A systematic review of the interventions to promote the wearing of hearing protection

ABSTRACT CONTEXT AND OBJECTIVE: Noise-induced hearing loss can only be prevented by eliminating or lowering noise exposure levels. When the source of the noise cannot be eliminated, workers have to rely on hearing protection equipment. The aim here was to summarize the evidence for the effectiveness of interventions to enhance the wearing of hearing protection among workers exposed to noise in the workplace. DATA SOURCE: Studies with random assignment were identified by an electronic search of the medical literature up to 2005. Data were double-entered into the Review Manager software, version 4.2.5. DATA SYNTHESIS: Two studies were found. A computer-based intervention tailored to individual workers’ risks and lasting 30 minutes was not found to be more effective than a video providing general information for workers. A second randomized controlled trial evaluated the effect of a four-year school-based hearing loss prevention program among schoolchildren working on their parents’ farms. The intervention group was twice as likely to wear some kind of hearing protection as was the control group (which received only minimal intervention). REVIEWERS’ CONCLUSIONS: The limited evidence does not show whether tailored interventions are more or less effective than general interventions among workers, 80% of whom already use hearing protection. Long-lasting school-based interventions may increase the use of hearing protection substantially. Better interventions to enhance the use of hearing protection need to be developed and evaluated in order to increase the prevention of noise-induced hearing loss among workers.


INTRODUCTION
Noise-induced hearing loss is one of the most common occupational diseases and the second most common self-reported occupational illness or injury.The condition is permanent and irreversible.Furthermore, there is no effective treatment for permanent hearing loss resulting from excessive noise exposure.However, the condition can be prevented by preventive measures, and sets of such measures are usually combined into hearing loss prevention programs (also called hearing conservation programs).Personal protective measures (e.g.earplugs or earmuffs) always form one part of a hearing loss prevention program. 1 Even though the effi cacy of these measures for shielding the inner ear from noise has been proven in laboratory settings, their effectiveness for preventing hearing loss from accumulated noise exposure depends mainly on how regularly they are used by workers.Studies have shown that if workers do not wear hearing protection for 100% of the time, its effectiveness will quickly diminish.For example, wearing hearing protection for only 90% of the time will decrease effectiveness to less than one third. 2Educational or behavioral interventions to promote its use are therefore important preventive measures.

Prevalence and etiology Prevalence and etiology
The risk of hearing loss due to noise exposure has long been recognized.Approximately 30 million workers in the United States alone are exposed to hazardous noise at work. 3 Early damage is typically sustained in the basal turn of the cochlea and affects hearing in the frequency range from 3000 to 6000 Hertz (Hz) (the frequency range of speech).One study calculated an annual worldwide incidence of noise-induced hearing loss of 1,628,000 cases. 4ith a worldwide population of 6.525 billion this is equal to 25 per 100,000 per year.Long-term exposure to noise levels greater than 80 dB(A) (i.e.situations in which you have to raise your voice if you want to communicate with someone who is within a distance of one meter) carries an increased risk of hearing loss, which increases exponentially with the noise level.The risk of hearing impairment (average hearing loss > 35 dB(A) at 1, 2 and 3 kHz) at age 60 due to 40 years of exposure to noise levels of 100 dB(A) has been estimated as 55%. 5 Concurrent exposure to ototoxic substances, such as solvents and heavy metals, may increase the potential for damage from noise. 6Individuals' susceptibility to the adverse effects of noise exposure is highly variable and cannot be accurately predicted.

Prevention
Hearing loss prevention programs in industry have been widely advocated.Occupational health and safety legislation obliges employers to take preventive measures in most countries. 1These have proven to be effective to some extent in countries like Finland, where the incidence of cases of noise-induced hearing loss halved between 1987 and 2002. 7In the European Union and the United States, assessment of exposure to noise is obligatory, as is periodic screening of workers exposed to certain noise levels.Employers are also obliged to follow a "hierarchy of hazard controls".This is designed to eliminate hazards in the workplace in a particular order, by establishing controls at the source of the hazard (engineering or administrative) before using less reliable human controls (in this case the wearing of personal protective equipment).However, technical or economic reasons may mean that human controls are heavily relied upon.Indeed, a recent study has shown that personal protective equipment is still a widely used hazard control. 8From laboratory studies, it is known that this equipment (earplugs and earmuffs) is effective in reducing exposure to noise, although this effectiveness can lessen under fi eld conditions. 9

Hearing protection programs Hearing protection programs
Studies in the United States indicate that there has been an increase in the use of hearing protection, but that there is still ample room for improvement. 105][16][17][18] Until now, there has been no systematic review that summarized the results from these trials.
The aim here was to summarize the evidence for the effectiveness of interventions to enhance the wearing of hearing protection among workers exposed to noise in the workplace.

