Blinded Sample Size Adjustment
- PDF / 477,003 Bytes
- 8 Pages / 504 x 720 pts Page_size
- 30 Downloads / 283 Views
Drug lnfonnofion Journal, Vol. 35. pp. 1123-1130, 2001 Printed in the USA. All rights reserved.
Copyright 0 2001 Drug Information Association Inc.
BLINDED SAMPLE SIZE ADJUSTMENT DAVIDR. BRISTOL* Director, Biostatistics and Statistical Programming, Purdue Pharma L.P., Stamford, Connecticut
LINDASHURZINSKE Senior Manager, Biostatistics, Wizer Global Research and Development, Ann Arbor, Michigan
The determination of the appropriate sample size is an important aspect of planning a clinical trial. In recent years, procedures for estimation of a nuisance parameter to adjust the sample size if necessary have been examined. Here, it is assumed that the clinical trial is conducted for the comparison of two treatments, where the observations are assumed to have normal distributions with a common unknown variance. For sample size determination, the variance is assumed known and the resulting sample size is sensitive to misspecification of the variance. An estimate of the variance, obtained while the clinical trial is ongoing, can ofien be used to assess the appropriateness of the assumed variance. The use of a blinded estimate of the variance to potentially adjust the sample size is examined. Key Words: Sample size; Blinding; Two-stage
INTRODUCTION THE DESIGN OF A comparative clinical trial includes specification of the sample size, which is usually accomplished by controlling the power. When the variable of interest is assumed to have a normal distribution, the comparison is usually made with respect to the difference in the means, and it is assumed that the variances for the two treatments are equal. The determination of the required sample size requires specification of the common variance, and the appropriateness of the sample size depends on how close the specified value is to the true variance. The value of the assumed variance may be based Based upon a poster presentation at the DIA Annual Meeting, June 2000, San Diego, California. Reprint address: David R. Bristol, Director, Biostatistics and Statistical Programming, Purdue Pharma L.P., One Stamford Forum, Stamford CT 06901. *This research was conducted while the first author was employed at F'tizer Global Research and Development, Ann Arbor MI.
on earlier studies of the same, or related, treatments, but the actual variance is unknown. It is desirable to assess the adequacy of the assumed variance, while having little effect on the type I and type I1 error rates. This can be accomplished by estimating the variance while the clinical trial is ongoing, and adjusting the sample size, if necessary, based on this estimate. Stein (1) presented a two-stage procedure for sample size determination when the variance is unknown. Similar problems have received considerable attention during the last 15 years. Wittes and Brittain (2) are often credited with having generated this renewed interest. Shih (3) considered the EM algorithm to maintain the blind. Bristol (4), Herson and Wittes ( 3 , Birkett and Day (6), Gould (7), Gould and Shih (8,9), and others have presented other
Data Loading...
