Early stopping of aerospace medical trials: application of sequential principles

A two-period, crossover trial was conducted in the hypobaric chamber on human subjects to compare the influence of inflight exercise (experimental) and restricted activity (control) on altitude decompression sickness (DCS) during simulated extravehicular activities. Out of 39 pairs (total of 78 exposures), 4 cases of DCS occurred under control and 5 occurred under experimental conditions. Analysis of the crossover results showed that the P values for differences in DCS occurrence was 0.56. Under these circumstances, it was necessary to decide whether additional information would be obtained by accruing more subjects. This problem was examined by using a skew sequential design in which the "stopping rule" was based on an alpha of 0.05 (one-sided) and power of 80%. The result of this analysis was in favor of the null hypothesis, and the trial was terminated. The authors recommend the use of similar stopping rules in aerospace trials to optimize sample size without compromising statistical validity.