A pretrial power analysis for the estimated required sample size was conducted using GPower38 version 3.1.9.6 and effect size was estimated based on published articles9 (link),39 (link),40 (link) for the primary endpoint assuming a normally distributed amputee population. It was therefore expected that 38 subjects were required to complete the protocol with a power of 0,95 and α at 0,05. Drop-out rate was estimated at proximally 20% and therefore 47 subjects were recruited.
We used R version 4.03 (R-Studio Version 1.2.5033) and lme441 to perform a linear mixed effects analysis of the relationship between the OPUS outcomes and clinical need. As fixed effects, we entered age, gender, and evaluation point (tested for interaction with “clinical need”) into the model. As random effects, we had intercepts for subjects and investigators, as well as by-subject and by-item random slopes for the effect of clinical need. P-values were obtained by likelihood ratio tests of the full model with the effect in question against the model without the effect in question. For comparison of individual OPUS items, the Benjamini & Hochberg method was used to control for Type I error due to multiple comparisons.42
We used R version 4.03 (R-Studio Version 1.2.5033) and lme441 to perform a linear mixed effects analysis of the relationship between the OPUS outcomes and clinical need. As fixed effects, we entered age, gender, and evaluation point (tested for interaction with “clinical need”) into the model. As random effects, we had intercepts for subjects and investigators, as well as by-subject and by-item random slopes for the effect of clinical need. P-values were obtained by likelihood ratio tests of the full model with the effect in question against the model without the effect in question. For comparison of individual OPUS items, the Benjamini & Hochberg method was used to control for Type I error due to multiple comparisons.42
Free full text: Click here
