Cochrane method for systematic review and meta-analysis of interventions to prevent occupational noise-induced hearing loss - abridged.

PURPOSE
Assess the effect of non-pharmaceutical interventions at work on noise exposure or occupational hearing loss compared to no or alternative interventions.


RESEARCH STRATEGIES
Pubmed, Embase, Web of Science, OSHupdate, Cochrane Central and Cumulative Index to Nursing and Allied Health Literature (CINAHL) were searched.


SELECTION CRITERIA
Randomized Controlled Trials (RCT), Controlled Before-After studies (CBA) and Interrupted Time-Series studies (ITS) evaluating engineering controls, administrative controls, personal hearing protection devices, and hearing surveillance were included. Case studies of engineering controls were collected.


DATA ANALYSIS
Cochrane methods for systematic reviews, including meta-analysis, were followed.


RESULTS
29 studies were included. Stricter legislation can reduce noise levels by 4.5 dB(A) (very low-quality evidence). Engineering controls can immediately reduce noise (107 cases). Eleven RCTs and CBA studies (3725 participants) were evaluated through Hearing Protection Devices (HPDs). Training of earplug insertion reduces noise exposure at short term follow-up (moderate quality evidence). Earmuffs might perform better than earplugs in high noise levels but worse in low noise levels (very low-quality evidence). HPDs might reduce hearing loss at very long-term follow-up (very low-quality evidence). Seventeen studies (84028 participants) evaluated hearing loss prevention programs. Better use of HPDs might reduce hearing loss but other components not (very low-quality evidence).


CONCLUSION
Hearing loss prevention and interventions modestly reduce noise exposure and hearing loss. Better quality studies and better implementation of noise control measures and HPDs is needed.


INTRODUCTION
Worldwide millions of workers are exposed to noise levels that increase their risk of hearing disorders (1) . While hearing loss prevention programs (HLPPs) are mandatory in many countries, the reportedly continuing high rate of occupational noise-induced hearing loss (NIHL) casts doubt upon their effectiveness (2) . Moreover, the broad range of interventions included in HLPPs makes it difficult to appraise the most effective strategy. A systematic review of studies that evaluated interventions to reduce occupational exposure to noise or to decrease occupationally induced hearing loss is therefore warranted. This paper summarizes the main results of the second update of the Cochrane review originally published in 2009.

Purpose
To assess the effectiveness of non-pharmaceutical interventions for preventing occupational noise exposure or occupational hearing loss compared to no or alternative interventions.

Research strategy
This is an abridged version of the second update of a Cochrane Review originally published in 2009 based on the methods originally described in the review protocol (3) . Systematic searches were conducted combining search words for the occupational setting, exposure, interventions, and effects on noise or hearing loss. No restrictions on language were used, publication year or publication status and were searched Pubmed, Embase, Web of Science, OSHupdate, Cochrane Central and Cumulative Index to Nursing and Allied Health Literature (CINAHL) databases up until September 2016 (search history in Appendix 1 and Appendix 2). To determine which studies fulfilled the inclusion criteria, pairs of the review authors independently scanned the titles and abstracts of every record retrieved from the databases. Full articles were retrieved for further eligibility assessment.
Data were independently extracted for each included study and resolved discrepancies by discussion. A standard form to extract information about was used: study design, randomisation methods, setting, participants, interventions, outcome measures, follow-up, and adverse events. To assess whether HLPPs are as good as not being exposed, it had to be made an assumption about the minimal clinically relevant hearing loss. Hearing loss was associated with exposure to 85 dB(A) as the minimum amount of damage that should be avoided by the interventions. Based on International Organization for Standardization (ISO) 1990 (4) , the amount of hearing loss after five years of exposure to 85 dB(A) was calculated for the median, 10th and 90th percentile would be 4.2 dB, 2.1 dB and 6.1 dB, respectively. This is equivalent to a mean of 4.2 dB hearing loss and represents clinically relevant hearing loss (5) . This means, the 95% CI from meta-analysis results on hearing loss can include zero, but not 4.2 to assure that the protected and non-exposed groups are equivalent (6) .

Selection criteria
We included studies that 1) used a randomised controlled, controlled before-after, or interrupted time-series study design, 2) included workers exposed to noise levels greater than 80 dB(A), 3) concerned interventions aimed at reduction of noise exposure to prevent NIHL, and 4) used noise exposure or NIHL as an outcome. Case studies on the effects of engineering control interventions without a control group could be included. The results of case studies for the conclusions of the review, as the study design did not fulfil our inclusion criteria, were not used.

