Speech and Hearing Sciences ETDs

Publication Date

Summer 7-15-2017

Abstract

Purpose: Calls for treatment and assessment fidelity strongly suggest the need to reduce treatment provider and assessor variance surrounding intervention research. The extent to which these sources of variance influence treatment outcomes in aphasia treatment research has yet to be examined. This simulation study sought to explore the relationships between quality of fidelity methods, sample size, power to detect treatment effects, and aphasia treatment effect sizes.

Methods: Individual participant outcomes collected from previous aphasia treatment research studies were used to simulate 200,000 participant outcomes, from which 8,000 sample treatment trials were simulated. Effect sizes were calculated for treatment outcomes related to four total assessment and treatment fidelity methods - treatment provider training, treatment provider monitoring, assessor blinding, and assessor training. Results from calculations were applied to 80,000 simulated participant trials of varying sample sizes, fidelity levels, and outcome assessments to determine effect size and power to detect effects.

Results: Simulated results found: positive effect sizes and increased power to detect effects for high fidelity treatment provider training and monitoring, with reduced effect sizes and ability to detect effects from high fidelity assessor blinding, and no effects for assessor training. Increased power was observed as sample size increased. Multidimensional assessment outcomes resulted in higher treatment effect sizes and power to detect effects than unidimensional outcomes.

Conclusions: Simulations generally support findings from previous research. With the exception of treatment provider training, few studies reported calculable outcomes related to fidelity, validating the need for this simulation and future research. High fidelity treatment provider training and monitoring are simple methods to increase ability to detect treatment effects and effect size overall, and blinding assessors helps to reduce biased reporting. Recommendations for researchers with limited resources are provided to reduce variance from assessors and treatment providers and increase confidence in results.

Degree Name

Speech-Language Pathology

Level of Degree

Masters

Department Name

Speech and Hearing Sciences

First Committee Member (Chair)

Jessica Richardson

Second Committee Member

Richard Arenas

Third Committee Member

Janet Patterson

Language

English

Keywords

aphasia, assessment, treatment, fidelity, integrity, simulation

Document Type

Thesis

Share

COinS