Ecological Momentary Assessment of Real-World Functional Behaviors in Individuals With Stroke: A Longitudinal Observational Study.

ABSTRACT

OBJECTIVE: To validate and characterize real-world functional behaviors in individuals after stroke. DESIGN: Longitudinal observational study using ecological momentary assessment (EMA) as a real-time assessment of functional behaviors in natural contexts. Wilcoxon rank-sum tests, Fisher exact tests, and Spearman correlations were used to analyze data. SETTING: Community. PARTICIPANTS: Individuals with mild to moderate stroke (N=212). INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURES: Individuals were assessed 5 times daily for 14 days with EMA surveys to determine what, with whom, and where individuals were doing activities and appraise mental, somatic, and cognitive symptoms. Individuals also completed standardized assessments during laboratory visits, including Lawton Instrumental Activities of Daily Living Scale, Activity Card Sort, Patient-Reported Outcome Measurement Information System, and Quality of Life in Neurological Disorders. RESULTS: Most individuals (median age, 60 years; 55% male) were ischemic stroke (90%) and had mild stroke severity (median National Institutes of Health Stroke Scale, 2). A total of 14,140 EMA surveys were analyzed. Individuals were home 78% of the time; primarily participated in passive, unproductive activities (27%), especially watching television and resting; and participated least in physical activities (4%). EMA was sensitive to indicators of poststroke disability; unemployed individuals reported fewer vocational activities but more activities of daily living (ADL) and passive activities than employed counterparts. Users of mobility devices and individuals with cognitive problems spent significantly less time on vocational activities and more on ADL than nonusers and those without cognitive problems. Our data supported the validity of EMA methods in stroke, with small to moderate correlations of EMA with in-laboratory measures of daily functioning (r=-0.30 to 0.35, P<.05) and very large correlations between EMA and in-laboratory measures of symptoms, especially those measuring same constructs (r=-0.64 to 0.79, P<.0001). CONCLUSIONS: Our findings reveal that EMA tracked poststroke functioning precisely. EMA may be beneficial in examining poststroke functional recovery, in monitoring patients for home-based interventions, and for longitudinal research.