Classification performance of Spanish and English word recognition testing to identify cochlear implant candidates.

Abstract

To explore the effects of language on word recognition ability and associated access to cochlear implants in Spanish- and English-speakers with hearing loss. Retrospective chart review. The study included 1411 English and Spanish-speaking adults with sensorineural hearing loss. Word recognition as a function of pure-tone average was evaluated by test language. Sensitivity and specificity of word recognition testing performed using the NU-6 and Spanish Bisyllables in discriminating between cochlear implant and non-cochlear implant candidates were analysed. Spanish-speakers had significantly poorer thresholds compared to English-speakers, yet significantly better word recognition, t(2763) = 1.86, p < 0.05. The effect of test language on word recognition increased with increasing hearing loss severity, F(3, 2762) = 6.03, p < 0.001, partial η2 = .007. The sensitivity and specificity of word recognition testing were 93% and 74% respectively for Spanish, and 95% and 75%, respectively, for English word recognition tests in meeting referral criteria for cochlear implant evaluation. Based on classification performance analysis, Spanish and English word recognition measures exhibit comparable effectiveness in referring patients for cochlear implant evaluation. While Spanish-language measures yielded better performance, this does not result in a significant difference in referral rates for Spanish versus English speakers for unaided word recognition scores of 60% or less.