A Mixed-Methods Assessment of Hearing Conservation Program Effectiveness

Transcription

1 A Mixed-Methods Assessment of Hearing Conservation Program Effectiveness Rick Neitzel, PhD, CIH, FAIHA Department of Environmental Health Sciences University of Michigan Peter Rabinowitz, MD, MPH Department of Environmental and Occupational Health Sciences University of Washington

2 What s in a hearing conservation Exposure limit e.g., OSHA 85 dba 8-hour TWA Action Level program? Assessment, control of noise Training Hearing protection devices Audiometric testing Record keeping Introduction 2

3 Problems Despite HCPs, NIHL common Many workers not covered HCPs may not be effective Need methods to assess HCP effectiveness E.g., American National Standards Institute S , Standard Evaluating the Effectiveness of HCPs (not adopted) Year-to-year variability of sequential audiograms; lagging indicator Introduction 3

4 In the absence of agreed-upon methods of assessing program quality and effectiveness, m The need No comprehensive, evidence-based methodology for assessing HCP effectiveness and quality Managers of HCPs may instead Use agency program audit checklists Compliance with minimum regulatory program requirements Track annual occurrence of HL indicators (standard threshold shifts, early flags ) Assess changes in noise over time Introduction 4

5 The opportunity Explore value of different elements of HCP in preventing NIHL Opportunity Work with large metals manufacturing company to assess HCPs at multiple sites At time of study, company had ~40 US sites Criteria: HCP, noise/audiometry for >=5 years 15 sites contacted, 14 participated, one used for pilot testing All information confidential Introduction 5

6 Site visits ( ) Site visits of 2 3 days duration by 4-6 research staff Walkthroughs Manager interviews with HCP and medical staff (~50 items, 2 hours), including program costs Worker surveys (8 items, 60 sec, self-administered during meetings, breaks; ~20-30% of workforce per site) Dosimetry (full-shift, OSHA HCA) Focus groups (4-12 people, min, 8-15 per site) Audiometric data analysis Methods 6

7 Audiometric analysis For each site, calculated Slope of annual change at 3, 4, 6 khz for individuals with >=3 audiograms over >= 5-years in past 10 years Annual % of tests with improvement or worsening >=15 dbhl from Hz) over past 5 years Corrected, uncorrected shift rates Methods 7

8 Noise analysis Calculated average noise levels by site Measurements made by site over past 10 years Our measurements during site visits Methods 8

9 Statistical analyses program characteristics vs. NIHL risk Multilevel statistical model Outcome was each individual s annual rate of change in binaural average of 3, 4, 6 khz Individual-level predictors: worker age, race/ethnicity, sex, baseline HTL, tinnitus Site-level predictors: management survey items, worker survey items, average noise exposures, STS rates, and audiogram test-test variability To account for unmeasured between-site variation, site included as random effect Methods 9

10 Site and worker characteristics Complete data from 13 sites Characteristic mean/n sd/% Plant characteristics (n=13) No. of workers - mean, sd Better/worse > 15 db statistic- mean, sd (%) Annual overall STS rate - mean. Sd Employee characteristic (n=8,028) Age - mean,sd Gender - n,% Male Female Race/Ethnicity - n,% White Black Hispanic Other Hearing at baseline (mean 3,4,6KHz) - mean,sd Annualized hearing loss rate (db/yr avg 3,4,6KHz) - mean,sd Better-worse statistic ranged from 31.3 to 44.5% Results 10

11 NIHL (average change 3, 4, and 6 khz, ) Mean annual age-corrected STS rate 3.5 ± 2.0% Results 11

12 Manager survey responses Lack of variability between sites for many items Response No Yes n % n % Annual audiograms conducted before (vs during/after workshift) Hearing protection devices required full shift exposure >85 (vs other than >85 or >90) Noise monitoring conducted by personal measurement (compared to personal and area) Work-relatedness determined by Professional supervisor OR Physician or Audiologist (vs other) Have any workers complained that HPDs interfere with their ability to do their jobs at your facility (yes vs no) Results 12

13 Worker survey responses More variability than manager survey Question mean sd Minumum Maximum When you had your most recent hearing test, did anyone talk to you about your hearing? (% yes) When you had your most recent hearing test, did anyone talk to you about how to best use your hearing protective device? (% yes) What percentage of time does your supervisor wear his/her hearing protective device while on the floor?(% %) In areas where hearing protection is required, how often do other workers wear their hearing protective devices? (% always) How would you rate the level of commitment to preventing hearing loss at your workplace? (% high or exceptional) How much of a personal concern do you have about losing your hearing because of on the job noise exposure? (% high or extreme) How much of a personal concern do you have about injuries or accidents occurring because of noise, hearing loss or wearing hearing protective devices? (% high or On average, how many times during each work day do you have to remove your hearing protective device in order to communicate? (% never) Results 13