Data analysis
Eight authors of recent studies were contacted regarding missing or unclear information and were obtained additional data from three (7)(8)(9) .
When authors reported results separately for participant groups (10,11) we combined these following the Cochrane Handbook for Systematic Reviews of Interventions guidance (12) . In two studies, multiple interventions were compared with one control group. To avoid using the same control group data more than once, the control group was split into three (13) or two (14) equal subgroups that were subsequently combined in the meta-analysis.
To evaluate the risk of bias, the quality criteria presented by Ramsay et al. (15) for ITS studies was used. For RCTs and cohort studies, the internal validity items by Downs and Black (16) were used that are mostly congruent with the Cochrane 'Risk of bias' tool (17) . We defined studies' overall risk of bias as low if they scored more than 50% of the maximum score.
Sufficiently homogeneous studies, regarding interventions, participants, settings and outcomes in a meta-analysis, were combined. When results were statistically heterogeneous according to the I 2 statistic, a random-effects model for the meta-analysis was used. A sensitivity analysis to assess the influence of risk of bias on the pooled effect sizes was conducted. We deemed change in hearing level at 4 kHz and Standard Threshold Shifts (STS) as similar outcome measures for hearing effects and calculated Standardized Mean Differences (SMD) to enable combination of both measures in the meta-analysis (18) . For easing interpretation, we transformed the pooled SMDs back to a mean change in hearing level in dB using the median standard deviation of the included studies.
The Grading of Recommendation, Assessment, Development and Evaluation (GRADE) approach to rate the quality of the evidence for each outcome was followed. The grading is based on study design, risk of bias, consistency, directness (generalisability), precision and publication bias across all studies (19) . Overall quality is considered high for RCTs and low for observational studies and can be further reduced or upgraded (20) (Table 1).
Ratings are interpreted as: 1) high-quality evidence is unlikely to change, moderate-quality evidence; 2) further research is likely to have an impact and may change estimates, low-quality evidence; 3) further research is very likely to have an important impact, and very low-quality evidence provides very uncertain effects estimates.
The results for the most important comparisons in 'Summary of Findings' (SoF) were presented tables.
While the participants in all studies were described as being exposed to noise at work, these descriptions were often based on measurement methods that were not clearly described. We assumed that the noise exposure was higher than 80 dB(A). Noise-exposed participants worked in construction, mining, manufacturing, agriculture, forestry, military, an orchestra, unspecified company or in various workplaces. One study did not describe workplaces (40) .
In most studies, only men were included or there were mostly male workers at the workplaces studied.
Most studies scored poorly on all aspects of the risk of bias checklist ( Figure 2) and only six studies scored an overall low risk of bias (13,22,28,29,37,40) .  The effect of engineering interventions (following legislation) on noise exposure was evaluated in one ITS study. The study (8) found that new legislation in the mining industry reduced the median personal noise exposure dose in underground coal mining by 27.7 percentage points (95% Confidence Interval (CI) -36.1 to -19.3 percentage points) immediately after the implementation of stricter legislation (Table 3). This roughly translates to a 4.5 dB(A) decrease in noise level. The intervention was associated with a favourable but statistically non-significant downward trend in time of the noise dose of -2.1 percentage points per year (95% CI -4.9 to 0.7, four-year follow-up, very low-quality evidence).
Studies showed immediate reductions in noise levels of machinery ranging from 11.1 to 19.7 dB(A) as a result of purchasing new equipment, segregating noise sources or installing panels or curtains around sources. However, studies lacked long-term follow-up, a control group, and in some cases the outcome was evaluated by an acoustical consultant or an employee at the firm where the intervention was evaluated and a conflict of interest was apparent (14 cases).  (1) and Adera (1993) ( (38) CBA Workers, n = 852, 1 company in the chemical industry, USA HLPP HL Long-term Royster (1980) (39) CBA Workers, n = 70, various occupations, USA HPD HL Immediate Salmani et al. (2014) (40) RCT Workers, n = 150, Iran HPD NE Immediate Seixas et al. (2011) (14) RCT Construction workers, n = 176, USA HPD NE Short term Simpson et al. (1994) (41) CBA Various occupations, n = 13283, 21 companies, USA HLPP HL Long-term Caption: CBA = controlled before after study, ITS = interrupted time series analysis, RCT = randomised controlled trial, HL = hearing loss, NE = noise exposure, HPD = hearing protection device, HLPP = hearing loss prevention program; n = number Caption: Adera 1993: Forouzanfar et al. (1) and Adera (21)  The review found no effects for administrative controls on environmental noise exposure. On-site training sessions giving instructions for HPD use and noise control techniques (sound barriers and distance) did not have an effect on personal environmental noise-exposure levels compared to information only in one cluster-RCT after four months' follow-up (Mean Difference (MD) 0.14 dB; 95% CI -2.66 to 2.38). Another arm of the same study found that personal noise exposure information had no effect on noise levels (MD 0.30 dB(A), 95% CI -2.31 to 2.91) compared to no such information (176 participants, low-quality evidence) ( Table 4).
HPDs reduced noise exposure on average over various frequencies measured by about 20 dB(A) in one RCT and three CBAs (57 participants, low-quality evidence). There was moderate-quality evidence that personal instructions for inserting earplugs into the ear canal have a considerable effect on the noise attenuation of the devices with an 8.6 dB (95% CI 6.9 to 10.3) higher protection averaged across frequencies (two RCTs (37,40) , 140 participants) ( Table 5).
