RTSS-Voice: Towards a unified system to classify treatments for muscle tension dysphonia

Project Summary Systematically improving upon current voice therapy outcomes is problematic as the specific clinician actions (i.e., ingredients) responsible for improved patient functioning (i.e., targets) are unknown. For example, Comparative Effectiveness Research (CER) can show that therapy A works better than therapy B on global outcome X. But why did therapy A provide better outcomes? Why did some patients in therapy B significantly improve with the “worse” program; and some patients in therapy A remain unchanged with the “better” program? A theory-based system is needed to scientifically identify a program’s ingredients associated with improved outcomes across patients. And standard labels are needed to make the identified active ingredients generalizable across therapy programs. Therefore, this project will use a theory-driven framework for describing the ingredients/targets of rehabilitation treatments—called the Rehabilitation Treatment Specification System (RTSS)—and standard voice-specific terminology/definitions—called the RTSS-Voice— to standardly describe (Aim 1) and compare (Aim 2) variations in treatment across 9 well-known and diverse voice therapies. Also, we will create/test an implementation toolkit to facilitate RTSS-Voice adoption in clinical care across 5 Voice Centers (Aim 3). It is hypothesized that the RTSS and RTSS-Voice will characterize all therapies without needing revisions (Aim 1) and identify ingredients/targets that are unique to one therapy and/or common across multiple therapies (Aim2). And since the RTSS-Voice will help clinicians think about their treatment more specifically and in relation to 9 evidence-based therapies, we hypothesize adoption will be associated with improved outcomes at all 5 Voice Centers. The resulting list of mutually exclusive ingredient/targets across therapy programs will obviously improve the state of CER by enabling the identification/comparison of active ingredients across therapy programs; instead of the current practice of studying/comparing entire programs. Clinical adoption will result in large datasets with standard ingredients linked to outcomes, which will facilitate innovative hypotheses and interpretable datamining/machine learning to realistically improve voice therapy effectiveness. This work is likely to generalize to other Centers due to the involvement of >20 influential clinicians and the implementation toolkit.