Speech Intelligibility Index (SII) as a Potential Referral Metric for Adult Cochlear Implant Candidacy Evaluation
- 1Vanderbilt University Medical Center
Abstract
ObjectivesThe purpose of this pilot study was to investigate (1) if the Speech Intelligibility Index (SII) could be used to identify ears that meet cochlear implant (CI) candidacy beyond the "60/60" referral guidelines and (2) investigate the relationship between SII from patients own hearing aids (HAs) and SII from optimally fit clinic-stock HAs such that a patients own HAs can act as a proxy for an optimally fit HA and thus as a referral metric.
DesignProspective study of 23 participants ([≥]18 years) with bilateral moderate to profound sensorineural hearing loss who were CI candidates in at least one ear. SII-Clinic (SII-C) values were recorded with probe-microphone measures using a Verifit II device (NAL-N2 targets) with speech stimuli presented at 60 dB A during patient CI evaluation in either their own HAs or clinic-stock HAs. SII values from participants own HAs were recorded as part of a larger research protocol, thus SII-Personal (SII-P), with probe-microphone measures using a Verifit I device (NAL-RP targets) with pink noise stimuli presented at 65 dB A. Participants that did not wear HAs did not have SII-P calculated. Data was collapsed across ears. Correlation analysis between SII and best-aided consonant-nucleus-consonant (CNC) was conducted. Benchmark SII values were extrapolated using a best-fit line and CI candidacy rates reflecting various aided CNC scores that could be applied across CI centers. Nested logistic regression models were used to determine if the SII-C could improve model fit beyond the revised ear-specific "60/60" referral guideline. The relationship between SII-C and SII-P was assessed with Wilcoxon signed-rank test and Spearman correlation.
ResultsCorrelation between SII-C and aided CNC score was r = 0.82 (p < 0.001), and the correlation between SII-P and aided CNC was r = 0.46 (p = 0.014). Extrapolated benchmarks of SII were calculated based on different CNC candidacy cutoffs, and benchmarks were able to capture at least 83% of the ears that met CNC candidacy cutoffs. SII was also found to improve model fit with word recognition score (WRS) and pure tone average (PTA) for 60% CNC candidacy cutoff ({chi} 2(1) = 9.7; p = 0.002) and increased discrimination (AUC = 0.89 to AUC = 0.99). There was no significant difference between SII-C and SII-P using the Wilcoxon signed-rank test (p = 0.559).
ConclusionsA larger study is needed that includes more ears that do not qualify for a CI; however, the results of this study demonstrate that (1) SII has the potential to be used in conjunction with the "60/60" referral guideline for CI candidacy evaluations and (2) SII can be used as a marker of audibility across different devices, including patients own HAs, making SII an accessible metric during HA fittings and fine tunings.
Related
- Human-calibrated automated intelligibility evaluation of speech neuroprostheses
- Exploring asymmetry of electrical compound action potential latencies in sequential cochlear implants as a possible effect of the lateral olivocochlear bundle.
- Evaluation of Fine Structure Coding in New Cochlear Implant Users with Tonotopic Fitting: a Prospective, Randomized, Double-Blind, Cross-Over Study
- Analytical validation of digital measure of intelligibility in progressive supranuclear palsy
- Screening and characterization of endophytic bacterial isolates associated with Echinops kebericho M. and Lepidium sativum L. and evaluation of their potential role
BibTeX
@misc{kong2026speech,
title = {{Speech Intelligibility Index (SII) as a Potential Referral Metric for Adult Cochlear Implant Candidacy Evaluation}},
author = {J. A. Kong and T. N. Tamati and A. C. Moberly and J. D. Neukam},
howpublished = {medRxiv},
year = {2026},
month = sep,
doi = {10.64898/2026.09.10.26362790},
url = {https://www.medrxiv.org/content/10.64898/2026.09.10.26362790v1},
note = {CC BY-ND},
}
RIS
TY - UNPB
TI - Speech Intelligibility Index (SII) as a Potential Referral Metric for Adult Cochlear Implant Candidacy Evaluation
AU - J. A. Kong
AU - T. N. Tamati
AU - A. C. Moberly
AU - J. D. Neukam
PY - 2026
DA - 2026/09/10
PB - medRxiv
DO - 10.64898/2026.09.10.26362790
UR - https://www.medrxiv.org/content/10.64898/2026.09.10.26362790v1
AB - ObjectivesThe purpose of this pilot study was to investigate (1) if the Speech Intelligibility Index (SII) could be used to identify ears that meet cochlear implant (CI) candidacy beyond the "60/60" referral guidelines and (2) investigate the relationship between SII from patients own hearing aids (HAs) and SII from optimally fit clinic-stock HAs such that a patients own HAs can act as a proxy for an optimally fit HA and thus as a referral metric. DesignProspective study of 23 participants ([≥]18 years) with bilateral moderate to profound sensorineural hearing loss who were CI candidates in at least one ear. SII-Clinic (SII-C) values were recorded with probe-microphone measures using a Verifit II device (NAL-N2 targets) with speech stimuli presented at 60 dB A during patient CI evaluation in either their own HAs or clinic-stock HAs. SII values from participants own HAs were recorded as part of a larger research protocol, thus SII-Personal (SII-P), with probe-microphone measures using a Verifit I device (NAL-RP targets) with pink noise stimuli presented at 65 dB A. Participants that did not wear HAs did not have SII-P calculated. Data was collapsed across ears. Correlation analysis between SII and best-aided consonant-nucleus-consonant (CNC) was conducted. Benchmark SII values were extrapolated using a best-fit line and CI candidacy rates reflecting various aided CNC scores that could be applied across CI centers. Nested logistic regression models were used to determine if the SII-C could improve model fit beyond the revised ear-specific "60/60" referral guideline. The relationship between SII-C and SII-P was assessed with Wilcoxon signed-rank test and Spearman correlation. ResultsCorrelation between SII-C and aided CNC score was r = 0.82 (p < 0.001), and the correlation between SII-P and aided CNC was r = 0.46 (p = 0.014). Extrapolated benchmarks of SII were calculated based on different CNC candidacy cutoffs, and benchmarks were able to capture at least 83% of the ears that met CNC candidacy cutoffs. SII was also found to improve model fit with word recognition score (WRS) and pure tone average (PTA) for 60% CNC candidacy cutoff ({chi} 2(1) = 9.7; p = 0.002) and increased discrimination (AUC = 0.89 to AUC = 0.99). There was no significant difference between SII-C and SII-P using the Wilcoxon signed-rank test (p = 0.559). ConclusionsA larger study is needed that includes more ears that do not qualify for a CI; however, the results of this study demonstrate that (1) SII has the potential to be used in conjunction with the "60/60" referral guideline for CI candidacy evaluations and (2) SII can be used as a marker of audibility across different devices, including patients own HAs, making SII an accessible metric during HA fittings and fine tunings.
ER -