Establishing Linguistic Categories for Hearing Aid Experiences in Adults: A Modified Delphi Study
None:
Background: Recent advancements in automated natural language processing (NLP) methods and tools have enhanced the efficacy and accuracy of quantitative analysis of natural language data. NLP offers significant potential for audiology by providing valuable insights from open-text responses about users' lived hearing aid experiences.Purpose: This study aimed to establish linguistic categories pertinent to the experiences of adults using hearing aids, with the ultimate goal of developing a specialized text processing module to facilitate natural language analysis of textual data on hearing aid experiences.Research Design: A modified electronic Delphi (e-Delphi) design was employed.Study Sample: A panel of 16 audiology experts from seven countries.Data Collection and Analysis: Two survey rounds were conducted. In round 1, experts rated categories from the Linguistic Inquiry and Word Count (LIWC) software and categories from principal component analysis of open-ended text data on hearing aid experiences. Experts also responded to open-ended prompts regarding categories derived from qualitative studies on hearing aid experiences. Responses were condensed and refined into items for rating during round 2. In round 2, experts reconsidered their ratings in light of the group answers of round 1. All ratings used a four-point Likert scale of importance. Measures of central tendency, levels of dispersion, and Cronbach's α reliability coefficients were conducted in both rounds for comparative purposes.Results: The open-ended section generated 26 items. Consensus was met on 53 linguistic categories encompassing social (e.g., social situations, social support), emotional (e.g., positive tone, negative tone), cognitive (e.g., cognitive processes, causation), lifestyle (e.g., work, leisure), hearing aid-related (e.g., sound quality, use and handling), and service delivery-related dimensions.Conclusions: The linguistic categories identified provide a foundation for developing a customized LIWC text processing module tailored to the analysis of hearing aid experiences. Future research is needed to refine and validate the custom text processing module.Clinical Relevance Statement: The findings provide a framework to enable the development of a customized hearing aid LIWC dictionary. Such a tool may enhance clinicians' ability to explore hearing aid outcomes and improve patient-centered care by integrating real-world experiences into clinical practice.
