Admission ambulation velocity predicts length of stay and discharge disposition following stroke in an acute rehabilitation hospital

Abstract

Background and purpose: Poststroke functional outcome and discharge disposition are influenced by age, lesion location and size, severity of neurological insult, prior functional ability, and social support. The effect of admission ambulation velocity on length of stay and discharge disposition has not been previously reported.

Design: Prospective, cohort study.

Setting: Designated acute stroke rehabilitation unit.

Method: 373 patients consecutively admitted to a designated inpatient stroke rehabilitation unit were studied. The study sample was divided into 2 groups, based on admission ambulation velocity (meters/second) during a 2-min timed walk test. Fast ambulators had an ambulation velocity of greater than 0.15 m/s, whereas slow ambulators had an ambulation velocity of 0.15 m/s or less, and also included nonambulators.

Interventions: None.

Outcome measures: Primary outcome measures were length of stay and discharge disposition. Secondary outcome measures were change in the Functional Independence Measure (FIM) scores and change in ambulation velocity per week. Continuous, ordinal, and nominal variables were analyzed using the Student t test, Mann-Whitney U test, and chi-square test, respectively. Sensitivity, specificity, and positive and negative predictive values assessed admission ambulation velocity as a predictor of discharge disposition.

Results: Patient variables for slow ambulators (n = 226) versus fast ambulators (n = 147) were as follows: age (68 +/- 13 SD vs. 69 +/- 12, P = 0.32), male-female ratio (100:126 vs. 78:69, P = 0.09), admission total FIM score (52 +/- 17 vs. 77 +/- 16, P < 0.0001), change in total FIM score (20 +/- 12 vs. 16 +/- 12, P < 0.003), change in ambulation velocity per week (0.05 +/- 0.06 vs. 0.13 +/- 0.30, P < 0.0001), length of stay in days (30 +/- 28 vs. 17 +/- 19, P < 0.0001), and discharge disposition: home/skilled nursing facility (133/91 vs. 133/13, P < 0.0001). Sensitivity, specificity, and positive and negative predictive values (based on chi-square analyses) for admission fast ambulators as a predictor of home discharge were 0.5, 0.87, 0.91, and 0.41, respectively. Admission ambulation velocity alone correctly identified discharge disposition in 78% of the patient population, based on logistic regression analysis (P < 0.0001).

Conclusion: Admission ambulation velocity can predict length of stay and discharge disposition poststroke. This effect is independent of age and admission total FIM score.