When does an interim analysis not jeopardize the type 1 error rate? The attached poster by Per Broberg of Lund University offers an answer based on a 2011 paper by Mehta & Pocock, and a 2013 paper by Broberg.
Broberg’s poster explains why type 1 error is preserved when employing adaptive sample size re-estimation designs. Such designs allow for one interim look that determines whether a trial ought to be stopped for futility, stopped for efficacy, continued as planned, or continued with a larger sample size. Such trials can be quite attractive both for the resulting high power, and the fact that potential investors can mitigate financial risks based on interim decisions.
According to Broberg: “The decision at the interim look whether or not to raise the sample size only requires calculation of a test statistic which (approximately) follows a standard normal distribution. The theory for this has shown that the type 1 error rate remains intact if the results show promise, meaning that the test statistic exceeds the threshold which depends on the number of observations at the interim, the planned final sample size and the increase considered.”
The poster was presented at ISCB 2014 in Vienna.
Related Items of Interest
Broberg, P. (2013). Sample size re-assessment leading to a raised sample size does not inflate type I error rate under mild conditions. BMC medical research methodology, 13(1), 94.
Mehta, C. R., & Pocock, S. J. (2011). Adaptive increase in sample size when interim results are promising: A practical guide with examples. Statistics in medicine, 30(28), 3267-3284.