Validation of the human activity profile in stroke: a comparison of observed, proxy and self-reported scores

PURPOSE: To assess the concurrent validity of the Human Activity Profile (HAP) with stroke and healthy control individuals, when reported by the subject or a proxy and compared to observed performance, used as gold standard; factors related to discrepancies; as well as to determine variables that could best predict physical activity levels. METHODS: A total of 24 stroke and 23 healthy control subjects took part in the study. Functional outcome measures included personal/demographic factors, HAP, the 10-meter walk, and cognition. RESULTS: Strong relationships were found between self-reported and observed scores for both groups (r=0.89-0.99). The relationships between proxy and observed scores were acceptable for stroke (r=0.80-0.87) but lower for control subjects (r=0.65-0.75), with intraclass correlation coefficients (ICCs) ranging from 0.69-0.89. Significant differences were found between proxy and observed scores for stroke, but not for control subjects. Discrepancies between scores showed little predictability from selected variables. For stroke subjects, the strongest variable explaining physical activity levels was walking speed (r2=43%) and the best combination of variables was walking speed plus cognition (r2=55%). CONCLUSIONS: Self-reported HAP was an excellent measure both for stroke and control subjects and related well to observed performance. Proxy reports appear to be valid for stroke subjects. HAP was significantly related to measures of walking speed.