**4. Discussion**

The goal of this study was to evaluate the efficacy of a construction worksite-based integrated intervention targeting both the conditions of work, and workers' health behaviors, simultaneously. We observed short-term improvements in ergonomic practices and in incidences of pain and injury after an injury prevention program. We also observed an improvement in physical activity and healthier dietary behaviors, such as increased consumption of fruits and vegetables, after a health promotion Health Week program.

At the individual level, we found a significant improvement in ergonomic practices, and a reduction in incidences of pain and injury, which supported the hypothesized pathways for the program. As promoted by NIOSH, ergonomic practices focus on workers modifying or establishing work procedures to reduce the risk of injuries [39,49]. While we did not quantify exposure to specific ergonomic hazards, the StIPP focused on workers' setting up their work more ergonomically. For example, working at knuckle level instead of on the ground, and using appropriate tools to reduce extreme postures associated with overhead work and manual materials handling. The program targeted the conditions of work directly controlled by the workers themselves (Figure 2) [7]. Giving such control to workers is important in reducing disability, as it gives workers opportunities to adapt their work in order to better manage their own musculoskeletal symptoms and health [50,51].

While we were encouraged that an improvement in ergonomics practices occurred, results also indicated that the program was not successful at addressing system level components. For example, while we saw ergonomics practices improve, we saw no significant change in the physical demands on the workers. Hence, we suspect that the intervention changed the way people completed their work but had limited effect on the physical demands of the job. In addition, the process evaluation revealed several important barriers and facilitators to our program at the organizational level. First, management and worker buy-in were identified to be integral to the success of the soft tissue injury prevention program. This was perceived to be key in a work environment that is fast-paced, unpredictable, and with tight production schedules tied to the requirements of the general contractor. For example, there was little time to complete task preplanning or for the safety manager to complete inspection protocols for injury hazards and ergonomic solutions. When management support for health and safety programs is not observed by the workers, other competing factors are often prioritized over health and safety, especially ergonomic practices [52]. This was quite evident on one site which had major delays due to the winter storms of 2015 in the Boston metropolitan area. Due to the loss of almost a month of production, competing safety and schedule priorities would supersede program delivery. Similar challenges to program delivery have been reported by others in the construction industry [53].

While we have had past success with a worksite safety program integrated within the complex structure associated with multiple employers, a large barrier to a systems approach ergonomics program was the challenge faced by subcontractors to make changes, even those changes that could improve site safety on their own worksite. Unlike our previous program that was designed to re-enforce existing safety practices [36], the ARM program required subcontractors to implement new, or modify existing practices and tools, that may be specific to their trade. Our program focused on simple ergonomics solutions that individual workers could implement to their own work [39]. However, more complex or system-level changes would require the involvement of multiple groups or stakeholders [54]. Ergonomics solutions in a fissured workplace require all site employers to take on elements of the program to effectively and systematically influence the overall conditions of work [17].

Moreover, system-level changes require better upfront planning before construction begins, such as during the bidding process for a job by setting out requirements from the multiple employers, and in the contracts for the jobs. The key informant interviews supported this concept. Expectations regarding safety programs in the contract is standard procedure in larger projects, especially owner-insured programs. An example is with respect to safety training, in which owners, especially public entities, require that construction workers have a minimum of OSHA-10 training to be onsite [55]. Whilst others require their contracting companies complete safety prequalification safety surveys, or have written safety management programs. Thus, including ergonomics in the contractual language may set up better expectations for a program.

Other researchers in the construction industry have also found mixed findings with respect to improvements in pain and injury and perceived physical effort after implementing ergonomic interventions, including participatory ergonomics programs [33,56–58]. In these studies, reasons for intervention failure were generally associated with the intervention not being delivered as intended or implemented at all of the sites [28,57]. In our study, intervention delivery occurred as per the protocol in four of the five sites during the intervention period. However, since the ergonomics program stopped after six weeks and workers often moved from sites before the follow-up data collection was completed, we also attributed this to our loss of significance at the six-month follow-up. Although we observed that on average the reduction in pain and injury incidences, and improvements in ergonomic practices were maintained, there was reduction in power due to loss to follow up.

In contrast, the Health Week had many successes in overcoming some of these barriers associated with the multiemployer structure. For one, it simply required the participation of the workers and little, if any, infrastructure. In theory, the Health Week might have addressed some conditions of work regarding psychosocial factors around health, like supervisor or co-worker support. Anecdotally, we observed foremen and co-workers being supportive of ensuring their co-workers signed up for health coaching or NRT. Some foremen would cover for their workers to allow them to participate in the week's activities. We also observed workers talking about eating healthier food with their coworkers during the week.

Another major strength of the Health Week was how it aligned with companies' current practices on the worksite and also with the interests and goals of the workers. This was also found in previous formative work we completed that found that policies, programs and practices are supported by management and workers alike if they can be easily integrated into company's business structures and align with workers' goals and needs [59,60]. Health Week was in a familiar format for the workers and companies alike. We modelled Health Week after the industry's standard practice of safety week where contractors have a specific theme and perform a series of outreach activities for workers to provide information on resources and best practices. Thus, due to the familiar format, workers may have been more receptive to the daily topics. Although the uptake on NRT and individualized health coaching was low, we did see improvements in workers' health behaviors, including higher intake of fruits and vegetables, and increased amount of time per week engaged in recreational physical activity. This finding is similar to the results of a health promotion intervention conducted in the Netherlands, which found that onsite group coaching sessions resulted in changes in physical activity, and dietary behaviors, but did not improve musculoskeletal symptoms [61].