The effects of HPDs on hearing loss were measured in short and long-term follow-up studies. Authors of two studies compared different devices and measured temporary threshold shifts at short-term follow-up but reported insufficient data for analysis. In two CBA studies, the authors found no difference in hearing loss from noise exposure above 89 dB (A) between earmuffs and earplugs at long-term follow-up (Odds Ratio we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of effect, but there is a possibility that it is substantially different; Low quality: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect; Very low quality: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect; *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI); 1 We downgraded by two levels from high to low because of high risk of bias and imprecision Caption: CI: Confidence interval; RCT: randomized controlled trial GRADE Working Group grades of evidence. High quality: we are very confident that the true effect lies close to that of the estimate of the effect; Moderate quality: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of effect, but there is a possibility that it is substantially different; Low quality: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect; Very low quality: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect; *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the absolute effect of the intervention (and its 95% CI); 1 We downgraded by one level from low to very low because there is only one study and it has a high risk of bias Caption: CI: Confidence interval; PEL: permissible exposure level; ITS: interrupted time series analysis (OR) 0.8, 95% CI 0.63 to 1.03, very low-quality evidence) ( Table 6). The long-term evaluation of the effect of earmuffs versus earplugs on hearing loss showed that earmuffs might perform better than earplugs in high noise levels, but worse in low noise levels (very low-quality evidence). Authors of another CBA study found that wearing HPDs more often resulted in less hearing loss at very long-term follow-up (very low-quality evidence).
Studies also evaluated the effects of the combination of interventions in a hearing loss prevention programmes on noise exposure and hearing loss. One RCT found no significant effect in lowering noise level with the use of noise level indicators plus basic information or plus intensive information compared to basic information only at two-and four-months follow-up. The noise level decreased 0.32 dB more in the control group at two months (95% CI -2.44, 3.08) but 0.14 dB more in the intervention group at four months (95%CI -2.66 to 2.38). Neither were statistically significant. Also, the comparison of intensive versus basic information showed no significant differences in noise levels at two (-1.7dB, 95% CI -1.24 to 4.64) and four months (0.3 dB, 95% CI -2.31 to 2.91).
One cluster-RCT found no difference in hearing loss at three-or 16-year follow-up between an intensive HLPP for agricultural students and audiometry only (moderate-quality evidence) (Table 7). One CBA study found no reduction of the rate of hearing loss (MD -0.82 dB per year (95% CI -1.86 to 0.22) for a HLPP that provided regular personal noise exposure information compared to a program that did not provide such information (Table 8).
There was very low-quality evidence in four long-term studies, that better use of HPDs as part of a HLPP decreased the risk of hearing loss compared to less well used HPDs in HLPPs (OR 0.40, 95% CI 0.23 to 0.69) ( Table 9).This could not be shown for worker training, audiometry alone or noise monitoring by very low-and moderate-quality evidences. More individualized information on daily noise exposure as part of a HLPP showed favourable but non-significant effects on hearing loss in one study. GRADE Working Group grades of evidence. High quality: we are very confident that the true effect lies close to that of the estimate of the effect; Moderate quality: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of effect, but there is a possibility that it is substantially different; Low quality: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect; Very low quality: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect; *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI); 1 We downgraded from high quality by one level because of imprecision due to small number of participants Caption: CI: Confidence interval; RCT: randomized controlled trial; RCTs: randomized controlled trials GRADE Working Group grades of evidence. High quality: we are very confident that the true effect lies close to that of the estimate of the effect; Moderate quality: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of effect, but there is a possibility that it is substantially different; Low quality: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect; Very low quality: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect; *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI), 1 We downgraded from low quality to very low quality because of high risk of bias in both studies; Caption: CI: Confidence interval; OR: Odds Ratio; STS: standard threshold shift; CBA: controlled before after study.
In the meta-analysis of four long-term CBA studies the difference in mean changes in hearing level at 4 kHz was 0.53 dB (95%CI -0.53 to 1.68) (10,25,26,30) . We performed a sensitivity-analysis and left out one study (10) that had a high risk of bias due to a 10-year age difference between the intervention and the non-exposed group, which could explain a difference of 7dB hearing thresholds (calculated based on ISO 1990 (4) ). Sensitivity analysis results showed workers in a HLPP had a statistically non-significant 1.8 dB (95% CI -0.6 to 4.2) greater hearing loss at 4 kHz than non-exposed workers (very low-quality evidence, Table 10). The confidence interval includes a possible hearing loss of 4.2 dB which is similar to the level of hearing loss resulting from five years of exposure to 85 dB(A), which means workers might still be at risk of a clinically relevant hearing loss.
In addition, out of three other CBA studies that could not be included in the meta-analysis, two showed an increased risk of hearing loss in spite of the protection of a HLPP compared to non-exposed workers (13,38) and one CBA did not (33) . GRADE Working Group grades of evidence. High quality: we are very confident that the true effect lies close to that of the estimate of the effect; Moderate quality: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of effect, but there is a possibility that it is substantially different; Low quality: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect; Very low quality: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect; *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI); 1 We downgraded one level from high to moderate due to lack of information on randomisation and allocation concealment Caption: CI: Confidence interval; HLPP; hearing loss prevention programme; OR: Odds ratio; STS: standard threshold shift; RCT: randomized controlled trial Matched for age, gender, baseline hearing loss and baseline hearing GRADE Working Group grades of evidence. High quality: we are very confident that the true effect lies close to that of the estimate of the effect; Moderate quality: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of effect, but there is a possibility that it is substantially different; Low quality: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect; Very low quality: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect; *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI); We downgraded by one level from low to very low because of high risk of bias Caption: CI: Confidence interval; HLPP: hearing loss prevention programme; CBA: controlled before after study

