OBJECTIVE: To develop a valid and reliable spinal cord injury (SCI) specific physical functioning (PF) scale for the Veterans Health Administration (VHA) version of the 36-Item Short-Form Health Survey. DESIGN: A mixed qualitative and quantitative research design was used. In phase 1, a pool of SCI-specific PF items was generated based on focus groups with patients and health care providers. In phase 2, the psychometric properties of the SCI-specific PF scale were established. SETTING: A VHA SCI center. PARTICIPANTS: The sample consisted of valid responses from 359 veterans with traumatic SCI who were seen at a VHA SCI center during the prior year (2002). INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURE: Physical functioning in people with SCI. RESULTS: Exploratory factor analysis was conducted separately on respondents with lower neurologic-level injuries (paraplegia, 53% [n=190]) and those with higher neurologic-level injuries (tetraplegia, 45% [n=163]) and identified 9 items loading on 1 factor in both groups. These 9 items were included in separate item response theory (IRT) model analyses for each subgroup. Based on the IRT analysis, 1 item was eliminated, resulting in an 8-item, SCI-specific PF scale. CONCLUSIONS: Although several of the items in the SCI-specific PF scale showed floor effects, particularly in people with tetraplegia, we found excellent reliability and strong support of convergent and divergent validity of the scale.
OBJECTIVE: To develop a valid and reliable spinal cord injury (SCI) specific physical functioning (PF) scale for the Veterans Health Administration (VHA) version of the 36-Item Short-Form Health Survey. DESIGN: A mixed qualitative and quantitative research design was used. In phase 1, a pool of SCI-specific PF items was generated based on focus groups with patients and health care providers. In phase 2, the psychometric properties of the SCI-specific PF scale were established. SETTING: A VHA SCI center. PARTICIPANTS: The sample consisted of valid responses from 359 veterans with traumatic SCI who were seen at a VHA SCI center during the prior year (2002). INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURE: Physical functioning in people with SCI. RESULTS: Exploratory factor analysis was conducted separately on respondents with lower neurologic-level injuries (paraplegia, 53% [n=190]) and those with higher neurologic-level injuries (tetraplegia, 45% [n=163]) and identified 9 items loading on 1 factor in both groups. These 9 items were included in separate item response theory (IRT) model analyses for each subgroup. Based on the IRT analysis, 1 item was eliminated, resulting in an 8-item, SCI-specific PF scale. CONCLUSIONS: Although several of the items in the SCI-specific PF scale showed floor effects, particularly in people with tetraplegia, we found excellent reliability and strong support of convergent and divergent validity of the scale.