Analysing and enhancing the MD Anderson Dysphagia Inventory: a mixed methods analysis of the head and neck cancer swallowing-related quality of life patient reported outcome tool

Abstract

Background Dysphagia, or difficulty swallowing, is a highly prevalent symptom of Head and Neck Cancer (HNC) and treatment, and has a marked impact on patients’ quality of life (QoL). Measurement of dysphagia-related QoL is therefore a key part of HNC clinical and research practice. Currently the only HNC-specific tool that caters to this need is the MD Anderson Dysphagia Inventory (MDADI). However, the MDADI is now more than 20 years old and has never undergone validation in the United Kingdom. Aim The aim of this study was to explore the psychometric properties and clinical utility of the MDADI. Methods This study followed a pragmatic, mixed methods approach to evaluate the MDADI, using the COSMIN methodology as a framework. Qualitative data from UK Speech & Language Therapists were gathered via an online survey, focussing on content validity and clinical utility of the MDADI, and analysed using a reflexive Thematic Analysis approach. Quantitative MDADI data from patients with HNC treated in NHS Lothian were used to analyse the MDADI’s structural validity and internal consistency, using Item Response Theory (IRT). Data generated by both arms of the study were then combined to explore the presence of Differential Item Functioning (DIF) in specific tool items, and to formulate suggestions for shortened versions of the tool. Results Analysis of the survey data uncovered issues with both the content validity and clinical utility of the MDADI. IRT analysis of tool structural validity and internal consistency showed these properties to be acceptable, however DIF analysis for the variables of age, sex and socioeconomic status indicated these variables all affected patient responses to specific MDADI items. Pre- and post-treatment versions of a shortened 5-item MDADI were generated, in addition to suggestions for future development of the MDADI tool.