BACKGROUND AND PURPOSE-Knowledge about robust and unbiased factors that predict outcome of activities of daily living (ADL) is paramount in stroke management. This review investigates the methodological quality of prognostic studies in the early poststroke phase for final ADL to identify variables that are predictive or not predictive for outcome of ADL after stroke. METHODS-PubMed, Ebsco/Cinahl and Embase were systematically searched for prognostic studies in which stroke patients were included ≤2 weeks after onset and final outcome of ADL was determined ≥3 months poststroke. Risk of bias scores were used to distinguish high-and low-quality studies and a qualitative synthesis was performed. RESULTS-Forty-eight of 8425 identified citations were included. The median risk of bias score was 17 out of 27 (range, 6-22) points. Most studies failed to report medical treatment applied, management of missing data, rationale for candidate determinants and outcome cut-offs, results of univariable analysis, and validation and performance of the model, making the predictive value of most determinants indistinct. Six high-quality studies showed strong evidence for baseline neurological status, upper limb paresis, and age as predictors for outcome of ADL. Gender and risk factors such as atrial fibrillation were unrelated to this outcome. CONCLUSIONS-Because of insufficient methodological quality of most prognostic studies, the predictive value of many clinical determinants for outcome of ADL remains unclear. Future cohort studies should focus on early prediction using simple models with good clinical performance to enhance application in stroke management and research. © 2011 American Heart Association, Inc.