Correction to: Minimum sample size for developing a multivariable prediction model: PART II - binary and time-to-event outcomes (Statistics in Medicine, (2019), 38, 7, (1276-1296), 10.1002/sim.7992)

Abstract

The authors have identified that Equation (13) in this paper contains an error, with a +E term being erroneously specified instead of a −E term. The correct Equation (13) is therefore (Formula presented.) where λ is the constant hazard rate, E is the total number of events, and T is the total time at risk (eg, total person-years). This equation is important toward criterion (ii) of the survival analysis sample size calculation but does not affect the binary outcome situation at all. When using this updated equation within the applied survival example in Section 5.2, the same sample size is identified as in the original paper. The authors would like to thank Glen Martin, Niels Peek, and Shane Collins from the University of Manchester for bringing an example to our attention that led to us identifying this error. Our sample size package pmsampsize in R and Stata has been updated accordingly. Richard Riley, on behalf of all the authors.