ABSTRACT
Background
Evidence regarding the effect of conversationally based communication group treatment on discourse production in aphasia is limited. Given the rich, complex communication experiences provided in these groups, it seemed plausible that participation in them could result in improvement of simpler aspects of discourse production.
Aims
To examine the effects of group communication treatment on the informativeness and efficiency of structured and conversational discourse tasks in adults with chronic aphasia.
Methods & Procedures
The data for this study were discourses elicited prospectively from 23 adults with chronic aphasia by Elman & Bernstein-Ellis (1999a) as part of their randomised controlled trial, but never previously transcribed or analyzed. We evaluated changes in discourse informativeness and efficiency at treatment exit and follow-up with Bayesian generalised linear mixed-effects models. Individual effect sizes at exit and follow-up were estimated and a region-of-practical-equivalence approach was used to evaluate whether the posterior distributions at each timepoint for each participant were clinically meaningful.
Outcomes & Results
Results at the group level revealed that structured discourses became more informative and efficient after treatment, and that this improvement was at least maintained at follow-up. Informativeness of conversational discourse did not change from treatment entry to exit, but there was modest evidence of improvement at follow-up. There was no evidence of change to the efficiency of conversational discourse at either timepoint. There was wide variability in individual response to the treatment.
Conclusions
Participation in conversation-based communication group treatment was associated with more informative and efficient structured discourse production and modestly improved informativeness in conversational discourse. Examining individual responses to treatment provided additional insight about the group-level outcomes and provided some clues about factors that might have influenced performance for some of the participants.
Acknowledgments
We thank Ellen Bernstein-Ellis, M.A., CCC-SLP, and the people with aphasia who participated in the study. Thanks also to the following students who contributed to various phases of this project: Alexa Bergenfeld, Alexandra Buchanan, Abby Cohen, Anne DeCaro, Vanessa Gomez, Sophia Hendrix, Sara Manzon, and Carly Truglio.
Disclosure statement
The original RCT was partially funded by a grant from the National Easter Seals Research Program (awarded to Roberta J. Elman).
Notes
1. For one instance of missing data at follow-up in the structured discourse samples (P11; medical reasons), model predictions of the individual effect size estimate included the residual (observation-level) variance. This approach is equivalent to generating model predictions for “new” data and can be used to impute values when data are only missing for the outcome variable.