Reliability and sensitivity to change assessed for a summary measure of lower body function

Abstract

Original Article

A summary performance measure comprised of a hierarchical balance task, a 4-meter walk, and five repetitive chair stands is increasingly being used as a predictor of independent living for older persons. The reliability and sensitivity to change of this summary performance measure have not been investigated, however. Because a measure can be reliable while being unresponsive to change, this study presents information on both the reliability and sensitivity to change for the summary performance measure. This is a 3-year prospective cohort study of 1,002 moderately to severely disabled older women. Short- and long-term reliability was assessed by intraclass correlation coefficients (ICC). Sensitivity to change was assessed by slope differences for three age categories (65–74, 75–84, and ⩾85) over six 6-month follow-up periods. Sensitivity to change was also assessed by summary performance change scores for those who did and did not suffer from one of four medical events [myocardial infarction (MI), stroke, hip fracture, or congestive heart failure (CHF)] at follow-up. The summary performance measure showed excellent reliability. Intraclass correlation coefficients ranged from 0.88 to 0.92 for measures made 1 week apart. The 6-month average intraclass correlation coefficient was 0.77 (range 0.72–0.79). The summary performance measure was also highly responsive to change. Subjects who suffered an incident MI, stroke, hip fracture, or CHF at follow-up were significantly more likely to have poorer summary performance change scores (−2.25) compared with those who did not have one of these medical events (−0.24). Additionally, subjects who suffered one of these events improved their summary performance scores in the following assessment period by 0.72. With increasing utilization of the summary performance measure by researchers and clinicians it is important that the measurement properties of this instrument are known. Our results show that the summary performance measure has excellent reliability and is highly sensitive to change.

Introduction

Measures of lower body function provide useful knowledge about preclincal functional decrements in older persons who report little to no disability, and enhance our understanding of the causal pathway from disease to disability 1, 2, 3, 4. One measure of lower body function is the summary performance measure, which is comprised of a hierarchical balance task, a 4-meter walk, and five repetitive chair stands [4]. This objective measure is increasingly being used on older subjects to predict risk for loss of ability to live independently in a community setting 1, 2. Guralnik et al. [1] showed the summary performance measure to be highly predictive of subsequent disability in a group of initially nondisabled individuals aged 70 years or older. Individuals with the lowest scores on the summary performance measure were more than four times as likely to have disability in Activities of Daily Living (ADLs) and mobility-related disability 4 years later compared with those with the highest scores [1]. Similarly, in a group of initially nondisabled older Mexican Americans, those with the lowest scores on the summary performance measure were more than six times as likely to have ADL disability, and were approximately five times as likely to have mobility-related disability 2 years later [2]. Additionally, lower scores on the summary performance measure are associated with an increased risk for nursing home admission, health care utilization, and death in older populations 4, 5, 6.

Objective measures of physical performance may offer several advantages over self-reports, including increased validity, and less confounding from culture, language, and educational level [7]. Additionally, they can provide information that complements a physician's clinical examination [3]. However, the summary performance measure's reliability and sensitivity to change have not been formally evaluated. The two objectives of this study, therefore, are to assess the reliability of the summary performance measure, and its sensitivity to change in a group of moderately to severely disabled older women.