Cognition assessments to predict inpatient falls in a subacute stroke rehabilitation setting

ABSTRACT

Background: Stroke-related falls occur at especially high rates in rehabilitation settings. Inpatient-hospital falls have been identified as one of the most common medical complications after stroke, negatively influencing recovery, nevertheless, the role of cognition in relation to falls during inpatient rehabilitation is largely unexplored.

Objective. We aim to predict inpatient falls in a subacute stroke rehabilitation setting using previously reported variables such as stroke severity, gender, age, ataxia, hemiparesis, and functionality in activities of daily living, further extending them with specific cognition variables assessing memory, verbal fluency, attention, and orientation.

Methods: This observational study included 158 stroke patients admitted to a rehabilitation center between 2007 and 2019, with less than 30 days since stroke onset to admission. Stroke severity was assessed using the National Institutes of Health Stroke Scale (NIHSS). Four logistic regressions were performed including NIHSS, age, sex, ataxia, and hemiparesis plus one of the following: (1) Functional Independence Measure cognitive (C-FIM) and motor (M-FIM) subtests. (2) individual C-FIM items, (3) Ray Auditory Verbal Memory Test (RAVLT) and (4) verbal fluency test (PMR), Digit Span from Wechsler Adult Intelligence Scale (WAIS III), and Orientation from Test Barcelona.

Results: Neither NIHSS, age, sex, ataxia nor hemiparesis predicted falls. C-FIM was a significant predictor (AUC:0.891), but not M-FIM. The problem solving C-FIM item (AUC:0.836), the RAVLT learning subtest (AUC:0.879), and PMR verbal fluency (AUC:0.871) were significant predictors for each model, respectively.

Conclusions: Cognition assessments, i.e., one FIM item, one RAVLT item, or a one-minute verbal fluency test are significant falls predictors.

KEYWORDS:

• Falls
• verbal memory
• rehabilitation
• verbal fluency
• predictors
• functional independence

Acknowledgments

Special thanks to Toni Ustrell from Institut Guttmann’s Nursery Department for his support with falls protocols and to Jaume Lopez from Institut Guttmann's Research and Innovation Department for data access.

Supplementary material

Supplemental data for this article can be accessed here.

Additional information

Funding

This research was partially funded by EU H2020 PRECISE4Q - Personalized Medicine by Predictive Modeling in Stroke for better Quality of Life [Grant Agreement 777107 – Research and Innovation Action].