DISCUSSION
We could not find any controlled studies, in which technical measures to reduce workers' noise exposures were evaluated at the company level. Some argue that control groups are not necessary because the effect can be measured immediately (55) . On the other hand, the measurement of noise levels in real working life can be biased by many operational and environmental factors. To address this issue, we systematically collected case studies.
The immediate results of those studies are similar to those of HPDs. Noise control can potentially make HPDs in workplaces unnecessary, along with other components of hearing conservation programs. However, for most case studies, it was unclear if the measured reductions also effected personal noise level exposure. Other case studies measured personal noise exposure of workers but did not report measurement protocols and the personal effect remains uncertain. Moreover, long-term follow-up is missing and it is unclear if these are lasting solutions. Many potential GRADE Working Group grades of evidence. High quality: we are very confident that the true effect lies close to that of the estimate of the effect; Moderate quality: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of effect, but there is a possibility that it is substantially different; Low quality: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect; Very low quality: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect; *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI); 1 STS used in two studies, change of mean 4 kHz threshold in one study; 2 Number of events based on median event rate in included studies; 3 Result from the meta-analysis of three studies; 4 One extra study provided similar evidence but could not be combined in the meta-analysis; 5 We downgraded by one level from low to very low because of risk of bias due to lack of adjustment for age and hearing loss Caption: CI: Confidence interval; HLPP: hearing loss prevention programme; OR: Odds ratio; STS: standard threshold shift; SMD: standardized mean difference Table 10. SoF table -HLPP versus non-exposed workers (hearing loss) Hearing loss prevention programme (HLPP) compared to non-exposed workers Patient or population: workers Settings: exposure to noise Intervention: HLPP Comparison: non-exposed workers ⊕⊝⊝⊝ very low 3,4 pooled effect size 0.17 (95% CI -0.06 to 0.40) recalculated into dBs GRADE Working Group grades of evidence. High quality: we are very confident that the true effect lies close to that of the estimate of the effect; Moderate quality: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of effect, but there is a possibility that it is substantially different; Low quality: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect; Very low quality: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect; *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI); Assumed increase of hearing threshold: median of three studies with respectively 3.4, 3.6 and 5.2 dB increase in hearing threshold at 4 kHz after five years' follow-up; 2 Results from three of five studies were included in sensitivity analysis because one study was at serious risk of bias and one other study showed that in spite of hearing protection workers were still more at risk than non-exposed workers; 3 We downgraded by one level from low to very low because three studies did not adjust for age and hearing loss at baseline; 4 We would have downgraded by one more level because the confidence interval does not exclude a risk of hearing loss similar to exposure to 85 dB(A) but we had already reached a rating of very low quality evidence Caption: CI: Confidence interval; HLPP: hearing loss prevention programme biases in the uncontrolled studies would be remediated by the use of control groups, better reporting of noise measurement protocols and long-term follow-up measurements.
No studies evaluated the effectiveness of recommendations from occupational health services, national agencies or occupational health professionals to reduce noise levels. Regulations regarding noise at work can make it difficult to challenge current practice in experiments.
For immediate effects of HPDs, we restricted our inclusion criteria to field studies among workers and excluded studies that made use of volunteers or were carried out in the laboratory. All excluded studies showed a benefit of extra instruction compared to less or no instruction (56)(57)(58)(59) . The increase in attenuation was similar to that found in our review. We only included studies that compared different devices worn by the same workers because the evaluation depends to a great extent on the wearer. That criterion excluded a great number of studies that evaluated different devices worn by different workers, but provided us with more reliable results.
