The maximum absolute difference between the patients allocated to each of the categories within a given factor summarises how well that factor has been balanced by the minimisation process and is used as a summary measure of balance for that factor.
Appropriate starting values for the minimisation process are determined by a standard continuation strategy.
The second part is the sum of the inverse length of all trajectories; hence, in the minimisation process, it favours long trajectories.
The study statistician (AMB) will conduct the minimisation process remotely via email.
Full details of the minimisation process will be published in a separate document with restricted access.
A stochastic element is included in the minimisation process to reduce the chances of predicting the randomisation outcome.
This implies that for minimisation, adjusting only for the covariates used in the minimisation process (and not for their interactions, or equivalently each strata) should give valid type I error rates.
In this method the minimisation process involves defining a trust region where the objective function SQE can be approximated with a simpler function q.
However, it should be noted that the decision of which variables to include in the minimisation process should be informed primarily by the clinical importance of variables and their impact on outcome.
The restoration achieved on the basis of a Wiener scheme is an optimum since the restoration filter is the outcome of a minimisation process.
