StatTools : Sample Size for Phase II Study (Fleming's Procedure) Explained and Table
 Introduction Sample Size Table References This page provides the table for sample size required for a Phase II study,based on the program from the Sample Size for Phase II Study (Fleming's Procedure) Program Page . The table provides sample size for : Powers of 0.8, 0.9, and 0.95 Probability of Type I Error (α) of 0.1, 0.05, and 0.01 p0 and pn from 0.05 to 0.95 at 0.05 interval To design a Phase II trial of a new drug or treatment, the existing probability of success (survival or cure) (p0) is defined, and the improved proportion of success with the new treatment being studied (pn) is also defined. Although the table provides for 3 levels of power and probability of Type I Error (α), in most cases, a power of 0.8 and α of 0.05 is used to calculate the sample size. The sample size is then used for the trial. At the end of data collection, if the proportion of success reaches or exceeds pn, then the treatment is deemed worthy of further study, and usually subjected to Phase III trials. If the proportion of success is less than pn, then the new treatment is considered not worth further study, and usually abandoned. Example We are treating the advanced stage of a particularly aggressive type of cancer, and after the current standard surgery and radiotherapy, only 20% (p0=0.2) of the patients survive 12 or more months. We are offered to trial a newly developed chemotherapy, and we decided that a formal large scale Phase III trial would be worth while if the Phase II trial find the 12 months survival rate is doubled (pn = 0.4). We use the statistical parameters α=0.05, power=0.8, p0=0.2, pn=0.4, and determined that the sample size is 29 cases. We administered the additional chemotherapy to 29 patients that fits the description, and 14 (0.48 or 48%) survived at the end of 12 months. As we define success as survival for more than 12 months, We concluded that the results are sufficiently encouraging for a formal phase III trial to be conducted. Had the 12 month survival rate be less than 12 out of our 29 cases (<0.4 or 40%), then we would have abandoned this new treatment and concluded that further study of it is unlikely to be productive.