Managing Psychosocial Health Risks in the Australian Construction Industry: A Holistic Hazard Management Intervention

Abstract

This study presents a case study of a holistic, psychosocial hazard management intervention program in a project-based, remote workforce in the Australian construction industry. There is a dearth of research on targeted, integrative, multi-level wellbeing interventions, and we seek to address this gap. Given the high rates of psychological distress and suicide in construction, understanding these hazards and the responses needed to manage them is critical for prevention. Data were collected from workers before and after the implementation of an intervention using an empirically validated measure of the work environment underpinned by the job demands–resources framework to evaluate exposure to psychosocial hazards, and mental health indicators, including psychological distress and suicidal ideation. Results revealed that job demands and resources improved following the change initiative, and workers reported significantly lower levels of psychological distress compared to workers on similarly diverse remote sites. The findings highlight the need for targeted mental health interventions addressing specific workplace psychosocial risks that adopt a holistic approach to change across all levels of an organisation. The study contributes to a nuanced understanding of psychosocial risks in construction and informs strategies to mitigate mental health harms in high-risk occupational settings.

1. Introduction

The construction industry is vital to Australia’s economy, employing almost ten percent of the national workforce and contributing seven percent of Gross Domestic Product [1]. The sector is widely recognised internationally as high-risk for both physical and psychological harm, with poor working conditions and elevated rates of mental ill-health and suicide among its workforce [2,3,4]. Remote and regional projects, particularly those reliant on Fly-In-Fly-Out or Drive-In-Drive-Out (hereafter, FIFO/DIDO) workforces, present additional psychosocial risks due to extended periods away from home, social isolation, and highly demanding work environments [5,6]. Controlling and managing these risks is not only critical for worker wellbeing but also organisational safety, productivity, and sustainability.

In response to these challenges, there are growing calls for examples of structured, evidence-based approaches to assess and mitigate psychosocial hazards in construction settings [7]. Despite these appeals, interventions and, consequently, published research usually suffer a fragmented approach; focus on individual-level change with disregard for barriers inherent to the work environment and culture; and fail to sustain management and leadership support [8,9]. This article addresses such shortcomings, presenting a case study of a holistic, multilevel, industry-led and research-driven employee mental health intervention. The intervention was guided by the job demands–resources (JD-R) model [10], which offers a robust theoretical basis for such interventions, proposing high job demands can negatively affect employee mental health and wellbeing, while sufficient job resources can buffer these effects [10,11].

This case study focuses on a major construction site—Rookwood Weir—with a project timeline of over 2 years in duration and a FIFO/DIDO workforce in regional Australia. The case organisation partnered with MATES in Construction (hereafter, MATES), an industry-led and supported, non-profit organisation focusing on suicide education and prevention within the construction industry. MATES was established in 2008 in response to a report identifying higher rates of suicide by males in the Queensland commercial construction industry compared to the general male population, and 2.38 times higher rates of youth suicide within the industry [12]. This article details the evolution of a partnership between these organisations, followed by the staged approach to improvements and how they were driven, evaluated and monitored. Drawing on the JD-R framework, the study analyses the impact of psychosocial hazards on the workforce pre- and post- intervention and, for mental health outcomes, benchmarks the case site against 11 comparable FIFO/DIDO sites at the completion of the intervention. The hypotheses predict that, following the intervention, perceived job demands will reduce and job resources for workers will increase, and psychological distress and suicidal ideation will be reduced compared to similarly diverse FIFO/DIDO sites.

The article is structured as follows. First, the literature review examines psychosocial challenges facing the construction workforce, which stem from the types of work performed, the working conditions, and the masculine culture typical of this sector [13]. The impact of these hazards on employee wellbeing and the risk management framework are detailed, highlighting the need for more research into holistic and integrative workplace interventions, whereby the study makes its key contribution. The JD-R model underpinning our research is explored as the mechanism to explain the relationship between psychosocial hazards, job demands and resources, and outcomes for employee mental health. Finally, the article outlines the case organisation, change intervention and partnership, followed by results from before the initiative (Time 1 and 2), after the initiative (Time 3 and 4), and compared to similarly diverse cases.

