accurate for the Fieller method than for the delta method. The simulations show that the role of validation analysis is likely to be limited in evaluating the reliability of using surrogate endpoints in clinical trials; however, it is likely to be a useful tool in epidemiology for identifying intermediate endpoints.

Original languageEnglish
Pages (from-to)143-153
Number of pages11
JournalJournal of Statistical Planning and Inference
Volume96
Issue number1
DOIs
StatePublished - 1 Jun 2001

Fingerprint

Dive into the research topics of 'Confidence intervals and statistical power of the 'Validation' ratio for surrogate or intermediate endpoints'. Together they form a unique fingerprint.

Cite this