14 Focus groups and noise level results 126 groups (6-14 per site), 767 workers Most sites: strong mgmt commitment to HCP, strong personal commitment to protecting hearing on job Fewer sites: strong personal commitment to protecting hearing off job Noise (513 measurements, per site) Average TWA 83.1 ± 3.1 dba Overall, 168 (32.7%) of TWAs >85 dba OSHA AL By site, 18-75% of TWAs >85 dba Results 14

15 Mean TWA level measured by company (dba) Noise level results Reasonably strong correlation between site measurements and ours Mean TWA level measured by research staff (dba) Results 15

16 Program cost results ($) Mean annual cost per plant $197,000, range $67,016-$396,638 Mean annual cost per worker: $294 Results 16

17 Costs by program element 17

18 Site costs by program element Results 18

19 Statistical modeling results Characteristic estimate p-value estimate p-value Worker Level Age < <.0001 Male (vs female) < <.0001 Race/Ethnicity Black (vs white) < <.0001 Hispanic (vs white) Tinnitus < <.0001 Tinnitus unknown Baseline hearing (mean 3,4,6KHz) <.0001 Location level Employee Survey When you had your most recent hearing test, did anyone talk to you about your hearing? (% yes >=85%) <.0001 Workplace level of commitment to HL (high or exceptional >=70% ) Manager Survey Partially adjusted * Annual audiograms conducted before (vs during/after workshift) Hearing protection devices required full shift exposure >85 (vs other than >85 or >90) Adjusted Noise monitoring conducted by personal and area measurement (compared to personal only) Work-relatedness determined by Professional supervisor OR Physician or Audiologist (vs other) Have any workers complained that HPDs interfere with their ability to do their jobs at your facility (yes vs no) Cost, noise, audiometric test-test variability not associated with NIHL risk Results 19

20 Conclusions: manager interviews Most questions did NOT correlate with NIHL rates Audiometry before rather than during workshifts associated with lower rates of NIHL During can detect temporary threshold shifts; best practice Sites that measured area and personal noise had less NIHL than did personal only Attention to noise exposure and control could play role Sites where managers reported employees complained about interference by HPDs had higher rates of NIHL Hearing protection might be less effective in such locations Conclusions 20

21 Manager interviews: possible reasons for limited variability All sites part of single company with harmonized approach to HCP Managers may have recognized correct answers, given those responses Managers not familiar with HCP elements, activities Variations in HCP activities not captured by interviews E.g., frequency of trainings may not reflect quality of training Discussion 21

22 Conclusions: worker surveys Sites where workers reported someone talked to them about hearing during audiometry had lower NIHL rates Counseling could have beneficial impact on worker behavior More effective use of HPDs Avoiding noise exposures at and away from work Perceived management commitment to hearing conservation associated with reduced NIHL risk Safety climate and culture can impact effectiveness of HCPs Conclusions 22

23 Conclusions: program evaluation Ask your employees their perception of the program! 23

24 Conclusions: noise exposures Even with long-standing HCP, overexposures common 1/3 of measurements exceeded OSHA AL overall At several sites, 1/2 to 2/3 of measurements exceeded AL Manager survey items on intensity of noise control efforts not predictive of NIHL rates Noise levels not predictive of NIHL risk Due to unmeasured individual differences in protector use? Worker survey question on how often workers removed protectors during work shift not correlated with NIHL risk Conclusions 24

25 Strengths and limitations Strengths Large number of facilities assessed In-depth, mixed methods approach at each site Limitations Retrospective nature of outcome data All sites part under company-wide HCP policy Inclusion of more diverse companies may have shown greater variation in HCP elements, manager survey assessments Discussion 25

26 Next Steps Web based assessment tool: 26

27 Acknowledgements Data collection/analysis Linda Cantley Deron Galusha Reginald Jennings Gemina Garland-Lewis Michael McTague Blanca Paccha Michael Park Stephanie Sayler Carol Merry Stephenson Kan Sun Alice Suter Assistance Workers, management at participating sites Funding National Institute for Occupational Safety and Health (5 RO1 OH010132) 27

28 Questions? Rick Neitzel Peter Rabinowitz Conclusions 28