Researchers who intended to evaluate a HLPP did not clearly define its implementation, which is especially important in studies comparing HLPPs. It is unclear if the results are applicable in other settings and what measures were taken in addition to HPDs (e.g. training).
The risk of bias was high (especially for long-term evaluation studies) because most studies were set up retrospectively and it is difficult to control confounders. Individual factors, such as skills necessary to correctly use HPDs or age, have an important effect on the outcome but only some studies used randomisation to ensure no baseline differences. Consequently, there is a need for better quality evidence. It has often been argued that randomisation of workers or workplaces is not possible but two studies that evaluated a HLPP (or components thereof) showed that randomisation was feasible, even in difficult sectors such as construction (14,22) . Evidence from more RCTs would eventually yield much higher-quality information on the effectiveness of hearing loss prevention programs.
Even though, we made significant efforts to search databases that would contain grey literature seeing that we did not go through conference proceedings. It is therefore possible that we missed retrospective cohort studies or controlled noise-reduction studies.
Publication bias could play a role in the results of the evaluation studies of HLPPs, with four of the studies being funded or carried out by people employed by the company responsible for the intervention, who could possibly have an interest in publishing studies demonstrating a preventative effect of HLPPs (13,33) .
Other authors drew similar conclusions to our review but mostly applied less systematic approaches.
One review located 22 studies that evaluated the field performance of many different types of HPDs worn by different workers (60,61) . The inclusion criteria of these studies were essentially different from ours because only studies comparing devices among the same subjects were included. However, the conclusions from all these studies are in agreement: under field conditions the noise attenuation of HPDs is much less than under laboratory conditions. Another review concluded that the evidence from long-term evaluation studies does not support HLPPs' effectiveness (62) , but the search for studies was not systematic. The review included five studies, of which four were also included in this review. His conclusions for the effectiveness of HLPPs are similar to ours.
Authors from other studies reviewed occupational NIHL data (63) , evaluated the quality of HLPPs in companies (64) , or performed a narrative review directed at the mining sector alone (65) . All studies concluded either that HLPPs are ineffective, or programs are commonly incomplete and miss noise control interventions.
There is very low-quality evidence that implementation of stricter legislation can reduce noise levels in workplaces. Case studies showed promising effects of engineering control on noise reduction at immediate follow-up but controlled studies and evaluation of the long-term effects are missing. It is unclear if results can be replicated in other workplaces and what the long-term effects are.
Under field conditions the average noise reduction of HPDs is lower than indicated ratings provided by the manufacturers. There is moderate-quality evidence that training of proper insertion of earplugs significantly reduces noise exposure at short-term follow-up but long-term follow-up is still needed.
There is very low-quality evidence that the better use of HPDs as part of HLPPs reduces the risk of hearing loss, whereas for other program components of HLPP we found no effect.
The absence of conclusive evidence should not be interpreted as evidence of lack of effectiveness. Rather, it means that further research is very likely to have an important impact.
Future studies should use randomised design for HPDs or comparisons of different HLPPs or single programme components, or different levels of implementation in a cluster-randomised design. The ITS design has potential for evaluating HLPPs because much data is collected routinely.

CONCLUSION
Hearing loss prevention interventions modestly reduce noise exposure and hearing loss. Better quality studies and better implementation of noise control measures and HPDs is needed.

Disclaimer
The findings and conclusions in this report are those of the author(s) and do not necessarily represent the views of the National Institute for Occupational Safety and Health.