# Table 3 Number of failures needed to stop a drug strategy group for each number of analysed patients

Analysed patients Minimum number of observed failures needed to stop groupa Maximum probability of stopping group if true cure rate equals: Minimum probability of stopping group if true cure rate equals:
90% 95% 90% 80% 70% 60%
3–7 3 0.026 0.004 0.001 0.008 0.027 0.064
8–13 4 0.034 0.003 0.005 0.056 0.194 0.406
14–20 5 0.043 0.003 0.009 0.130 0.416 0.721
21–26 6 0.040 0.002 0.014 0.231 0.637 0.904
27–33 7 0.042 0.001 0.015 0.287 0.744 0.958
34–39b 8 0.037 0.001 0.017 0.367 0.844 0.986
40–41b 8 0.048 0.001 0.042 0.563 0.945 0.998
42–48 9 0.046 0.001 0.021 0.469 0.920 0.997
49–55 10 0.044 0.0004 0.022 0.528 0.952 0.999
56–63 11 0.047 0.0003 0.021 0.580 0.971 1.000
64–71 12 0.048 0.0002 0.023 0.648 0.985 1.000
72–78 13 0.045 0.0001 0.025 0.705 0.993 1.000
1. Groups will recruit a minimum of 78 patients with 39 patients in each stratum if recruitment is not stopped into that group. If recruitment is stopped into one group, the maximum number of patients in each other group and the other stratum for that group will be higher
2. Maximum and minimum probabilities are for the range of analysed patients
3. a More than 0.95 posterior probability of the true cure rate being less than 90% (Pr(true cure rate < 0.9|x) > 0.95, where x is the data currently observed, with the beta (4.5, 0.5) prior which has mean = 0.9 and variance = 0.015
4. b These rows not pooled despite the same number of minimum failures to provide information about the fully recruited strata (n = 39)