Background and aims Choice of control comparator in RCTs of rehabilitation may increase the risk of bias. Pharmacological trials employ inert placebo-control comparisons whereas attention-control interventions aim to control for the benefits derived from the notice entailed in rehabilitation RCTs participation. We reviewed the delivery and content of attention-control comparisons in RCTs of speech and language therapy (SLT) for people with aphasia after stroke. Data Sources: Data from a Cochrane review of speech and language therapy for aphasia after stroke following a search of electronic databases (including CENTRAL, MEDLINE, EMBASE, CINAHL, AMED and SpeechBITE) (inception to September 2015). Methods: We conducted direct and indirect comparisons between SLT, attention-controls and no-therapy controls. We extracted intervention details using the template for intervention description and replication checklist. Language ability, drop-outs and non-adherence was double data extracted. Standardized mean differences and risk ratios (95% confidence intervals) were calculated. Results: Seven RCTs (n=447) compared SLT with attention-control interventions. Targeted rehabilitation of specific language structures was only described in SLT interventions. Control and SLT interventions however shared several active intervention components; social support, language stimulation, functional language practice and materials. Interventions’ format, location, frequency, intensity, duration and dose were matched. Distinctions between interventions were diminished. Conclusion: Social support interventions are unsuitable attention-control placebos in RCTs of aphasia rehabilitation. SLT utilises social interactions for therapeutic purposes to support the delivery of targeted language rehabilitation. The distinction between the interventions’ content and delivery was eroded to such an extent that the trials may have lacked sensitivity to detect differences.
No datasets are available for this submission.
No keywords are available for this submission
No license information is available for this submission.