A0208
Title: Arguments for the biological and predictive relevance of the proportional recovery rule
Authors: Jeff Goldsmith - Columbia University (United States) [presenting]
Abstract: The proportional recovery rule (PRR) posits that most stroke survivors can expect to reduce a fixed proportion of their motor impairment. The PRR explicitly relates change scores to baseline values as a statistical model. This approach arises in many scientific domains but can potentially introduce artefacts and flawed conclusions. Approaches are described that can assess associations between baseline and changes from baseline while avoiding artefacts due to mathematical coupling or regression to the mean. Methods that can compare different biological models of recovery are also described. Across several real datasets in stroke recovery, evidence for non-artifactual associations between baseline and change is found, and support for the PRR is compared to alternative models. A statistical perspective is also introduced that can be used to assess future models in the conclusion that the PRR remains a biologically relevant model of stroke recovery.
