Abstract 2131: Prediction of Discharge Walking Ability from Initial Assessment in a Stroke IRF Population

Abstract

The purpose of this study was to investigate which physical therapy assessments administered at admission to an inpatient rehabilitation facility (IRF) could predict discharge walking ability in a local population of people with stroke. The sample consisted of 111 IRF participants with stroke who consented to have records stored in the Brain Recovery Core database between January 2010 and January 2011. Independent variables from the admission assessment included: lower extremity Motricity Index, somatosensation, Modified Ashworth Scale, Functional Independence Measure (FIM) locomotor item, Berg Balance Scale, walking speed, age, history of previous stroke, and time from stroke onset to IRF assessment. The dependent variable was discharge walking ability, defined as walking speed on the 10 m Walk Test. Correlational analyses examined relationships with and between the admission variables and discharge walking ability. Step-wise multiple regression was used to determine the most parsimonious combination of variables that could explain variance in discharge walking ability. Logistic regression was used to determine the likelihood of achieving household (< 0.4 m/s) versus community (≥ 0.4 - 0.8, > 0.8 m/s) ambulation categories. Results from the stepwise model indicated that 2 admission variables, Berg Balance score and FIM locomotor score, explained 80% of the variance in discharge walking ability. For the logistic model, the odds ratio of achieving only household ambulation at discharge was 20 (95%CI: 6-63) when the combination of having a Berg Balance < 20 and a FIM locomotor score of 1 or 2 was present. The logistic model correctly classified 92% of subjects achieving only household ambulation and 64% of subjects achieving better than household ambulation. These findings suggest that performance on the Berg Balance Scale and the FIM locomotor item at admission to an IRF can explain most of the variance in discharge walking ability. Having the combination of a Berg Balance score < 20 and a FIM locomotor score of 1 or 2 at admission indicates that a person is highly likely to be only a household ambulator at discharge from the IRF. Knowing at the time of IRF admission that a person with stroke is not likely to achieve limited community or community ambulation status will allow for earlier discharge planning with respect to needed assistance at home, durable-medical equipment, and home modifications.