ABSTRACT
Fixed-sample-size, sequential, and stagewise, group sequential plans are presented. The one-sample case arises in two contexts. The first context is when the standard rate is externally imposed. For example, one requirement for efficacy of a candidate therapy regimen may be that test animals exposed to a toxicant dose level of 2 X LDso for unprotected animals and subsequently treated with the test regimen experience at most a 15 percent lethality rate. The second context is when a candidate therapy regimen is to be compared with a standard therapy regimen for which considerable experience has been accumulated. For example, a second requirement for efficacy of a candidate therapy regimen may be that test animals exposed to a toxicant dose level of LDso for animals treated with the standard regimen and subsequently treated with the candidate regimen test program experience a lethality rate no greater than that for animals given the standard regimen. If considerable past experience with the screen is available and if experimental conditions are stable over time, the lethality results for a given toxicant and standard treatment regimen can be pooled over time; the estimated lethality rate based on pooled results will be close to 50% and will have a very small standard error. The lethality rate observed in animals treated with the candidate therapy regimen will thus be compared with this 50% lethality yardstick.
