Confidence Limits of Word Identification Scores Derived Using Nonlinear Quantile Regression

Narne, Vijaya Kumar; Möller, Sören; Wolff, Anne; Houmøller, Sabina Storbjerg; Loquet, Gérard Sylvian Jean Marie; Hammershøi, Dorte; Schmidt, Jesper Hvass

doi:10.1177/2331216520983110

Cited by 2 publications

(1 citation statement)

References 31 publications

(75 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, in a foundational work, Dubno et al (1995) used regression modeling to determine when an observed WR score was not at PB max given the degree of sensorineural hearing loss. Additional examples have been reported by Yellin et al (1989) and Narne et al (2021). The Speech Intelligibility Index (SII) (American National Standards Institute & Acoustical Society of America 2017) is a more rigorous computational method.…”

Section: Introductionmentioning

confidence: 99%

Clinical Utility of the Standardized Word Recognition Score

et al. 2023

View full text Add to dashboard Cite

Objectives: An unexpectedly low word recognition (WR) score may be taken as evidence of increased risk for retrocochlear tumor. We sought to develop evidence for or against using a standardized WR (sWR) score in detecting retrocochlear tumors. The sWR is a z score expressing the difference between an observed WR score and a Speech Intelligibility Index–based predicted WR score. We retrospectively compared the sensitivity and specificity of pure-tone asymmetry-based logistic regression models that incorporated either the sWR or the raw WR scores in detecting tumor cases. Two pure-tone asymmetry calculations were used: the 4-frequency pure-tone asymmetry (AAO) calculation of the American Academy of Otolaryngology–Head and Neck Surgery and a 6-frequency pure-tone asymmetry (6-FPTA) calculation previously optimized to detect retrocochlear tumors. We hypothesized that a regression model incorporating the 6-FPTA calculation and the sWR would more accurately detect retrocochlear tumors. Design: Retrospective data from all patients seen in the audiology clinic at Mayo Clinic in Florida in 2016 were reviewed. Cases with retrocochlear tumors were compared with a reference group with noise- or age-related hearing loss or idiopathic sensorineural hearing loss. Two pure-tone–based logistic regression models were created (6-FPTA and AAO). Into these base models, WR variables (WR, sWR, WR asymmetry [WRΔ], and sWR asymmetry [sWRΔ]) were added. Tumor detection performance for each regression model was compared twice: first, using all qualifying cases (61 tumor cases; 2332 reference group cases), and second, using a data set filtered to exclude hearing asymmetries greater than would be expected from noise-related or age-related hearing loss (25 tumor cases; 2208 reference group cases). The area under the curve and the DeLong test for significant receiver operating curve differences were used as outcome measures. Results: The 6-FPTA model significantly outperformed the AAO model—with or without the addition of WR or WRΔ variables. Including sWR into the AAO base regression model significantly improved disease detection performance. Including sWR into the 6-FPTA model significantly improved disease detection performance when large hearing asymmetries were excluded. In the data set that included large pure-tone asymmetries, area under the curve values for the 6-FPTA + sWR and AAO + sWR models were not significantly better than the base 6-FPTA model. Conclusions: The results favor the superiority of the sWR computational method in identifying reduced WR scores in retrocochlear cases. The utility would be greatest where undetected tumor cases are embedded in a population heavily representing age- or noise-related hearing loss. The results also demonstrate the superiority of the 6-FPTA model in identifying tumor cases. The 2 computational methods may be combined (ie, the 6-FPTA + sWR model) into an automated tool for detecting retrocochlear disease in audiology and community otolaryngology clinics. The 4-frequency AAO-based regression model was the weakest detection method considered. Including raw WR scores into the model did not improve performance, whereas including sWR into the model did improve tumor detection performance. This further supports the contribution of the sWR computational method for recognizing low WR scores in retrocochlear disease cases.

show abstract

Section: Introductionmentioning

confidence: 99%

Clinical Utility of the Standardized Word Recognition Score

et al. 2023

View full text Add to dashboard Cite

show abstract

Accuracy and Consistency of Confidence Limits for Monosyllable Identification Scores Derived Using Simulation, the Harrell–Davis Estimator, and Nonlinear Quantile Regression

Narne,

Mohan,

Avileri

et al. 2024

Diagnostics

View full text Add to dashboard Cite

Background: Audiological diagnosis and rehabilitation often involve the assessment of whether the maximum speech identification score (PBmax) is poorer than expected from the pure-tone average (PTA) threshold. This requires the estimation of the lower boundary of the PBmax values expected for a given PTA (one-tailed 95% confidence limit, CL). This study compares the accuracy and consistency of three methods for estimating the 95% CL. Method: The 95% CL values were estimated using a simulation method, the Harrell–Davis (HD) estimator, and non-linear quantile regression (nQR); the latter two are both distribution-free methods. The first two methods require the formation of sub-groups with different PTAs. Accuracy and consistency in the estimation of the 95% CL were assessed by applying each method to many random samples of 50% of the available data and using the fitted parameters to predict the data for the remaining 50%. Study sample: A total of 642 participants aged 17 to 84 years with sensorineural hearing loss were recruited from audiology clinics. Pure-tone audiograms were obtained and PBmax scores were measured using monosyllables at 40 dB above the speech recognition threshold or at the most comfortable level. Results: For the simulation method, 6.7 to 8.2% of the PBmax values fell below the 95% CL for both ears, exceeding the target value of 5%. For the HD and nQR methods, the PBmax values fell below the estimated 95% CL for approximately 5% of the ears, indicating good accuracy. Consistency, estimated from the standard deviation of the deviations from the target value of 5%, was similar for all the methods. Conclusions: The nQR method is recommended because it has good accuracy and consistency, and it does not require the formation of arbitrary PTA sub-groups.

show abstract

Confidence Limits of Word Identification Scores Derived Using Nonlinear Quantile Regression

Cited by 2 publications

References 31 publications

Clinical Utility of the Standardized Word Recognition Score

Clinical Utility of the Standardized Word Recognition Score

Accuracy and Consistency of Confidence Limits for Monosyllable Identification Scores Derived Using Simulation, the Harrell–Davis Estimator, and Nonlinear Quantile Regression

Contact Info

Product

Resources

About