Selected studies Selected studies
A literature search covering the period from January 1996 to June 2005 was conducted using the following databases: Medical Literature Analysis and Retrieval System Online (Medline), Excerpta Medica database (Embase), Literatura Latino-Americana e do Caribe em Ciências da Saúde (Lilacs), National Institute for Occupational Safety and Health (NIOSHTIC), International Occupational Safety and Health Information Centre (CIS-DOC), Cumulative Index to Nursing & Allied Health (CINAHL) and Cochrane Ear, Nose and Throat Disorders Group Specialized Register, the Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library, Issue 2 2005).The following exhaustive list of synonyms for hearing protective devices was used: ear protective device, hearing protective device, hearing protector, hearing protection, ear muff, ear plug, ear defender, protective equipment, (noise, occupational).In addition, the references given in the selected papers were examined, irrespective of the publication year.
Approximately 1500 titles were requested.After the reading of all titles, the reviewers (Regina El Dib and Régis Andriolo) selected 150 potential full articles to be included in the review if they had a randomised design, if they were among noise-exposed (> 80 dB(A)) workers or pupils, if there was some kind of intervention to promote the wearing of hearing protection (compared to another intervention or no intervention), and if the outcome measured was the amount of use of hearing protection or a proxy measure thereof.Of this total, 138 articles were excluded from the review because they were classifi ed as case reports, narrative reviews or letters to editors.Thus, following this assessment of full articles, only 12 publications were considered for inclusion in this review.Six studies were then excluded because they did not meet the inclusion criteria of the review (retrospective studies or non-randomized studies): Davis and Sieber (2002), 19 Ewigman et al. (1990), 20 Roeser et al. (1983), 21 Toivonen et al. ( 2002), 22 Walker (1972) 23 and Williams (2004). 24A further three are awaiting assessment because of the poor quality of the reports (Lusk et al. 1999, 14 Sadler and Montgomery 1982 17 and Zohar et al.1980 18 ).Thus, only two studies (three publications) that met the minimum methodological requirements were included in this review (Knobloch and Broste 1998, 25 Lusk et al. 2003 15 and Lusk et al. 2004 16 ).

Methodological quality Methodological quality assessment
In order to assess the methodological quality of the selected studies, a quality assessment list was developed (Appendix 1).The items incorporated are generally accepted methodological criteria.The methodological quality of the trials included in this review was measured using the criteria described in the Cochrane Handbook, 26 since scales and checklists are not a reliable method for assessing the validity of a primary study. 27wo reviewers (Regina El Dib and Régis Andriolo) independently assessed the trial quality of each study, in accordance with this assessment list.Subsequently, disagreements between the examiners, which were small, were discussed to reach a consensus.

Statistical procedure Statistical procedure
Using the available reported data, 2 x 2 tables were constructed to relate determinants to outcomes.In these studies, the outcomes assessed were the proportion of participants who were wearing hearing protection devices relative to the proportion in the control group; intention to use the devices; perceived benefi ts from using hearing protection; barriers to the use of hearing protection (self-reported use of hearing protection); and awareness of risk.Relative risk (RR) and weighted mean difference (WMD) were used to make estimates of the effects from the treatment in the included studies.Furthermore, the 95 per cent confidence intervals (CI) of these RR and WMD were calculated.

