The System Usability Scale (SUS) is widely used in the rehabilitation field to assess the satisfaction dimension of digital technologies, although it was not specifically developed for this context.

This systematic review aims to synthesize and appraise the psychometric properties of the SUS in the context of rehabilitation digital technologies, identify its strengths and limitations, and provide evidence-based recommendations for researchers and clinicians.

We assessed methodological quality using the COSMIN Risk-of-Bias checklist (4-point scale), applying the "worst score counts" principle. The certainty of evidence was synthesized using a modified GRADE approach, downgrading for risk of bias, inconsistency, imprecision, and indirectness.

We included 33 studies. Hypothesis testing for construct validity (n = 27) was rated Sufficient (+) with high certainty, showing expected correlations with satisfaction (rs = 0.38 to 0.49) and task performance (r = 0.91). Internal consistency (n = 8) was rated Sufficient (+) with reliability coefficients (α or ω) ranging from 0.63 to 0.91; however, evidence was not graded due to structural inconsistencies. Structural validity (n = 3) yielded an indeterminate result due to conflicting Confirmatory Factor Analysis indices (CFI from 0.816 to 0.960) and a lack of assessment for local independence and monotonicity. Notably, content validity assessment was entirely absent across all included studies.

While the SUS demonstrates robust construct validity, uncertainty regarding its structural validity in rehabilitation contexts suggests that the total sum score cannot be fully trusted as a unidimensional measure. Furthermore, the lack of evidence for content validity is a major gap, as the scale's relevance for patients with cognitive or communication impairments remains unverified. Clinicians should interpret SUS results with caution, triangulating them with qualitative data to ensure specific clinical goals of effectiveness and efficiency are met. Future research must prioritize content validation and rigorous structural analysis using modern psychometric techniques.