Rasch Analysis of the Listening Effort Questionnaire-Cochlear Implant

Listening effort may be defined as the attentional and cognitive resources needed to understand an auditory message, modulated by motivation. Despite the use of hearing devices such as hearing aids or cochlear implants (CIs), the requirement for high listening effort remains a challenge for individuals with hearing loss. The Listening Effort Questionnaire-Cochlear Implant (LEQ-CI) is a hearing-specific patient-reported outcome measure (PROM), which has been designed for use in the CI candidacy and rehabilitation process to assess perceived listening effort in everyday life in adults with severe-profound hearing loss. The LEQ-CI has been developed in line with international consensus-based standards for best practice in PROM construction. The aim of this study was to improve the measurement precision of the LEQ-CI and to assess its psychometric measurement properties.

A field test was undertaken with 330 CI patients from five National Health Service auditory implant centers in the United Kingdom. Participants were adults (≥18 years of age), had a severe-profound hearing loss, and met the UK candidacy criteria for cochlear implantation specified by the National Institute for Health and Care Excellence (NICE). Participants completed and returned an anonymized 29-item (each with a 5- or 7-point response option), draft version of the LEQ-CI (LEQ-CI29) and a demographic questionnaire. Rasch analysis was undertaken using Winsteps software and the partial credit model to assess rating scale function and item fit. Results informed refinements to produce a 21-item version (LEQ-CI21), which underwent a further Rasch analysis.

The sample was predominantly female: 60.3% (n = 191). Median age of participants was 66 (range 21 to 89) years, with 7.3% (n = 24) of respondents being CI candidates and 92.7% (n = 306) being CI recipients. Mean duration of implantation was 3.8 (SD = 4.8) years. Initial Rasch analysis of the LEQ-CI29 revealed poor rating scale functioning. Collapsing the 5- and 7-point rating scales to 3- and 4-point scales and removing eight items produced a 21-item PROM (LEQ-CI21). Rasch analysis of the LEQ-CI21 showed good fit to the Rasch measurement model. No items showed misfit and dimensionality analysis supported the existence of a single Rasch dimension, defined as perceived listening effort in daily life. Person reliability was 0.91 and the person separation index was 3.28, establishing four levels of person ability. The item separation index was 9.69, confirming the item hierarchy. No items showed differential item functioning for gender or age. The item difficulty range was -0.81 to 1.05, the person ability range for nonextreme persons was -3.54 to 2.49, and the mean person ability was -0.31.

Overall, the LEQ-CI21 was found to meet the Rasch model criteria for interval-level measurement. The LEQ-CI21 is the first PROM to be developed specifically for the measurement of perceived listening effort and one of the first patient-reported outcome measures for use with CI patients to be developed using Rasch analysis. The LEQ-CI21 has the potential to be used as a research tool and in clinical practice to evaluate perceived listening effort in daily life. Further psychometric evaluation of the LEQ-CI21 is planned.