Quality assessment Quality assessment
Lusk et al. 15 and Lusk et al. 16 described their allocation method as "computer-generated".There was otherwise no mention of allocation concealment.These studies were therefore graded B (unclear) with regard to quality of allocation concealment.Knobloch and Broste 25 made use of a method to avoid contamination of participants: randomization by clusters.With regard to allocation concealment and generation of allocation, Knobloch and Broste 25 was graded B, because these details were not described in the paper.Randomization by clusters, as used by Knobloch and Broste, 25 is thought to be the best way to circumvent the problem of "contamination" inherent to interventions of an institutional nature, in which participants randomized to distinct interventional approaches have the chance to exchange their experiences in a common occupational environment.
Detection bias was present in the Knobloch and Broste 25 study, since the intervention and data collection were carried out by the researchers.Knobloch and Broste 25 did not allow for the cluster effect, which could have been estimated by providing intracluster correlation coeffi cients. 28e corrected for the cluster effect by calculating a possible design effect.The intracluster correlation coeffi cient from other school-based interventions can be estimated as 0.006. 23The mean cluster size was 22.1.The design effect (1+(m-1)*r) is thus 1+(22.1-1)*0.006= 1.1266.We divided the number of events and the sample sizes by the design effect and entered them into RevMan.
In the Lusk et al. 15 and Lusk et al. 16 studies, a rigorous method of randomization was used, which was generated by computer and sealed at the time of allocation.Nonetheless, these studies present an indication of selection bias because the trialists did not use a cluster randomization process.
Contrary to what could be expected from the long-term study by Knobloch and Broste, 25 only 6.4% of the total number of participants in the intervention group and 10.05% of the control group dropped out from the study.The low dropout rates give this study a low risk of bias.The Lusk et al. 15 and Lusk et al. 16 studies were considered to present a high risk of bias, because they did not meet the attrition criterion.The withdrawal rate within their relatively short study periods was 53.2%.

Pooling
The clinical and methodological diversity found in the included studies meant that it was not possible to combine studies in a meta-analysis.Therefore, we only performed representations of meta-analyses, as follows: Participants with personal tailored Participants with personal tailored information versus participants information versus participants with non-tailored (general) with non-tailored (general) information versus control information versus control (commercial videotape) group (commercial videotape) group Graph 1 shows a representation of a meta-analysis in relation to the outcome "Mean percentage of time actually using hearing protection device in required areas, over the preceding week and month".There was a statistically signifi cant difference favoring the participants who received tailored information rather than non-tailored information in the Lusk et al. 15 study, with a weighted mean difference (WMD) of 5.70% and 95% confi dence interval (CI) of 1.79 to 9.61.Comparing the tailored information strategy versus control (commercial videotape), there was also a statistically signifi cant difference favoring the participants who received the tailored intervention (WMD 6.40%; 95% CI: 2.42 to 10.38), as shown in Graph 2. Finally, there was no statistically signifi cant difference between the non-tailored information group and the control group (WMD 0.70%; 95% CI: -3.63 to 5.03), as shown in Graph 3.
Boosters (mailed flyers): Boosters (mailed fl yers): All participants with tailored All participants with tailored information versus all participants information versus all participants with non-tailored information with non-tailored information Graph 4 shows a representation of a meta-analysis in relation to the outcome "Mean percentage of time actually using hearing protection device in required areas, over the preceding week and month".There was a statistically signifi cant difference favoring the participants with tailored information in the subgroup that received a booster after 30 days (WMD -11.40%; 95% CI: -18.31 to -4.49; p = 0.001).No difference was demonstrated between the tailored and non-tailored infor-Graph 1. Participants with personal tailored information versus participants with non-tailored (general) information.mation groups in any of the other subgroups in Lusk et al. 16 (booster after 90 days; booster after 30 and 90 days; no booster).
Boosters: All participants with Boosters: All participants with tailored information versus all tailored information versus all control participants control participants In relation to the outcome "Mean percentage of time actually using hearing protection device in required areas, over the preceding week and month", two particular interventions seemed to favor the participants included in the tailored group over those included in the control group: booster after 30 days (WMD -10.30%; 95% CI: -18.09 to -2.51; p = 0.01) and boosters after 30 and 90 days (WMD -9.50%; 95% CI: -17.86 to -1.14; p = 0.03), as shown in Graph 5.

