Principal Investigators
Other Investigator(s)
Project Abstract/Overview
Progress and Outcomes

Justin Carter, MS
(National Rehabilitation Hospital)

Michael J. Rosen, PhD
(University of Vermont)

Linsey Baker, BS
(Virginia Polytechnic Institute)

Project Abstract/Overview

Assessment of patients' functional status in a rehabilitation setting is typically done by an observing therapist using a subjective rating scale such as the FIM, Functional Independence Measure. As a method to introduce objectivity and "ecological validity" into the assessment process, ATRC engineers at NRH installed sensors into the Independence Square (a real-life home and community setting) which was constructed in the hospital in March 1998. The aim of the long-term project is to add objective data to a patient's chart without requiring additional time or effort from the physical, occupational, or speech therapist. The current project is a pilot for future work.

Progress and Outcomes

Main accomplishments:

•  completion of data collection (5 subjects)

•  identification of significant Performance Metrics

•  completion of data reduction and analysis

•  initiation of paper to be submitted to peer-reviewed journal

In 2004, project investigators recruited ited and tested the remaining two subjects needed to complete the 5-subject pilot study. Following the testing of the final two subjects, investigators moved to reducing and analyzing the data. The first step of this data reduction was the time and labor intensive removal of all identifiable characteristics from the videos taken while the subjects partook in the study. Once this was successfully completed, reducing the sensor data was begun.

While the sensor data was being reduced, researchers defined a preliminary set of six potential Performance Metrics (PMs). The six PMs are presented below. All of the values were designed so that a higher score indicated better performance; hence 1/(total time) as opposed to simply total time is the metric used.

Preliminary data analysis indicates that PM3 regressed best against the AMPS (Assessment of Motor Performance) Motor score, while PM1 and PM6 were both statistically significant with respect to the AMPS Process score. All six PMs showed promising correlation coefficients for both Motor and Process scores.