2. Psychosocial Hazards in the Construction Industry

Psychosocial hazards are defined as any aspect of the design or management of work that increases the risk of work-related stress, while risk is the likelihood of harm occurring from psychosocial hazards [14]. Unlike physical hazards, which are often visible and well understood, psychosocial hazards have historically been overlooked across all industries, given the perception they are more complex and difficult to identify and manage [14,15]. The impact of psychosocial hazards on employee wellbeing, however, is a serious health and safety issue, which in recent years has begun receiving due consideration through increased academic research, and amendments to relevant workplace legislation [7].

While harm from psychosocial factors poses a challenge in all workplaces, some industries are exposed to a larger number of hazards at greater levels of risk than other industries. The construction industry, particularly remote building sites, are often fraught with challenges such as compressed timelines, financial pressure and poor communication [16]. The psychosocial hazards of the industry can be considered as pertaining to either the nature of the work itself and the physical environment in which the work is carried out; the employment conditions of the organisation or project; and the industry culture and norms inherent in the workplace.

The nature of work in the construction industry is physically demanding, with much of the workforce engaged in manual labour and facing exposure to various hazards such as extreme weather, high noise levels, chemical and dust exposure, and operation of heavy equipment [2]. Physical characteristics of work may be a source of acute discomfort, stress and physical injury; however, they also overlap with hazards in the employment environment. Illness and injury caused by physical or mental health conditions can affect productivity and lead to absenteeism or reduced likelihood of future employment opportunities, particularly in short-term project work. The temporary nature of work projects means many construction workers are casually employed without access to paid sick leave, making recovery harder and contributing to a lack of job security [17]. Job uncertainty is a work-related stressor for all industries but is unique in its impact on construction workers, particularly in FIFO/DIDO work [16,18].

Compounding this problem, in the Australian construction industry, the culture is heavily shaped by a workforce comprising 87% males [19]. In male-dominated industries, standards of traditional masculine norms can lead to greater stigma around discussing mental health and, consequently, less help-seeking behaviour [20]. Bullying and harsh initiation practices are typical in the industry, particularly affecting young workers and apprentices [21]. Long working hours reduce workers’ energy and impact their ability to engage in healthy behaviours outside of work, such as exercising and eating nutritious food [2]. This unhealthy lifestyle is exacerbated by the tendency for excessive alcohol consumption, smoking and illicit drug taking behaviours normalised in the masculine industry culture [22].

2.1. The Burden of Poor Employee Wellbeing

The combination of work and employment hazards, combined with a challenging culture, creates multiple strains and risk factors for poor physical and mental health. The physical health consequences of such stressors have been linked to a range of negative health outcomes including cardiovascular disease [23], immune system dysfunction [24], and higher incidence of diabetes [25], among other health issues. Workers facing psychosocial pressures are also more likely to adopt maladaptive coping mechanisms such as excessive alcohol consumption and use of illicit drugs, behaviours consistently identified at much higher prevalence in the construction industry [9,26].

Most concerning, however, is the association between psychosocial hazards and suicide, a leading cause of death in the Australian construction industry [27]. The rate of male suicide in the industry is estimated to be over 25% higher in construction than in other industries [28]. Additionally, estimates indicate for every death by suicide, 20 other individuals attempt suicide [29]. Characteristics of work considered to increase suicide risk in construction include bullying, long working hours, job uncertainty limited job control, workplace culture and self-stigma [20,21].

Beyond individual outcomes, psychosocial hazards in the workplace have broader implications for businesses and communities. Organisations face increased costs through reduced productivity, higher absenteeism, presenteeism, staff turnover, and workers’ compensation claims [9]. At the community level, the burden of work-related mental health issues contributes to greater demand on healthcare services, social support systems, and public health resources, while also negatively affecting the wellbeing of workers’ families and social networks [30,31]. The cost of mental illness and suicide attributable to work related psychosocial hazards in Australia has been estimated at $30 billion in lost productivity annually [32]. Addressing psychosocial risks in the workplace is therefore critical not only for individual health but also for sustaining healthy, productive businesses and resilient communities.

