Purpose:
The purpose of the study proposed in this ethics application is to examine the impact of an adaptive speech-language treatment app, created using principles of user-centred design, on language and communication outcomes of people with aphasia (PwA) within a randomized controlled trial.
Background:
Nearly one third of individuals who have a stroke will present with aphasia, a communication impairment that affects the ability to speak, understand, read and write. People with aphasia (PwA) have been shown to demonstrate more limitations in their participation in daily activities, lower quality of life and a higher incidence of depression relative to post-stroke individuals without aphasia, even when other factors such as physical ability, social support and well-being were similar.
The overall goals of the VoiceAdapt research project are to: 1) conduct a needs survey with PwA and speech-language pathologists (SLPs); 2) develop an app employing a user-centred design process; and 3) conduct a randomized controlled trial. Using user-centred design, end users and stakeholders (PwA, therapists and formal/informal caregivers) will be engage din a needs survey during the technology development process. Outcomes will be analyzed to elaborate implications for decision makers as to how to implement such adaptive training environments in stroke rehabilitation. This aspect of the research (i.e., Goals 1 \& 2) will be carried out by consortium members in Austria and Germany. Only the randomized controlled trial (Goal 3), which is the subject of the present application, will be conducted at the University of Alberta and the University of Toronto.
PwA will be provided with a personalized speech-language treatment app (i.e., on a tablet), which automatically adapts the exercises to the users' actual skill level. Individual learning progress of PwA will be supported by personalized and gamified training elements to achieve higher compliance and adherence to the exercises. As a primary outcome, the adaptive assistive technology will examine language performance related to the specific intervention of the personalized speech-language treatment app. Additional objectives will examine the change in the overall language and communication functioning of PwA; the well-being and quality of life of PwA and their caregiver's perception of communication effectiveness after intervention with the treatment app.
The research will be carried out by the VoiceAdapt Consortium, which brings together the resources of five participating international institutions. Research partners have expertise in the fields of technology/e-health and graphical design skills (Nurogames, NG), personalization and adaptvitiy (Technical University of Berlin, TUB), user-centred design and gamification/persuasion (Austrian Institute of Technology, AIT), and clinical research and intervention studies (University of Toronto, UoT; University of Alberta, UoA).
Research Methods/Procedures:
Researchers at the University of Alberta and University of Toronto will be the primary investigators of the randomized controlled trial (RCT) portion of the VoiceAdapt Project. The RCT will involve 80 PwA in total across University of Alberta and University of Toronto sites.
Design of the RCT: The RCT will be a two arm, randomized, waitlist-controlled, crossover group design. In Phase 1, PWA randomized to the intervention group will train with VoiceAdapt. The control group will be a waitlist treatment-deferred group. In Phase 2, the control group will train with VoiceAdapt. The randomization of participants will be achieved by computer-generated assignment.
Participants: All participants will have suffered a left hemisphere stroke, be at least 6 months post-onset, and will have aphasia as defined by performance on the Western Aphasia Battery - Rev. (Aphasia Quotient greater than or equal to 30, with a prominent verbal expression impairment), but type of aphasia will not be an inclusion criterion. Participants will speak English as their primary language, pass screenings of visual perception, hearing and basic cognitive functioning and be willing to commit to participation for the study duration.
Intervention: Each intervention phase will last for 5 weeks. During this period, participants in the experimental condition will be asked to do speech language tasks 5 days a week for 60 minutes via the VoiceAdapt app (intervention group). The control group will not be engaged in any one on one speech therapy during the first intervention phase (Phase 1). During the second intervention phase (Phase 2), the control group will participate in training with the VoiceAdapt app (60 minutes/day x 5 days/week) and the intervention group from Phase 1 will crossover into receiving no treatment. Research staff will check in with individuals in the intervention group on a weekly basis to monitor and adjust treatment exercises. Research staff will check in with participants in the control group twice during their respective control phase.
Outcome Measures: In line with the recommended core outcome measures for aphasia research (Wallace et al., 2019), participants in both groups will be tested before and after each condition on the following measures: the primary outcome measure will be the Boston Naming Test (BNT) to measure naming; secondary outcome measures will be: Western Aphasia Battery-Revised (WAB-R); Stroke and Aphasia Quality of Life Scale (SAQOL-39) to measure quality of life; and Communication Effectiveness Index (CETI) to measure the person with aphasia and the partners' perception of the participant's communication. In addition to these measures, the Situational Motivation Scale (SIMS) and an end of study questionnaire will be provided to all participants.
Plan for Data Analysis:
The primary analysis will be a one-sided hypothesis test of the null hypothesis of the change in naming on the Boston Naming Test (BNT) from baseline to week 6, tested at the 0.1 level of significance. The observed change on the BNT as well as the baseline, 6 week and follow up values will be summarized for each group as means and standard deviations. Secondary outcomes will be summarized using means (SD), medians (IQR) or proportions (CI), and analyzed using t tests, as appropriate. All analyses will be presented in aggregate.
