Northwest Biotherapeutics Numerical Analysis: - published 28th February 2020
Here, I use a markov model to simulate the trial by randomly picking 331 patients with replacement from the reverse engineered data sets from the blended DCVax-L phase III data set (liau et al., 2018) and a approporate control data set (Lai et al., 2011, Brnades et al., 2008 or both). After the simulation I perform the log-rank statistical test to ask whether the difference between the NULL data and the published interim DCVax-L phase III data (Reverse engineered from Liau et al., 2018) is signifciant. The log-rank stat is applied to the resuling KM data. In the near future I will add the Fleming-Harrington 0,1 and Max-Combo test statisitics to these results. Furthermore, I will restrict the markov modelling to the null data comparing to the fixed DCVax-L phase III blended data since this is a given ground truth and does not need to be modelled.