Computerised speech and language therapy or attention control added to usual care for people with long-term post-stroke aphasia: the Big CACTUS three-arm RCT

People with aphasia may improve their communication with speech and language therapy many months/years after stroke. However, NHS speech and language therapy reduces in availability over time post stroke. This trial evaluated the clinical effectiveness and cost-effectiveness of self-managed computerised speech and language therapy to provide additional therapy. A pragmatic, superiority, single-blind, parallel-group, individually randomised (stratified block randomisation, stratified by word-finding severity and site) adjunct trial. Twenty-one UK NHS speech and language therapy departments. People with post-stroke aphasia (diagnosed by a speech and language therapist) with long-standing (> 4 months) word-finding difficulties. The groups were (1) usual care; (2) daily self-managed computerised word-finding therapy tailored by speech and language therapists and supported by volunteers/speech and language therapy assistants for 6 months plus usual care (computerised speech and language therapy); and (3) activity/attention control (completion of puzzles and receipt of telephone calls from a researcher for 6 months) plus usual care. Co-primary outcomes - change in ability to find treated words of personal relevance in a bespoke naming test (impairment) and change in functional communication in conversation rated on the activity scale of the Therapy Outcome Measures (activity) 6 months after randomisation. A key secondary outcome was participant-rated perception of communication and quality of life using the Communication Outcomes After Stroke questionnaire at 6 months. Outcomes were assessed by speech and language therapists using standardised procedures. Cost-effectiveness was estimated using treatment costs and an accessible EuroQol-5 Dimensions, five-level version, measuring quality-adjusted life-years. A total of 818 patients were assessed for eligibility and 278 participants were randomised between October 2014 and August 2016. A total of 240 participants (86 usual care, 83 computerised speech and language therapy, 71 attention control) contributed to modified intention-to-treat analysis at 6 months. The mean improvements in word-finding were 1.1% (standard deviation 11.2%) for usual care, 16.4% (standard deviation 15.3%) for computerised speech and language therapy and 2.4% (standard deviation 8.8%) for attention control. Computerised speech and language therapy improved word-finding 16.2% more than usual care did (95% confidence interval 12.7% to 19.6%;