A Statistical Learning Model for Predicting Noise-Induced Hearing Loss in Humans

Details

• Personal Author:
  Davis R ; Hamernik RP ; Qiu W
• Description:
  The specific aims of the proposal were: (1) To obtain noise exposure and audiometric data on a large number of workers in heavy industry; (2) To develop and test a statistical learning model [Support Vector Machine (SVM)] for the prediction of hearing loss in humans exposed to high levels of complex noise. The effectiveness of the kurtosis metric in evaluating the risk of hearing loss from complex noise exposures was addressed. A cross-sectional approach was used in this study. The four main study procedural elements were to: 1) select workplaces based upon consistent noise and employment characteristics; 2) select and recruit subjects based upon strict quality criteria (i.e., health history and stable employment criteria); 3) obtain shift-long temporal waveforms of the noise that workers were exposed to for evaluation of noise exposures on selected subjects; and 4) obtain audiometric testing on all selected subjects. Using the data from (3) and (4), in industries that have high and low kurtosis noise environments, we evaluated several approaches to determine if the kurtosis metric is a reasonable candidate for use in modifying exposure level calculations to estimate the risk of NIHL from any type of noise exposure environment. An SVM model with a nonlinear radial basis function kernel was also built on a database consisting of individual subject shift long (8 hrs) industrial noise recordings and the associated hearing threshold levels. The results were: 1. A large worker database (N=1,150 subjects) was acquired. The database consists of individual subject shift long (8 hrs) industrial noise recordings and their associated hearing threshold levels. 2. Current hearing risk assessment criteria (ISO-1999) underestimated the mean and median noise-induced hearing loss (NIHL) by up to 10 dB for the higher kurtosis level group [â(t) > 4]. 3. The SVM model successfully predicted the NIHL resulting from complex industrial noise exposures. The inclusion of the kurtosis metric, which increased the predictive accuracy of the model by 19%, supports the need to incorporate this metric in the evaluation of a complex noise exposure for hearing conservation purposes. 4. A new approach to characterize the hazardous effects of complex noise was developed in which an energy based metric [cumulative noise exposure (CNE)] was combined with a temporal correction term (i.e., kurtosis) to evaluate human noise exposure data. The use of the kurtosis metric more accurately assessed the risk of developing high frequency NIHL in workers exposed to high level Gaussian (G) and non-Gaussian (non-G) noise. The results above lend support for the kurtosis metric as an important variable in determining the hazards to hearing posed by a high-level complex industrial noise for hearing conservation purposes. This finding suggests that an energy metric (Leq) alone is not sufficient to predict noise-induced permanent threshold shift (NIPTS). Energy is insensitive to the effects of the temporal characteristics of a noise exposure known to be important in affecting hearing. These results allow us to better understand the role of the kurtosis metric and SVM model in NIHL which may lead to its incorporation into a new generation of more predictive hearing risk assessment for noise exposure. [Description provided by NIOSH]
• Subjects:
  Database Hearing Loss, Noise-Induced Noise Noise, Occupational Occupational Exposure Occupational Health Safety
• Keywords:
  Audiological Testing Audiometry Databases Humans Models Noise Exposure Noise Induced Hearing Loss Occupational Exposure Risk Assessment Statistical Analysis
• Series:
  Grant Final Reports
• Publisher:
  National Institute for Occupational Safety and Health
• Document Type:
  Text
• Funding:
  Grant
• Genre:
  Final Grant Report
• Place as Subject:
  New York ; OSHA Region 2
• CIO:
  National Institute for Occupational Safety and Health (NIOSH)
• Division:
  OEP - Office of Extramural Programs
• Topic:
  Workplace Safety & Health
• Location:
  North America
• Pages in Document:
  1-17
• NIOSHTIC Number:
  nn:20059071
• NTIS Accession Number:
  PB2021-100153
• Citation:
  Atlanta, GA: U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention, National Institute for Occupational Safety and Health, R01-OH-008967, 2012 Nov; :1-17
• Contact Point Address:
  Dr. Roger P. Hamernik, State University of New York at Plattsburgh, Auditory Research Laboratories, 101 Broad St, Plattsburgh, N.Y. 12901
• Email:
  Roger.Hamernik@plattsburgh.edu
• Federal Fiscal Year:
  2013
• NORA Priority Area:
  Manufacturing
• Performing Organization:
  Plattsburgh State University, Plattsburgh, New York
• Peer Reviewed:
  False
• Start Date:
  20080901
• Source Full Name:
  National Institute for Occupational Safety and Health
• End Date:
  20120831
• Collection(s):
  National Institute for Occupational Safety and Health
• Main Document Checksum:
  urn:sha-512:9fd48c9a736a01b8429a58a4a16d6a4bd23b7a094a19b1909dd9f4f8b2159c0db9de40003a8dae33e1a8d8772e7adb33c2591ea9eb10bc4b8c27d410c2ca14e9
• Download URL:
  https://stacks.cdc.gov/view/cdc/218965/cdc_218965_DS1.pdf
• File Type:
  [PDF - 168.72 KB ]