Group sequential design

A group sequential design evaluates accumulating data at several planned interim analyses instead of committing to one final test. At each analysis, a prespecified rule can stop for efficacy, stop for futility, or continue collecting observations. The design specifies both the timing of analyses and the sequential stopping boundaries.

Repeated inspection changes the probability of rejecting a Null hypothesis. Even two independent tests, each at significance level $0.05$, reject at least once with probability $1-0.95^2=0.0975$ under their common Null hypothesis. Tests on accumulating observations are dependent, so their rejection boundaries require their actual joint distribution rather than that independence calculation. An Alpha-spending function distributes the permitted type I error over information fractions; it does not make unchanged fixed-sample critical values valid at every analysis.

For a simple valid construction with $K$ planned analyses, choose separate rejection events whose null probabilities are each at most $\alpha /K$. The union bound makes the probability of any rejection at most $\alpha$, whatever their dependence. More efficient boundaries exploit the dependence between successive test statistics. A Futility boundary removes rejection paths, and a Conditional error principle can justify a later design adaptation while preserving the original error guarantee. The possibility of early stopping also affects point estimation and confidence intervals; fixed-sample formulas do not automatically retain their usual interpretation after selection.

A Futility boundary specifies an interim result below which a clinical trial stops without rejecting the null hypothesis. Removing only potential rejection paths cannot increase the type I error. With a finite boundary and a nonsingular joint normal test-statistic distribution, it strictly reduces that error relative to the corresponding fixed-final-boundary test.

For independent stage means $X,Y\sim N(\delta ,s^2)$ and combined mean $S=(X+Y)/2$, conditioning on the first-stage sample mean $X$ exceeding $c$ raises that mean. In this equal-stage setting, the estimator bias after continuation is $E(S\mid X\ge c)-\delta =(s/2)\lambda ((c-\delta )/s)$, where $\lambda$ is the upper-tail inverse Mills ratio. It is positive and tends to zero as continuation becomes virtually certain.

Consider independent stage means $X\sim N(\delta ,1/I_1)$ and $Y\sim N(\delta ,1/I_2)$, with positive known information $I_1,I_2$. Continue to stage two only if $X\sqrt{I_1}\ge f$. The pooled mean is $M=wX+(1-w)Y$, where $w=I_1/(I_1+I_2)$. Write $m$ for its observed value. Its conditional mean under a candidate parameter $d$ isHere $\varphi$ and $\Phi$ are the standard normal probability density and cumulative distribution functions. The Conditional bias-corrected normal mean estimate solves $g(d)=m$. This inverts the selection-shifted mean instead of ignoring the fact that the first-stage mean exceeded a threshold.

For the upper-tail inverse Mills ratio $\lambda (a)=\varphi (a)/(1-\Phi (a))$, $g^{\prime }(d)=1-w{\lambda }^{\prime }(f-d\sqrt{I_1})$. The variance of a truncated standard normal variable gives $0<{\lambda }^{\prime }(a)<1$, hence $1-w<g^{\prime }(d)<1$. Since $0<w<1$, $g$ is strictly increasing and ranges over the whole real line; the root exists uniquely. Differentiating the joint normal likelihood conditional on continuation shows that the same root maximizes that conditional likelihood. Inverting a conditional mean does not itself make an estimator conditionally unbiased. The separate Rao-Blackwell estimator after interim selection provides that property in the equal-information setting.

For two independent stage means $X,Y\sim N(\delta ,s^2)$ and continuation $X\ge c$, let $S=(X+Y)/2$ and $v=s/\sqrt{2}$. The second-stage mean is conditionally unbiased. The Rao-Blackwell theorem gives $E(Y\mid S,X\ge c)=S-v\lambda ((c-S)/v)$, using $X\mid S\sim N(S,v^2)$ and truncation. This is a conditionally unbiased estimator using both stages through their combined sample mean.

The Conditional error principle permits adaptation of a testing design if, conditional on the observed interim information, the modified future rule's Type I error does not exceed the original rule's conditional future rejection probability. Averaging then preserves the original unconditional error bound. For a completed single-look test that has not rejected, there is no future conditional error budget under that original design.

An Alpha-spending function assigns the cumulative Type I error budget to successive information fractions in a group sequential design. Its total at the final analysis is the overall significance level. The rejection boundaries must be calibrated using the joint distribution of the sequential test statistics; independent fixed-level tests do not automatically spend the assigned budget correctly.

An interim analysis examines data before a study reaches its planned maximum Sample size.

A Sequential stopping boundary specifies values of an accumulating test statistic that trigger early stopping for efficacy, futility, or harm.

A common approximation for group sequential test statistics represents the accumulating score by $S(I)=\theta I+B_I$, where $I$ is information and $B$ is standard Brownian motion. At fixed information levels $0<I_1<\cdots <I_J$, put $Z_j=S(I_j)/\sqrt{I_j}$. The vector $(Z_1,\dots ,Z_J)$ has a multivariate normal distribution withIndeed independent Brownian increments give $\mathrm{Cov}(B_{I_j},B_{I_k})=I_j$. The statistics share observations, so later and earlier tests are correlated. For two equally sized stages, the interim/final correlation is $1/\sqrt{2}$, not zero. This joint law is used to calculate the probability of crossing any efficacy boundary in a group sequential design. It is exact for the Brownian score model; its use for a general trial statistic is a normal approximation requiring the appropriate information and independent-increment limit assumptions.