Moshammer 2015.
Methods | CBA (arc sin transformed linear regression analysis of HPD use on NIHL) Austria Type of industry: steel factory |
|
Participants | Fitters and welders at a steel factory Age mean 16.4 years, range 14‐19 years Gender, hearing ability not reported Average noise exposure: 90.8 dB(A) (range 85.4‐107.4 dB(A) over 13.3 years (range 2‐23 years) n = 125 |
|
Interventions | Use of hearing protection, self‐reported percentage of use, geometric mean of responses on questionnaire at 3 annual health examinations when hearing level was measured | |
Outcomes | Noise‐induced hearing loss at 4 kHz, measured as hearing level at baseline adjusted for age minus hearing level at end of follow‐up adjusted for age (information from the study authors); also the average hearing loss at 2, 3 and 4 kHz was measured | |
Notes | The study was set up to predict hearing loss based on TTS at start of the study. Participants were selected from a cohort of workers that started as apprentices at the firm between 1982 and 1989, who had at least 2 years of noise exposure and who worked at places that were noisy | |
Risk of bias | ||
Bias | Authors' judgement | Support for judgement |
14. Blinding (subjects) | High risk | No blinding; participants chose themselves about wearing or not wearing HPD |
15. Blinding (outcome assessors) | Unclear risk | Not clear if audiometrists were aware of HPD use |
16. Retrospective unplanned subgroup analyses | High risk | different data analysis and results because of comments on journal article |
17. Follow‐up | Low risk | regression analysis includes adjustment for different time of follow‐up (noise years) |
18. Statistical tests | Low risk | multivariate analysis |
19. Compliance | Unclear risk | self‐reports of use of hearing protectors in noisy areas, unclear how valid |
20. Outcome measures | Low risk | NIHL was calculated from thresholds measured with audiogram minus age related HL |
21. Selection bias (population) | Low risk | same workplaces, type of work (closed cohort in 1 steel plant) |
22. Selection bias (time) | Low risk | workers in I and C started between 1982‐1989 |
23. Randomization | High risk | no randomisation, participants chose themselves how often they used HPDs in noisy areas |
24. Allocation concealment | High risk | no randomisation |
25. Adjustment for confounding | Low risk | adjusted confounders |
26. Incomplete outcome data | High risk | > 60% lost to follow‐up |