Research Article: Regression techniques employing feature selection to predict clinical outcomes in stroke

Date Published: October 19, 2018

Publisher: Public Library of Science

Author(s): Yazan Abdel Majeed, Saria S. Awadalla, James L. Patton, Yih-Kuen Jan.


It is not fully clear which measurable factors can reliably predict chronic stroke patients’ recovery of motor ability. In this analysis, we investigate the impact of patient demographic characteristics, movement features, and a three-week upper-extremity intervention on the post-treatment change in two widely used clinical outcomes—the Upper Extremity portion of the Fugl-Meyer and the Wolf Motor Function Test. Models based on LASSO, which in validation tests account for 65% and 86% of the variability in Fugl-Meyer and Wolf, respectively, were used to identify the set of salient demographic and movement features. We found that age, affected limb, and several measures describing the patient’s ability to efficiently direct motions with a single burst of speed were the most consequential in predicting clinical recovery. On the other hand, the upper-extremity intervention was not a significant predictor of recovery. Beyond a simple prognostic tool, these results suggest that focusing therapy on the more important features is likely to improve recovery. Such validation-intensive methods are a novel approach to determining the relative importance of patient-specific metrics and may help guide the design of customized therapy.

Partial Text

Recovering from stroke is a highly variable process [1] that is difficult to influence or predict. There are many clinical assessments to evaluate the state of a patient and gauge his or her long-term prognosis. Some assessments are sufficiently reliable [2], though there is no widely accepted gold standard [3, 4]. In practice, a battery of clinical evaluations are conducted, each used to assess a different aspect of a patient’s condition. Common assessment areas include: (1) motor ability, such as Fugl-Meyer [5]; (2) functional performance, such as Wolf Motor Function Test [6]; and (3) self-reported motor activity, as in the case of the Motor Activity Log [7] and the Functional Independence Measure [8]. There is no consolidated outcome measure that encompasses these disparate evaluations, and there is general consensus that a combination of assessments provides the best profile of a patient [9].

Only some models were able to effectively predict clinical outcomes (WMFT and UEFM) using quadratic polynomials of the features given in Table 1, and pairwise interactions (see Methods). Next, we used only those successful models to identify and rank salient predictors of these clinical outcomes, and these rankings were consistent in 4-fold cross-validation with 100 repeats. These two steps are described in more detail in the sections below.

Changes in motor ability (UEFM) and motor function (WMFT) can both be predicted by our models. Change in WMFT is easier to predict since it is a continuous measure. Both changes in UEFM and WMFT can be linked to specific movement features as well as patient demographics and clinical characteristics. Our validation approach also allowed us to measure the certainty of our findings. Since we are unable to affect demographics or clinical characteristics, features that we can influence during rehabilitation are the most critical. This work suggests that speed would be a good first target for further study.




0 0 vote
Article Rating
Notify of
Inline Feedbacks
View all comments