Operational Difficulties with Internal Pilot Studies to Update Sample Size
- PDF / 555,760 Bytes
- 8 Pages / 504 x 719.759 pts Page_size
- 44 Downloads / 196 Views
0092-8615/2000 Copyright 0 2M)o Drug Information Association Inc.
Printed in the USA. All rights reserved.
OPERATIONAL DIFFICULTIES WITH INTERNAL PILOT STUDIES TO UPDATE SAMPLE SIZE SIMONDAY,BSc CSTAT Medical Department, Leo Pharmaceuticals, Buckinghamshire, United Kingdom, and Applied Statistics Group, Napier University
Internal pilot studies used for establishing the required sample size for a clinical trial part way through that trial are attractive when one isfaced with difficulties in determining underlying variances or underlying event rates. By extrapolating a known curio of posthoc sample size calculations, it is demonstrated how a sample size calculation may be obtained from a blind internal pilot that increases the study size even though, had a decision been made to stop the study and unblind it, enough information existed to reject the null hypothesis. The questions of whether to continue recruiting and how much further to recruit may be considered. This will be influenced by the determination of thefinal (relative to the initial) study size determined from the internal pilot and the current perceived interest in the research question. Ifa study is stopped at the point of the recalculation, the reasons f o r the early stopping need to be carefully considered to decide if any bias may result. The decision to stop may be unconditional or conditional on the data. The two cases are distinguished through a variety of examples and their possible biases are considered. Key Words: Clinical trial; Internal pilot study; Operational bias; Sample size; Sequential methods
INTRODUCTION AN INTERNAL PILOT study can be used to determine (or, strictly speaking, redetermine) the required size of a clinical trial part way through recruitment for that trial. The basic methodology has been outlined by Wittes and Brittain (1) and by Gould (2). The procedure they each describe is as follows: 1. From initial considerations, determine the sample size (No)required in each treatment group, 2. Recruit only n patients to each group,
Reprint address: Simon Day, Leo Pharmaceuticals, Longwick Road, Princes Risborough, Buckinghamshire. HP27 9RR United Kingdom. E-mail: simon. [email protected].
where n =7tNo ( x typically being about 0.5), and calculate the observed variance of the sample data 3. Use this estimate of 0’ to recalculate the group sizes as N , , 4.Continue recruiting; the required final recruitment figure for each group is the maximum of No and N , , and 5 . The analysis proceeds in the usual manner as if no recalculation had taken place and uses all data collected (those prior to and after the recalculation of sample size).
(a2), a2
Birkett and Day (3) took the idea a stage further and while investigating what n might optimally be showed that no initial sample size (No)was necessarily required and that x was not relevant. The sample size calculation could be performed at any point after a rea-
461
Downloaded from dij.sagepub.com at NORTHERN KENTUCKY UNIV on June 4, 2016
462
Simon Day
sonable amount of data ha
Data Loading...
