Efficacy of elaborated semantic features analysis in Aphasia: a quasi-randomised controlled trial

ABSTRACT

Background: Word finding difficulty is one of the most common features of aphasia. Semantic Features Analysis (SFA) directly aims to improve word finding in people with aphasia. Evidence from systematic reviews suggests that SFA leads to positive outcomes, yet the evidence comprises single case studies and case series. There is a need to evaluate the efficacy of SFA in controlled group studies/trials.

Aims: To evaluate the efficacy of Elaborated Semantic Feature Analysis (ESFA) for word finding in people with aphasia. We investigated: (a) the efficacy of ESFA versus a delayed therapy/control, (b) the efficacy of two therapy approaches – individual versus a combination of individual and group therapy.

Methods and procedures: We ran a multi-centre, quasi-randomised controlled trial, nested in a larger study (Thales-Aphasia). Participants were recruited from community settings. They had to be people with aphasia due to stroke at least four months post-onset. Participants were randomized to individual vs combination vs delayed therapy/control groups. Both therapy groups had 3 h of ESFA per week for 12 weeks. Delayed therapy/control group had no intervention for 12 weeks and were then randomized to either individual or combination therapy. The primary outcome was confrontation naming. Secondary outcomes were the Boston Naming Test, Discourse, the Functional Assessment of Communication Skills for adults (ASHA–FACS), the Stroke and Aphasia Quality of Life scale (SAQOL-39g), the General Health Questionnaire-12 item, and the EQ-5D.

Outcomes and Results: Of the 72 participants of the Thales-Aphasia project, 58 met eligibility criteria for speech-language therapy and 39 were allocated to ESFA. The critical p-value was adjusted for multiple comparisons (.005). For the therapy versus control comparison, there was a significant main effect of time on the primary outcome (p < .001, η²_p = .42) and a significant interaction effect (p = .003, η²_p = .21). An interaction effect for the SAQOL-39g (p = .015, η²_p = .11) and its psychosocial domain (p = .013, η²_p = .12) did not remain significant after Bonferroni adjustment. For the individual versus combination ESFA comparison, there were significant main effects of time on the primary outcome (p < .001, η²_p = .49), the BNT (p < .001, η²_p = .29) and the ASHA-FACS (p = .001, η²_p = .18). Interaction and group effects were not significant.

Conclusion: Though underpowered, this study provides evidence on the efficacy of ESFA to improve word finding in aphasia, with gains similar in the two therapy approaches.