Purpose: Patient-generated narratives provide valuable insight into the lived experience of hearing aid outcomes. However, conventional qualitative analysis is resource-intensive, and existing Natural Language Processing (NLP) tools lack domain-specific constructs for audiology. This study aimed to develop and formatively evaluate a custom hearing aid dictionary within the Linguistic Inquiry and Word Count (LIWC) framework to enable automated analysis of user-generated narratives. Methods: A multi-phase development process was used. An eight-member committee of hearing healthcare professionals and hearing aid users generated an initial corpus of 1,597 entries across 51 clinically relevant categories identified through a prior e-Delphi consensus process. Psychometric properties were evaluated using Cronbach’s alpha and Kuder-Richardson Formula 20 (KR-20) on textual survey responses and transcribed interview data from prior studies, followed by iterative refinement, removing entries that failed to meet pre-specified internal consistency thresholds, resulting in 1,013 retained entries. The dictionary was applied to open-text responses from a cross-sectional survey of hearing aid users (n = 1,077). Associations between linguistic categories and self-reported outcomes (e.g., overall benefit, overall satisfaction, and daily helpfulness of hearing aids) were examined using Spearman correlations. 1 IHCON 2026 Results: Iterative refinement improved internal consistency, with post-refinement Cronbach’s α ranging from −.01 to .35 for survey data and −.02 to .53 for interview data, consistent with established distributional properties of sparse linguistic data and comparable to the LIWC-22 framework. The custom dictionary captured domain-specific aspects of hearing aid use (e.g., speech understanding, device management, social situations, perceived benefit), whereas general LIWC categories reflected broader cognitive and affective processes. Correlations between custom linguistic categories and self-reported outcomes were statistically significant, with modest effect sizes typical of linguistic analysis studies (rₛ = .10–0.19). For example, positive self-perceived satisfaction and quality-of-life language were associated with greater overall benefit, satisfaction and daily helpfulness, whereas negative benefit-related and quality-of-life language were associated with poorer self-reported outcomes. Conclusions: The custom hearing aid LIWC dictionary provides evidence of construct validity and enhances the interpretability of patient narratives. Domain-specific linguistic tools may provide a scalable complement to traditional outcome measures; however, further validation across diverse datasets and contexts is warranted.