Purpose: Exercise-based treatments may improve swallowing safety and efficiency; yet, it is not clearly understood which factors predict nonadherence to recommended treatment protocols. The aim of this study was to construct an algorithm for stratifying risk of nonadherence to a lingual strengthening dysphagia treatment program.
Method: Using recursive partitioning, we created a classification tree built from a pool of sociodemographic, clinical, and functional status indicators to identify risk groups for nonadherence to an intensive lingual strengthening treatment program. Nonadherence, or noncompletion, was defined as not completing two or more follow-up sessions or a final session within 84 days.
Results: The study cohort consisted of 243 Veterans enrolled in the Intensive Dysphagia Treatment program across six sites from January 2012 to August 2019. The overall rate of nonadherence in this cohort was 38%. The classification tree demonstrated good discriminate validity (C-statistic = 0.74) and contained eight groups from five variables: primary diagnosis, marital status, Penetration-Aspiration Scale (PAS) severity score, race/ethnicity, and age. Nonadherence risk was categorized as high (range: 69%-77%), intermediate (27%-33%), and low risk (≤ 13%-22%).
Conclusions: This study identified distinct risk groups for nonadherence to lingual strengthening dysphagia rehabilitation. Additional research is necessary to understand how these factors may drive nonadherence. With external validation and refinement through prospective studies, a clinically relevant risk stratification tool can be developed to identify patients who may be at high risk for nonadherence and provide targeted patient support to mitigate risk and provide for unmet needs.