Home  /  Products  /  Features  /  Group sequential designs

Group sequential designs New

Why wait until you finish collecting data to analyze the results of your clinical trial? Group sequential designs (GSDs) allow you to preplan interim analyses of clinical-trial data while controlling the familywise error rate. If interim results are conclusive, you can stop the trial early for efficacy or futility, saving resources and sparing participants from receiving inferior treatments. Use Stata's gsdesign commands or interactive Control Panel to calculate efficacy- and futility-stopping boundaries, compute sample sizes for interim analyses, compare GSDs with equivalently powered fixed studies, and graph the stopping boundaries for your trial.

Learn about Group sequential designs.

Watch Group sequential designs.

GSD methods

  • One-sample mean test
  • Two-sample means test
  • One-sample proportion test
  • Two-sample proportions test
  • Log-rank test of survivor functions
  • User-defined test

Stopping boundaries

  • Classical O'Brien–Fleming bounds
  • Classical Pocock bounds
  • Classical Wang–Tsiatis bounds
  • Error-spending O'Brien–Fleming-style bounds
  • Error-spending Pocock-style bounds
  • Error-spending Kim–DeMets bounds
  • Error-spending Hwang–Shih–de Cani bounds

Stopping for

  • Efficacy
  • Futility
  • Both efficacy and futility

Futility boundaries

  • Binding
  • Nonbinding

Control Panel interface

  • Select analysis type from descriptive summaries
  • Specify analysis interactively
  • Automated tables and graphs
  • Customizable tables and graphs
  • Command interface

Graph boundaries

  • Show critical values at each look
  • Compare with fixed-study critical values
  • Customizable

Interim analyses

  • Evenly spaced
  • Unevenly spaced at user-specified increments

Additional resources