JMIR Publications

ABSTRACT

Background:

Traumatic Brain Injury (TBI) is a leading cause of disability across all ages worldwide. TBI patients are characterized by a wide heterogeneity, considered one of the most significant barriers to finding effective therapeutic interventions. Cluster analysis (CA) allows for identification of homogeneous subgroups based on similarities in performance on neuropsychological tests. CA has been scarcely applied in web-based rehabilitation treatments, considering different CA approaches and cluster validity indices (CVI).

Objective:

1) apply state-of-the-art CVIs to different cluster strategies (hierarchical, partitional and model based) to identify meaningful classes 2) Apply combined strategies of dimensionality reduction and clustering by means of principal components analysis and random forests to improve obtained CVIs and perform stability assessment to the final profiles 3) Characterize the identified profiles by means of demographic and clinically relevant variables 4) Study the external validity of the obtained clusters considering three relevant aspects of TBI rehabilitation: Glasgow Coma Scale (GCS), functional independence (FIM) in activities of daily life assessment at admission of rehabilitation and cognitive rehabilitation tasks executed all along the rehabilitation process

Methods:

CA was performed with cluster, factoextra and mclust R packages. We run combined strategies using the factominer and random forest R packages. We performed stability analysis using fpc R package. Normality of the data was checked by Shapiro-Wilk normality test, visual inspection of QQ-Plots and histograms, as well as skewness and kurtosis. Homogeneity of Variance was checked by Levene's Test. Between groups comparisons were performed using T-test, chi-square test or Mann-Whitney U test as appropriate.

Results:

We analyzed 574 adult (mostly severe) TBI patients undergoing web-based rehabilitation, the whole period under study is September 2008 to April 2019. We identified and characterized 3 clusters with strong internal validation and stability: 1) moderate attentional impairment and moderate disexecutive syndrome with mild memory impairment and normal spatio-temporal perception being 66% highly educated (P<.05) 2) severe disexecutive syndrome with severe attentional and memory impairments and normal spatio-temporal perception being 49.2% highly educated (P<.05) 3) very severe cognitive impairment, being 45.2% highly educated (P<.05). We externally validated them with severity of injury (P<.006), and functional independence assessments: cognitive (P< .001), motor(P<.001) and total(P<.001). We mapped 151,763 web based cognitive rehabilitation tasks performed by the 574 participants during the whole period (all cognitive functions) to the 3 obtained clusters (P<.001) and confirmed the identified patterns

Conclusions:

CA in web-based cognitive rehabilitation treatments allows for increasing the sample of participants in relation to state-of-the-art publications, identifying and characterizing strong patterns of response to neuropsychological tests and externally validating the obtained clusters using important aspects of TBI rehabilitation, such as the set of cognitive web based tasks available in the web platform, in order to tailor them to each profile.