Objective: Determine association of advancements in speech processor technology with improvements in speech recognition outcomes.
Study design: Retrospective cohort.
Setting: Tertiary referral center.
Patients: Adult unilateral cochlear implant (CI) recipients.
Intervention: Increasing novelty of speech processor defined by year of market availability.
Main outcome measures: Consonant-Nucleus-Consonant (CNC) and Hearing in Noise Test (HINT) in quiet.
Results: From 1991 to 2016, 1,111 CNC scores and 1,121 HINT scores were collected from 351 patients who had complete data. Mean post-implantation CNC score was 53.8% and increased with more recent era of implantation (p < 0.001, analysis of variance [ANOVA]). Median HINT score was 87.0% and did not significantly vary with implantation era (p = 0.06, ANOVA). Multivariable generalized linear models were fitted to estimate the effect of speech processor novelty on CNC and HINT scores, each accounting for clustering of scores within patients and characteristics known to influence speech recognition outcomes. Each 5-year increment in speech processor novelty was independently associated with an increase in CNC score by 2.85% (95% confidence limits [CL] 0.26, 5.44%) and was not associated with change in HINT scores (p = 0.30).
Conclusion: Newer speech processors are associated with improved CNC scores independent of the year of device implantation and expanding candidacy criteria. The lack of association with HINT scores can be attributed to a ceiling effect, suggesting that HINT in quiet may not be an informative test of speech recognition in the modern CI recipient. The implications of these findings with respect to appropriate interval of speech processor upgrades are discussed.