2.2. Managing Psychosocial Hazards in the Workplace

In countries like Australia, legislation for managing psychological hazards in the workplace relies on a four-stage risk management approach [14]. Managers are required to first identify hazards, assess the likelihood of harm, control the risk (through elimination or minimisation of risks) and finally, evaluate the effectiveness of the implemented control measures. While tools exist for identifying psychosocial hazards and assessing their associated risk [33], addressing the third stage of the risk management cycle—using the hierarchy of controls to determine the most management options—is an aspect of the model receiving limited academic discussion [16].

Following the identification and assessment of psychosocial hazards, management has a duty to eliminate risks as far as is reasonably practicable [14]. Determination of what is reasonably practicable encompasses considerations around the likelihood of occurrence and degree of harm that may result from the hazard; what is or should be known about the hazard and its elimination; the availability of suitable hazard management approaches; and the associated costs and whether they are proportionate to the risk [34]. With the fast pace of research on workplace wellbeing, determination of the degree of risk will change as new information becomes available, and as new risks are discovered and understood [14]. This context makes the control of hazards an enduring and constantly evolving endpoint [34].

Despite the importance of understanding the process of controlling psychosocial hazards in the workplace, this body of literature remains piecemeal. Studies in the construction industry tend to focus on the identification of psychosocial risks, prevalence of mental health conditions or suicide, and characteristics that increase workforce vulnerability [6,35]. Research on interventions to address psychosocial hazards, is often fragmented, targeting individual behaviour change or awareness in isolation from the broader organisational environment, culture, and workplace barriers [8], or without sustained integration of upper and frontline leadership [9]. There has been a call for more studies examining sustainable, embedded interventions, with evidence of measurable impact on psychosocial conditions and mental health outcomes [8,31].

A small and emerging body of literature has highlighted the importance of integrated and multilevel approaches to workplace mental health interventions, which address a broad spectrum of employee needs rather than focusing on isolated outcomes. For instance, LaMontagne et al. [9] proposed that effective interventions should simultaneously prevent harm via reduction in psychosocial risk factors, develop the positive aspects of work including building on worker strengths, and manage mental ill-health in the workplace regardless of its cause. Building on this proposition, Petrie et al. [36] outline a five-pronged approach to integrated mental health promotion, that focuses on enhancing organisational and personal resilience, facilitating help-seeking behaviour, and promoting recovery and return to work.

Such “integrated” or holistic interventions must operate across multiple levels, targeting individuals (e.g., coping and health behaviours), as well as the broader organisational environment (e.g., leadership, culture and policy reform) [11,18]. Despite their conceptual strength, multilevel interventions can present challenges in implementation and evaluation and are, therefore, lacking in extant research. This article addresses that gap by presenting a case study of a holistic, multilevel, industry-led and research-driven intervention, designed to improve psychosocial conditions and mental wellbeing at a major Australian construction site.

2.3. Misalignment Between Job Demands and Resources

The mechanism by which psychosocial hazards impact worker wellbeing is commonly explained using the job demands–resource model [10]. The JD-R model proposes psychological features of work can be categorised into two dimensions: job demands and job resources. Job demands refer to components of work requiring sustained effort or energy expenditure, leading to potential physiological or psychological costs for employees. In contrast, job resources encompass factors that facilitate goal attainment, mitigate demands, or foster learning, growth, and development. An underlying assumption of the JD-R framework is that both demands and resources influence worker wellbeing. Negative outcomes can result from differences between demands an individual has placed on them at work and the resources they have available to complete their job [37,38].

