Background: Discourse assessment and treatment in aphasia rehabilitation is a priority focus for a range of stakeholder groups. However, a significant majority of speech and language therapists (SLTs) infrequently conduct discourse analysis, and do not feel competent in doing so. Known barriers identified in other countries, specifically a lack of time, training, expertise and resources, affect use of discourse analysis in clinical practice. Aims: To investigate UK SLTs' reported practices and views of discourse analysis, barriers and facilitators, and clinical feasibility in aphasia rehabilitation. Methods & Procedures: An online survey of 52 questions adapted from existing research and incorporating behaviour change literature was created for the study and piloted. UK SLTs working in aphasia rehabilitation for at least 6 months were invited to participate. Potential participants were contacted through national and local clinical excellence networks, a National Health Service (NHS) bespoke e-mail list, and national magazine advertisement, and the study was also advertised on social media (Twitter). Therapists read an online participant information sheet and submitted individual electronic consent online; then progressed to the Qualtrics survey. Descriptive, correlational and inferential statistical analyses were conducted, and content analysis was carried out on the questions requiring text. Outcomes & Results: A total of 211 valid responses were received from primarily female SLTs, aged 20-40 years, working full-time in the NHS in England, in community, inpatient and acute/subacute multidisciplinary settings. A total of 30% SLTs collected discourse analysis often, were mostly very experienced, and working part-time in community settings. Years of experience was predictive of use. Discourse was most often collected using standardized picture descriptions and recounts during initial assessment. Samples were infrequently recorded, and typically transcribed in real-time. Most SLTs (53-95%) reported making clinical judgements or manually counted words, sentences, communication of ideas and errors, and were confident in doing so. Barriers included time constraints; lack of expertise, confidence, training, resources and equipment; and patient severity. Discourse 'super-users' were distinguished by significantly higher professional motivation for discourse and workplace opportunity than other SLTs, and 'non-users' were distinguished by significantly less knowledge and skills in discourse analysis than other SLTs. SLTs reported a desire and need for training, new/assistive tools and time to do more discourse analysis in practice. Conclusions & Implications: Clinicians were highly engaged and relatively active in at least some aspects of discourse analysis practice. Interventions that target individual clinicians as well as organizations and systems are needed to improve the uptake of discourse analysis in practice.