Towards Personalized Web-based Cognitive Rehabilitation for Patients with Ischemic Stroke: An Elo Rating Approach (Preprint)

BACKGROUND Stroke is a worldwide cause of disability, 40% of stroke survivors sustain cognitive impairments, most of them follow inpatient rehabilitation at specialized clinical centers. Web-based cognitive rehabilitation tasks are already integrated into clinical settings. The impact of a task execution depends on the ratio between the skills of the treated patient and the challenges imposed by the task itself. Thus, treatments personalization requires a trade-off between patients’ skills and tasks difficulties, which is still an open issue. In this work we propose Elo ratings to support clinicians in representing patients’ skills and supporting tasks assignations to optimize rehabilitation outcomes. OBJECTIVE i) perform a stratification of patients with ischemic stroke at early stage of rehabilitation in three levels according to their Elo rating ii) show the relationships between the Elo rating levels, tasks difficulty levels and rehabilitation outcomes iii) determine if Elo rating obtained at early stages of rehabilitation is a significant predictor of rehabilitation outcomes. METHODS The PlayerRatings R library was used to obtain the Elo rating for each patient. Working memory was assessed using the DIGITS subtest of Test Barcelona and the Rey Auditory Verbal Memory Test (RAVLT) was used to assess verbal memory. The three subtests of RAVLT were used: RAVLT learning (RAVLT075), free-recall memory (RAVLT015) and recognition (RAVLT015R). Memory predictors were identified using forward stepwise selection to add covariates to the models which were evaluated by assessing discrimination using the area under the receiver operating characteristics curve (AUC) for logistic regressions and adjusted R2 for linear regressions. RESULTS Three Elo levels (Low, Mid and High) with the same number of patients (n=96) in each Elo group, were obtained using the 50 initial tasks executions (from a total of 38,177) for n=288 adult patients consecutively admitted for inpatient rehabilitation in a clinical setting. The highest proportion of patients that improved in all 4 memory items were from Mid Elo level: 56.7% of them improved in DIGITS, 67.1% in RAVLT075, 58.8% in RAVLT015 and 53.7% in RAVLT015R (p < 0.001). The proportion of patients from the Mid Elo level that performed tasks at difficulty levels #1, #2 and #3 were: 32.1%, 31.0% and 36.9% (p < 0.001) respectively, showing the highest match between skills (represented by Elo level) and tasks difficulties, considering the set of 38,177 tasks executions. Elo ratings were significant predictors in 3 of the 4 models and quasi-significant in the other. When predicting RAVLT075 and DIGITS at discharge we obtained R2=0.54 and R2=0.43 respectively, meanwhile in RAVLT075 and DIGITS improvement predictions we obtained AUC= 0.73, 95% CI(0.64-0.82) and AUC= 0.81 95%CI(0.72-0.89). CONCLUSIONS Elo ratings can support clinicians at early rehabilitation stages in identifying cognitive profiles that can be used to assign tasks’ difficulty levels.