The construction industry is defined by high job demands and low job resources [7,13,39]. Exposure to high levels of job demands—such as work overload, long hours, role ambiguity, and conflict—depletes workers’ physical and emotional energy, leading to burnout and poor mental health [40]. Reducing excessive job demands has been associated with lower levels of psychological distress and improved employee wellbeing across a range of construction and industrial settings [6,9]. Interventions aiming to enhance leadership and cultural support, and provide greater awareness about the impact of psychosocial hazards on mental health, can enhance “upstream resources” that positively impact workers’ perceived or actual job demands and resources. Furthermore, enhancing support and awareness at an organisational-level can mitigate the adverse impact of high demands and low resources on mental health outcomes. This aligns with research on the extended JD-R model [41,42]. As detailed in the methods, this case study outlines an intervention designed to address workers’ experience of high, or increasing, job demands and low job resources. Based on extant research on the JD-R framework, it is hypothesised:

Hypothesis 1. 

Following the implementation of a holistic psychosocial risk management intervention, perceived job demands (e.g., role conflict and role overload) will significantly reduce.

Insufficient job resources, such as job control, co-worker support, or supervisor recognition, in the presence of high job demands lead to mental ill-health, low morale, and poor performance [2,11]. Research applying JD-R theory has indicated role conflict and ambiguity exhibit the strongest association with mental health problems [7]. There is also growing recognition of the importance of safety competencies and attitudes of leadership as a job resource that can improve worker safety behaviours [4,43]. Supervisor support, change consultation, and procedural fairness have also been identified as key resources that help buffer high job demands and protect against psychological strain [11]. Accordingly, it is proposed:

Hypothesis 2. 

Following the implementation of a holistic psychosocial risk management intervention, perceived job resources (e.g., supervisor support, job control, praise and recognition) will significantly increase.

JD-R theory also predicts that improvements in the balance between job demands and resources will result in improvements to employee wellbeing outcomes, such as reduced psychological distress and increased engagement [10,11]. In high-risk industries like construction, elevated job demands, and low job resources have been associated with increased rates of mental ill-health, including suicidal ideation [2,18]. Interventions that address psychosocial hazards by improving working conditions and strengthening social support have been shown to reduce suicide risk and psychological distress among construction workers [18]. This led to the third hypothesis:

Hypothesis 3. 

Workers’ psychological distress and suicidal ideation will be lower compared to workers at comparable sites when an organisation has implemented a holistic psychosocial risk management intervention.

The hypotheses are guided by JD-R theory, as this theoretical perspective provides the most integrative capacity to consider both the negative influences on employee experiences and factors that can have a positive effect on workers. When the inherent demands of the construction industry are considered, it becomes increasingly important for organisations to find ways to both mitigate those demands but also provide well considered and structured resources for employees. Over time, JD-R has enabled accommodation of a wide variety of job characteristics [37], allowing a more comprehensive understanding of both worker wellbeing and performance within the context of safe working [44]. Furthermore, JD-R facilitates exploration of direct associations between various demands and resources on employee health and wellbeing [45].

The following section outlines the case study of Rookwood Weir, a major construction site operating between November 2021 and August 2023 in regional Queensland. The methods describe the case study context, and the role of the industry-driven support agency MATES that assisted in the delivery of the targeted psychosocial risk management intervention. Consideration is given to the role of site leaders and the third-party support organisation that initiated implementation changes on the ground. Survey results are presented from two timepoints before the intervention, and two points following the intervention. Employee mental health outcomes after the intervention are then compared with eleven additional FIFO/DIDO sites.

3. Materials and Methods

3.1. Methodology Overview

The case study data was obtained via four repeated cross-sectional survey waves across the project build with 538 survey responses in total: Time 1 (N = 138)—November 2021 and Time 2 (N = 118)—August 2022 (prior to the substantive rollout of initiatives), and Time 3 (N = 176)—March 2023 and Time 4 (N = 106)—August 2023 (post the rollout of initiatives). The study timeframe (November 2021–August 2023) reflects the period of active engagement between the Rookwood Weir Alliance and MATES (Queensland & Northern Territory division), during which on-site access and survey administration were feasible.

