Background: Self-reported functional status is a commonly used health measure in clinical settings, yet the optimal approach for assessing function is often debated.
Objective: To examine the agreement between a self-reported and a performance-based measure of function and the relative ability of each measure to predict long-term health outcomes.
Design: Prospective cohort study.
Setting: 20 hospitals in Connecticut and Massachusetts.
Participants: 620 postmenopausal women (46 to 91 years of age) who had experienced a stroke or transient ischemic attack.
Measurements: A self-reported and a performance-based measure of function were assessed at baseline (before intervention) by using the Barthel index and the Physical Performance Test.
Results: Disagreement between the self-reported and performance-based measure of function was common (slight disagreement, 55.0%; substantial disagreement, 19.3%). Most women (95.4%) overreported their level of function. Women who were more clinically impaired (risk ratio [RR] for more comorbid conditions, 1.52 [95% CI, 1.17 to 1.97]; RR for recent stroke, 2.33 [CI, 1.45 to 3.73]; and RR for cognitive impairment, 1.76 [CI, 1.34 to 2.32]); who were less educated (RR = 1.30 [CI, 1.02 to 1.67]); and who were of nonwhite ethnicity (RR 1.43 [CI, 1.07 to 1.91]) were more likely to overreport their level of function. An impaired performance-based measure of function predicted subsequent stroke or death (hazard ratio, 1.50, [CI, 1.06 to 2.11]); however, an impaired self-reported measure of function was not likely to predict these outcomes.
Conclusions: Clinicians should be aware that results of self-reported and performance-based measures of function can differ in women who have experienced a recent cerebrovascular event. Although more difficult to collect, results of a performance-based measure may provide information about long-term health outcomes that is not available from a self-reported measure.