Loading...
Please wait, while we are loading the content...
Similar Documents
Making sense of biostatistics: group sequential designs and interim analyses.
| Content Provider | CiteSeerX |
|---|---|
| Author | Turner, J. Rick |
| Abstract | The trial designs we have discussed so far have been fixed-sample designs. The sample size is estimated during the trial’s planning stage. The resulting number of subjects is stated in the study protocol. These subjects are then enrolled and randomized. Statistical analyses of the data are then conducted at the end of the trial. The focus of this column is another design, called the “group sequential design, ” in which interim analysis plays a crucial role. The fundamental purpose of interim analysis is to facilitate informed decision-making at various points about whether the trial should be terminated early. The word “group ” in the name “group sequential design ” does not refer to a treatment group, or “arm. ” It refers to the group of all subjects who have completed a trial before an interim analyses is conducted. Imagine a study in which a maximum of 1,000 subjects will be recruited into each of two treatment arms. The study’s design calls for three potential interim analyses and one final analysis if the trial has not been terminated. The first interim analysis will be conducted once 500 subjects (approximately 250 subjects in each arm) have completed their participation. If the trial continues, another 500 subjects will complete their participation and join the interim analysis group. The second interim analysis will use the data from all completed subjects, i.e., 1,000 subjects. If the trial continues, another 500 subjects will complete their participation and join the interim analysis group. If the trial continues, the final 500 subjects will complete their participation. At this point, all 2,000 subjects will have participated in the trial, and a final analysis will be performed. Why would we want to terminate a trial before all of the potential subjects have completed their participation? There are several reasons, all of which relate to the fact that choosing the number of subjects for a clinical trial is a matter of estimation, not precise formulaic determination. A chapter from my book New Drug Development (Turner, 2007) that focuses on sample size estimation is available at |
| File Format | |
| Access Restriction | Open |
| Subject Keyword | Interim Analysis Group Sequential Design Final Analysis Interim Analysis Group Name Group Sequential Design Sample Size Statistical Analysis Design Call Sample Size Estimation Several Reason Fundamental Purpose Word Group Second Interim Analysis First Interim Analysis Study Protocol Various Point Clinical Trial Precise Formulaic Determination Informed Decision-making Crucial Role Fixed-sample Design Book New Drug Development Treatment Group Treatment Arm Potential Subject Trial Design Trial Planning Stage Potential Interim Analysis |
| Content Type | Text |
