The lack of a unified approach to outcome assessment in stroke rehabilitation limits our ability to interpret evidence provided by randomized controlled trials (RCTs). The purpose of this review was to identify outcomes and assessment tools reported in RCTs of stroke rehabilitation interventions as a first step toward consistent assessment of outcomes. Given that the validity of research is linked to reliability and validity of measurement, the relationship between the use of previously developed outcome measures and the methodological quality of RCTs was explored. Electronic literature searches identified RCTs examining stroke rehabilitation therapies from 1968 to 2005. The Physiotherapy Evidence Database (PEDro) scale was used to assess methodological quality. Cited outcomes were recorded and assessment tools identified as previously published or study specific. Four hundred ninety-one RCTs cited the assessment of 1447 outcomes using 489 measurement tools. Two hundred fifty-four of these were previously published, and 235 were study specific. A core of 30 frequently cited tools was identified. The use of previously published assessment tools to evaluate primary study outcomes was associated with higher PEDro scores. Significant heterogeneity in outcome assessment was demonstrated, although a core of 30 frequently cited tools could be identified. Appropriate evaluation and selection of outcome measures would enhance the methodological quality of randomized controlled trials.