High noise levels are ubiquitous in the construction industry. As a result, although it is completely preventable, noise induced hearing loss (NIHL) is very common among construction workers. Factors such as intermittent and highly variable noise levels, lax regulatory enforcement, lack of awareness and knowledge of harmful noise levels, attitudes and beliefs towards using hearing protection devices (HPDs) and the improper use of HPDs contribute to construction workers' NIHL. To evaluate the effectiveness of different strategies to increase use of HPDs among construction workers, we designed and implemented a large-scale four year long study (July 2005 - May 2009) across eight construction sites in the Puget Sound area. The aims of this study were to assess the effectiveness of a three-component intervention in motivating construction workers to effectively wear HPDs during periods of high noise exposure. HPD use was captured via two self-reported instruments, a questionnaire and a validated combination of task card (including HPD use) with simultaneous dosimetry measurements. All study methods and protocols were approved by the University of Washington Institutional Review Board (IRB). Study design and development: The study consisted of three interventions: 1. A one-hour long hearing conservation training, 2. Use of an innovative new noise level indicator (NLI), and 3. Follow-up reinforcement tool-box trainings (TB). Hearing Conservation Intervention: The content of the one-hour long hearing conservation training was intended to address common barriers to HPD use, and was guided by a modified version of Pender's Health Promotion Model (HPM). The HPM has previously been shown to predict construction workers' use of hearing protection (Lusk et al, 1997). Training concepts were specifically matched to the constructs of the modified HPM along with meeting OSHA training requirements. Eight constructs in total were covered in the training and incorporated five belief factors (perceived susceptibility, severity, benefits, barriers, and self-efficacy), two HPD use cofactors (interpersonal norms and situational influences), and one knowledge factor. The training incorporated two "hands-on" demonstrations, including instruction on the selection and proper use of HPDs and teach-back elements for trainees to respond to questions and real-life experiences. For maximum flexibility, the training was designed to be delivered as either a one-time 1 hour session or two 30-minute segments, and used flip chart visual materials for logistical ease. Additionally, the design of the training delivery was intended for peer trainers, or trainers trained by an expert, to be able to give the training as effectively as an expert. Development and Testing of Training Program We conducted a pilot test of the hearing conservation training was done on one construction site with 23 workers (Neitzel et al, 2008). HPD use increased significantly post-training, with the percent of time that HPD worn in noise levels above 85 dBA nearly doubling. However, there were virtually no statistically significant differences seen in survey responses in relation to the survey items. The results of this pilot demonstrated the potential short-term effectiveness of the training program. We also tested the flexibility of the training delivery by comparing the effectiveness of the training when delivered by an expert versus a trainer who had been trained in a "train-the trainer" program (Trabeau et al, 2008). One-hundred and three workers were trained by four trainers who had received the "Train-the-trainer" training (n=56 workers), and one expert trainer (n=45 workers). Effectiveness was assessed by comparing pre- and post-training changes in beliefs regarding HPD use, knowledge of noise, and self-reported HPD use in high noise. There were no statistically significant differences found between the expert and the train-the-trainer groups except for two belief items, one pertaining to susceptibility and the other pertaining to self-efficacy. As a result, we believe the training program could be used in a 'train-the-trainer' mode, assuming the trainers receive appropriate guidance and instruction on the use of the materials. The images used for the training are available at: Noise Level Indicator The NLI used in the study was a prototype, smaller than a cell phone, which provided real-time noise levels to construction workers by flashing different colored lights and vibrating at noise levels at or above 85 dBA. The unit was designed to be worn on workers' lapel or fall protection harness for maximum visibility. Toolbox Trainings Additionally, Toolbox (TB) trainings were given every other week for two months (four TB trainings total per worker). Each of the four TB trainings was 5 minutes long and covered a single key topic from the hearing conservation training (see example in Figure 3). Intervention activities on each site started with the one-time hearing conservation training but lasted two months with the additional NLI intervention(for some subjects) and TB interventions (for some sites). Intervention evaluation measures Evaluation components of the interventions included a survey given at four different time periods, the assessment of worker's shift noise exposure using data-logging dosimeters, and task cards reporting HPD use during the shift in which noise levels were measured at three different time periods. The survey had sections covering the beliefs associated with the HPM model constructs listed above, as well as current and intended use of HPDs. The percent of time during a shift which equivalent continuous sound pressure levels (e.g., Leq levels, measured according to the NIOSH Recommended Exposure Limit, or REL, for noise) exceeded 85 dBA was calculated using the dosimetry data and matched with when a worker reported wearing HPDs from the associated task card. Prior to intervention activities (e.g., the baseline time period), study participants received a survey and one exposure assessment with a matching task card indicating HPD use. Immediately following the training, participants completed another survey. Once all of the intervention activities were completed (two months after training), all participants regardless of the intervention group completed another survey and an exposure assessment with a matching task card indicating HPD use. Finally, a follow-up exposure assessment with activity card and a survey occurred four months after training. Site and participant allocation to intervention groups All participants across the eight participating construction sites received the one-hour hearing conservation training. Four of the sites received no additional training while the other four sites received the bi-weekly TB training. Workers from all eight sites were randomized into two groups, those who received NLI and those who did not. In total, there were four different intervention groups: those who received the training only, those who received the training and NLI, those who received the training and TB, and those who received all three, training, NLI and TB. All participants provided informed consent with a form approved by the University of Washington IRB prior to enrolling in the study. Results and key findings: Participants, Two hundred seventy-one construction workers (roughly thirty to forty construction workers per site) were enrolled into the study. The trades represented in the study were carpenters, laborers, electricians, iron workers, operating engineers, sheet metal workers, plumbers, pipefitters and cement masons. Ninety-seven percent of the subjects were male and on average were 37.7 years old. Sixty-three percent of the subjects had a high school education level and 30% had more than a high school education. The most common trade group was carpenters which made up 43% of the subjects, with laborers (17%) and electricians (11%) being the other largest represented trades. A few differences were seen in the demographic data between the four intervention groups with TB sites reporting slightly better hearing and less years working in construction. Noise exposures, Our study documented exposures among the participating construction workers which routinely exceeded the NIOSH REL of 85 dBA. The mean full-shift Leq level prior to intervention activities was 89 +/- 5 dBA, with an average of 40 +/- 20% of time in each shift spent above 85 dBA. Average noise levels dropped slightly at the two and four month follow-up, though both were still above 85 dBA (88 +/- 5 dBA and 87 +/- 5 dBA, respectively, and about 33% of each shift spent over 85 dBA on average). HPD use, We found that self-reported HPD use was found to be more accurate from task cards than from surveys; HPD use was typically overstated via survey (Trabeau et al, 2008; Edelson et al, 2009). As a result, we used HPD use from the task card method as the primary outcome measure for the intervention. A clear bi-modal distribution of HPD use was seen from the task card data. At baseline, eighty-one percent of subjects reported using HPDs either almost always (>90% of time >85 dBA) or almost never (<10% of time >85 dBA). We have restricted the HPD use data analyses presented here to the 176 construction workers who completed exposure assessments in all three time periods of the study (at baseline, at the end of the two-month intervention period, and at a 4-month follow-up). Table 1 shows HPD use at the three time periods. Overall, the percent of time HPDs were used in high noise increased from 35% pre-intervention to almost 50% at the end of the 2-month intervention period. At the four month follow-up, HPD use was just over 40%. A statistically significant increase in HPD during both intervention follow-up periods, compared to pre-intervention levels, were observed.