Intelligibility of the Patient's Speech Predicts the Likelihood of Cochlear Implant Success in Prelingually Deaf Adults

Objectives: The objective of this study was to determine the validity and clinical applicability of intelligibility of the patient's own speech, measured via a Vowel Identification Test (VOW), as a predictor of speech perception for prelingually deafened adults after 1 year of cochlear implant use. Specifically, the objective was to investigate the probability that a prelingually deaf patient, given a VOW score above (or below) a chosen cutoff point, reaches a postimplant speech perception score above (or below) a critical value. High predictive values for VOW could support preimplant counseling and implant candidacy decisions in individual patients. Design: One hundred and fifty-two adult cochlear implant candidates with prelingual hearing impairment or deafness took part as speakers in a VOW; 149 speakers completed the test successfully. Recordings of the speech stimuli, consisting of nonsense words of the form [h]-V-[t], where V represents one of 15 vowels/diphthongs ([alpha, a, epsilon, e, I, i, omicron, o, Y, y, u, phi, epsilon i, oe y, alpha u]), were presented to two normal-hearing listeners. VOW score was expressed as the percentage of vowels identified correctly (averaged over the 2 listeners). Subsequently, the 149 participants enrolled in the cochlear implant selection procedure. Extremely poor speakers were excluded from implantation, as well as patients who did not meet regular selection criteria as developed for postlingually deafened patients. From the 149 participants, 92 were selected for implantation. For the implanted group, speech perception data were collected at 1-year postimplantation. Results: Speech perception score at 1-year postimplantation (available for 77 of the 92 implanted participants) correlated positively with preimplant intelligibility of the patient's speech, as represented by VOW (r = 0.79, p < 0.00001): the more intelligible the patient's speech, the higher the predicted postimplant speech perception score. This correlation is explained by the hypothesis that the two variables have a common driving force, i. e., (in) adequacy of auditory speech input in the earliest years of life. With a 60% cutoff point, VOW can discriminate between individuals with "above-chance" postimplant speech perception and those with "chance level" postimplant speech perception with sensitivity and specificity of 0.84 and 0.86, respectively. The probability that a patient with a VOW score = 60% achieves "above-chance" speech perception after implantation is 0.91. Conversely, the probability that a patient with VOW < 60% reaches "above-chance" speech perception is 0.25. Conclusions: For prelingually deaf adults, intelligibility of the patient's speech-as represented by VOW-is a valid predictor of postimplant speech perception. A patient with a VOW score above a preset cutoff is much more likely to develop acceptable speech perception after implantation than a patient with a VOW score below that cutoff. The binary classification based on VOW and the associated probabilities of cochlear implant success in terms of speech perception can be used-in addition to existing criteria-to support the clinician in guiding patient expectations and in considering implant candidacy for individual patients.