Investigating the Minimal Clinically Important Difference for Bimodal Benefit in Speech Recognition Scores

To determine the minimal clinically important difference (MCID) values of bimodal benefit (ie, the additional benefit from use of hearing aid in the non-cochlear implant ear) in speech recognition scores in adult cochlear implant (CI) patients. Retrospective chart review. Tertiary referral CI center. Five hundred ninety-nine adult bimodal (CI with contralateral hearing aid) users, (20 to 93 y old; 336 males). Addition of contralateral hearing aid to CI listening. Bimodal benefit speech recognition scores for Consonant-Nucleus-Consonant (CNC) words and Arizona Biomedical (AzBio) sentences in quiet and noise were used to calculate MCID values. Bimodal benefit scores were calculated by subtracting CI-only performance from bimodal performance. Subjective measures of hearing abilities were assessed using the Speech, Spatial, and Qualities (SSQ-12) index. MCID values were calculated using distribution-based (half-standard deviation, standard error of the mean, Cohen d, and minimum detectable change) and anchor-based methods using the SSQ-12. After averaging across all 4 distribution-based methods, the MCID values for CNC, AzBio in quiet, and AzBio in noise were 7.8%, 6.4%, and 7.5%, respectively. The anchor-based method utilizing SSQ-12 scores yielded greater values than the null area under the curve value (>0.5), demonstrating a poor fit using anchor-based methods. The bimodal benefit MCID values from this study will aid in clinician and patient decision making (eg, pursuing a second CI and/or continued hearing aid use) and future bimodal benefit intervention studies. Results also suggest a need for better patient-reported measures to assess bimodal benefit.