The sample population was made up of site-based workers (direct employees and subcontractors over the age of 18) engaged on the Rookwood Weir project during each survey window. Due to workforce churn inherent to project work, cohorts differed and were treated as independent groups for analysis. Surveys were paper-based and administered during toolbox/pre-start meetings to all workers on shift; participation was voluntary and anonymous. Psychosocial hazards (People at Work—Construction [PAW-CON]) were measured at all four timepoints. Mental health outcomes (General Health Questionnaire [GHQ-12]; two suicidal ideation items adapted from the Suicidal Behaviors Questionnaire-Revised [SBQ-R]) were added at Time 4 following an instrument revision. We evaluated change across phases with Multivariate Analyses of Variance (MANOVAs) and benchmarked Time 4 against 11 comparable FIFO/DIDO sites (n = 1286) collected July 2023–March 2024 that engaged with MATES during normal operational activities. MANOVAs were conducted using listwise deletion for missing values. Consequently, analytic sample sizes vary by model and may be smaller than the total wave counts. The N used in each analysis is reported in the corresponding table captions. Wave-specific sample and procedural details are described in Section 4.1, Section 4.2 and Section 4.3, adjacent to the corresponding analyses.

3.2. Rookwood Weir Case Context

The Rookwood Weir represents a significant milestone in Australian water infrastructure, being the largest weir constructed in the country in over 70 years and the most substantial addition to water infrastructure in more than a decade. Originating from the Lower Fitzroy River Infrastructure Project, the initiative aimed to tackle emerging water supply challenges in the Rockhampton and Gladstone regions. Failure to proceed with the project would have resulted in limited growth opportunities for high-value agriculture, ongoing water security concerns for Rockhampton’s water users and Gladstone Area Water Board and potential hindrances to industrial development, posing adverse effects on the Queensland economy.

To realise the project’s objectives, the Rookwood Weir Alliance, comprising Sunwater, construction partners ACCIONA and McCosker Contracting, and design partner GHD, was established. The endeavour received a total funding allocation of AUD 568.9 million, with contributions from both the Queensland and Australian Governments amounting to AUD 183.6 million, while Sunwater financed the remaining portion [46]. Positioned on the Fitzroy River, west of Rockhampton, the weir spans 460 m in length and sits 17 m above the riverbed.

Construction commenced in late 2020, culminating in the project’s completion in November 2023. The endeavour involved more than 2.189 million work hours dedicated to building the weir and associated infrastructure projects [47]. The project relied on a FIFO/DIDO workforce. At its peak, the total workforce operating across Rookwood Weir reached 356, with a total of 384 workers across all supporting projects.

Over the duration of the project, the workforce faced difficult conditions. The site experienced flooding three times over the course of three years and night operations took place for almost a full year [48]. The site was also locked down for two weeks during the COVID-19 pandemic. These conditions, as well as the remote location of the site, resulted in unique psychosocial risks for the workforce including increased job pressure, sleep disruptions, and psychological isolation.

3.3. MATES in Construction Overview

To mitigate these risks, the health and safety team at Rookwood Weir partnered with MATES in 2021 to begin a process of identifying, assessing, controlling and reviewing the effectiveness of control measures designed to reduce psychosocial hazards in the workplace. MATES offers an integrated industry intervention program that raises awareness of suicide as a preventable problem, builds stronger and more resilient workers, connects workers to the most suitable available help, and supports and partners with researchers to inform industry on best practice [18]. MATES has partnered with around 4000 worksites to provide non-clinical case management, outreach services, mental health training, and an all-hours support service for industry employees for more detail see [49]. The MATES Program Logic includes various short-term, medium-term, and long-term outcomes, including the reduction of psychological distress and suicide in the construction industry in 5–10 years [18]. Peer-reviewed evaluations of the MATES program report positive impacts on suicide prevention literacy, help-offering intentions and related outcomes among construction workers [18].

