Data mining for response shift patterns in multiple sclerosis patients using recursive partitioning tree analysis

Abstract

PCS and MCS change scores are obfuscated by response shifts. The contingent true scores for PCS change scores are not comparable across patient groups.

Response shift in patients with multiple sclerosis: an application of three statistical techniques

“…As noted above, three different statistical techniques were used for response shift detection: (1) Structural Equation Modeling [44]; (2) Latent Trajectory Analysis [43]; and (3) Recursive Partitioning and Regression Tree modeling [45]. These statistical techniques were used to capture response shift because they are the statistical secondary analytic approaches that are the current state of the art in response shift detection.…”

“…For all analyses, the response shift effects were intended to be detected in the overall PS score and the SF-12 composite scores (mental and physical health) because both of these are multiple-item scales, yielding continuous scores that would likely be more reliable and yield better power in statistical comparisons [58]. For more detail, the interested reader is referred to the preceding three papers in this special section [43][44][45].…”

“…The objective of this article is to discuss the application of the three statistical methods, highlighting differences in requirements, use of data, findings, and interpretability. While we will describe the three methods in some detail, a fuller description of three methods can be found in the articles preceding this article [43][44][45]. On the basis of this comparison, directions for future research will be suggested.…”

Gait Features and their Relationships to Physiological Attributes of Multiple Sclerosis

“…Bejarano et al [65] assess the diagnostic accuracy of clinical variables, including expanded disability status scale (EDSS), using various computational classifiers and employ data mining for feature selections, for predicting the MS course. Li and Schwartz [12] use a partitioning tree analysis technique to identify response shift patterns in persons with MS. Other applications of data mining in bioinformatics include probe selection for gene-expression arrays [66], plant genotype discrimination [67], experiments with automatic cancer diagnosis [68], protein annotation [69], and identification and prediction of drug-induced nausea [70].…”

“…Each type of assessment has important limitations, such as being qualitative, prone to floor and ceiling effects, dependent on physicians' judgment in interpreting outcomes of imaging and clinical symptoms, affected by the differences among individuals regarding the notion of disability or improvements, time-consuming, being imprecise (over-/under-reporting, missing data being difficult to interpolate), having low variance across ratings, limited sensitivity to subtle changes in functions, affected by the response shifts [12] due to changes within individuals regarding health standards, and being burdensome for the persons with MS. Furthermore, severity and disability assessments are neither conclusive nor based on a simple yes or no criterion [6].…”