Introduction: This paper evaluates audits and feedback as methods to increase implementation of evidence in stroke rehabilitation. Method: The study used an action research approach and theories of knowledge translation. A sample of 22 occupational therapists participated from two Danish hospitals that admitted stroke patients. Data collection methods included audits of occupational therapy medical records, documentations of daily practice, and collaborative discussions. Active feedback and discussions of the findings took place, at a group level in four local clinical audits. Data analysis of daily self-reported recordings and audits was descriptive. Audit data were analysed using descriptive statistics. A phenomenological hermeneutical interpretive methodology was used for analysing qualitative data. Findings: Audits and feedback were based on clear standards and contextual developing action plans. Daily practice in both settings adapted to the clinical guidelines. Implementations of the standardized assessment tools seemed to be the most successful. Conclusion: The effects of audit and feedback profited from the active participation of the therapists, as well as local gatekeepers having formal responsibilities for implementing change. The process was strengthened by providing the audits and feedback more than once. The effect of audits and feedback was positively influenced by being in line with current conceptual frameworks, local policies, and values. Key words: Stroke rehabilitation, evidence-based practice, occupational therapy.