Davies 2008.

Methods	CBA study
Participants	Workers in lumber mills during 1979‐1996 who had at least 2 hearing tests n = 22,376 Canada, British Columbia
Interventions	Intervention: hearing conservation programme; n = 16,347 Control: those exposed to < 80 dB‐years plus those at their first hearing test following baseline; n = 6002 estimated from the number of person‐years of 41,357 with 6.8‐year follow‐up
Outcomes	STS: ≥ 10 dB at 2, 3 or 4 kHz in the better ear
Notes	Long‐term Comparability ‐ intervention/control: proportional hazards model to adjust for age and hearing ability at baseline
Risk of bias
Bias	Authors' judgement	Support for judgement
14. Blinding (subjects)	High risk	not blinded
15. Blinding (outcome assessors)	High risk	not blinded
16. Retrospective unplanned subgroup analyses	Low risk	no subgroup analyses
17. Follow‐up	Low risk	same time period for cases and controls (dB‐years)
18. Statistical tests	Low risk	multivariable (Cox) regression analyses, HR (95% CI)
19. Compliance	Unclear risk	no information provided
20. Outcome measures	Unclear risk	audiometry quality not reported (hearing thresholds, STS)
21. Selection bias (population)	Low risk	same industry
22. Selection bias (time)	High risk	different time period
23. Randomization	High risk	no randomisation
24. Allocation concealment	Unclear risk	no randomisation, not applicable
25. Adjustment for confounding	Low risk	adjusted for age and hearing loss
26. Incomplete outcome data	Unclear risk	not reported