Multicomponent educational Multicomponent educational intervention versus control intervention versus control (baseline hearing tests and (baseline hearing tests and follow-up tests) follow-up tests)
Knobloch and Broste 25 measured the percentage of students who used hearing protection devices "at least sometimes".After three years of the study, it was possible to detect a statistical difference favoring the participants in the intervention group over those in the control group (relative risk (RR) 0.42; 95% CI: 0.36 to 0.49; p < 0.00001) (intention-to-treat analysis), as shown in Graph 6.The difference was maintained after four years (RR 0.51; 95% CI: 0.45 to 0.58; p < 0.00001) (completer analysis), as shown in Graph 7. Results from a more rigorous analysis using the intention-to-treat approach revealed that the hearing conservation program trialed by these researchers was effective for both time points, i.e. after three and four years (p < 0.0001).At the start of the study, only 23% of the intervention group and 24% of the control group wore hearing protection "at least sometimes".After three years, this had increased to 83% in the intervention group and 35% in the control group: an absolute difference of 48%.
Graph 5.All participants with tailored information versus all control participants.Graph 6. Multicomponent educational intervention versus control (intention-to-treat analysis).

DISCUSSION
This systematic review offers up-todate but limited evidence supported by randomized controlled trials, regarding the effectiveness of interventions to promote the wearing of hearing protection devices.Two of the studies 15,16 included used the same participants but studied different interventions that both yielded a negative result among workers who had already reported substantial use of hearing protection.The methodological quality of the studies was reasonable, even though there was a substantial risk of attrition bias due to the loss from follow-up of half of the workers.This could imply that only those who were interested and motivated turned up for the post-intervention measurement.Moreover, it could mean that in a highly motivated group of workers it is diffi cult to increase the percentage of use of hearing protection.However, the risk of hearing loss increases exponentially with the amount of time for which protection is not worn.This means that there is a need to develop interventions that are capable of motivating workers in such a way that the percentage of use will also cover the remaining 20%.Applying the intervention to the whole group, including those that already show perfect behavior in wearing hearing protection, does not seem to be very effi cient.It might be worthwhile to explore further the possibilities of targeting the interventions at those that do not use hearing protection.
We found one study that applied the intervention to schoolchildren who were exposed to noise through their work at their parents' farms. 25The methodological quality was reasonable and the number of clusters and the number of participants were suffi cient to demonstrate a signifi cant difference.The outcome measurement used in that study (the percentage of participants who used hearing protection devices "at least sometimes") is diffi cult to interpret in the light of the axiom that hearing protection should be worn for 100% of the time a person is exposed to hazardous noise levels.However, it provides evidence that a school-based program maintained for several years can substantially increase the use of hearing protection, in comparison with a minimal intervention control group.In the United States, it has been suggested that there is substantial room for incorporating occupational health and safety information in vocational training schools. 29Furthermore, in occupational health and school settings, cluster randomization seems to be a reasonable approach in designing randomized controlled trials.However, methods to allow for intracluster correlation must be used in the analysis of such studies.
There are many qualitative studies that have examined the reasons why workers do not use hearing protection.Comfort 2,30-31 and the level of enforcement of the requirement to wear hearing protection 11,13,30 have been reported as factors in not using hearing protection.Several models have been used to explain the variation in the use of hearing protection, such as the health promotion model and the protection-motivation theory. 11,12esearchers should make use of these studies to develop new interventions that might be more effective.
There are several studies showing that instructions on how to use earplugs are needed in order to properly insert them.In non-randomized studies that compared the noise attenuation of earplugs used with and without receiving instructions, a signifi cant and important reduction in noise reduction between the groups was found. 22,24here is limited evidence that longterm school-based programs can effectively increase the use of hearing protection among students at vocational schools.To date, the limited evidence available does not demonstrate that the use of personalized information to motivate workers to use hearing protection is better than more general information.The limited evidence also does not show that the use of reminders after the intervention increases the use of hearing protection.It could be that this is due to a ceiling effect and only applies in situations in which the majority of workers already use hearing protection.
There are only a few good quality studies evaluating the effectiveness of interventions to promote the wearing of hearing protection devices.More randomized controlled trials are needed.To avoid the risk of contamination, cluster randomized trials are to be preferred.Proper adjustments should be made for the cluster effect and intracluster correlation coeffi cients should be reported.Future trials should have standardized outcome measurements such the endpoint proportion of participants who wear hearing protective devices in relation to the endpoint proportion in the control group, intention to use the devices, perceived benefi ts and barriers to the use of hearing protection (self-reported use of hearing protection).Dropouts and losses from follow up should be reported.

Graph 2 .Graph 3 .Graph 4 .
Participants with personal tailored information versus control (commercial videotape) group.Participants with non-tailored (general) information versus control (commercial videotape) group.All participants with tailored information versus all participants with non-tailored information.