The definition of appropriate treatment effect parameters, corresponding minimal clinically important differences and the implementation of appropriate sample size estimation and inference procedures is one of the important open issues in the design and analysis of spinal cord injury trials [Steeves et al., 2015]. The problem arises from the ordinal nature of many potential primary endpoints, such as the upper extremity motor score (UEMS).
The objective of this work package is to develop a statistical model for the conditional UEMS distribution in two-armed randomised clinical trials. Based on this model, minimal clinically important differences shall be discussed, and methods for sample size estimation, statistical inference and subgroup analyses shall be developed. In this project, we develop a statistical framework allowing future clinical trials using UEMS as primary endpoint to be planned and analysed with procedures taking the specific properties of this endpoint into account.