An effective partnership between Rookwood Weir and MATES hinged on the joint dedication of site leaders and MATES field officers. Leaders at Rookwood Weir proactively engaged with MATES throughout the project, drawing from prior positive experiences with the organisation within the industry. Their advocacy expedited the completion of The Blueprint for Better Mental Health and Suicide Prevention (hereafter, The Blueprint). The Blueprint is the overarching framework used by the construction industry to gauge the effectiveness and need for workplace mental health and suicide prevention initiatives. The audit tool was used to assess the organisational-level policies and practices for addressing mental health and suicide prevention, identifying potential gaps that could contribute to psychosocial hazards on site. Analysis of the survey highlighted areas where improvements could be made, prompting targeted initiatives from the Rookwood leadership team and MATES. Working in tandem, the Health and Safety Team at Rookwood and MATES field officers developed a strategy to foster positive mental wellbeing among the workforce throughout the project’s duration.

Initially, six leaders from different project sectors underwent Applied Suicide Intervention Skills Training (ASIST) to equip them with the necessary skills to intervene if a worker showed signs of suicidal thoughts. These leaders, including positions such as Health and Safety Manager, Construction Manager, Leading Hand, and Peggy (administration), were strategically chosen to ensure diverse coverage across the project. By having ASIST-trained individuals in various roles, workers across different trades would have someone familiar to turn to for support, and awareness of the training could permeate various departments. Prioritising ASIST training ensured that workers had access to support pre-emptively, before sensitive topics like mental health and suicide were addressed in toolbox talks and training sessions.

Next, MATES conducted an introductory toolbox talk for approximately 100 workers, addressing the importance of mental health awareness and offering practical guidance on how to assist those in need. Additionally, the Health and Safety Manager promptly arranged for MATES to deliver General Awareness Training (GAT) on-site, providing workers with a deeper understanding of mental health issues and empowering them to initiate conversations about it. Following the GAT, several workers volunteered to become “Connectors”, trained to support individuals in crisis while connecting them with professional help.

Throughout the project’s duration, periodic evaluations of psychosocial hazards on site were conducted using the PAW-CON survey [6]. The survey was designed to investigate psychosocial hazards in the work environment with a specific focus on the building and construction industry. PAW-CON is a construction-specific adaptation of Australia’s validated People at Work survey [50]. Item generation was informed by data and research by the Australian Government, Beyond Blue (a national mental health organisation), and MATES. Development of the tool was funded and supported by MATES (more detail on the instrument is provided in the Quantitative Results).

The PAW-CON surveys were conducted across four time periods to assess shifts in workers’ perceptions of job demands and available resources in the work environment. These surveys unearthed psychosocial hazards of moderate concern in the work environment, stimulating the development of targeted interventions, such as the training discussed above, to reduce their impact and the subsequent analysis of their effectiveness in future PAW-CON surveys.

3.4. The Role of Site Leaders

Site leaders played a pivotal role in connecting MATES with the workforce, enabling the successful achievement of all major milestones and enhancing MATES’ impact on site. A crucial element of the strategy developed by site leaders and MATES involved providing MATES training and conducting PAW-CON surveys. Whenever key events like training sessions or surveys were scheduled, leaders mandated attendance and ensured everyone on site was well-informed. Additionally, leaders maximised productivity during downtimes to meet milestones efficiently. In the face of unforeseen obstacles, like site flooding, leaders utilised such periods to enhance worker wellbeing through training sessions, surveys, or Toolbox Talks. Furthermore, site leaders nurtured a robust bond between MATES and the workforce by advocating for the regular presence of MATES Field Officers on site. This approach fostered a sense of familiarity and trust among workers, making it easier for them to seek and offer assistance when needed.

3.5. The Role of MATES at the Site

The strong partnership between site leaders at Rookwood Weir and MATES meant that the MATES Field Officers were highly involved in strengthening the wellbeing of the workforce during the project, both on site and at an individual level. Field Officers were readily available for consultations regarding the mental wellbeing of the workforce. Site leaders frequently reached out to MATES to explore new initiatives and their potential impact on the workforce. Additionally, supervisors could seek guidance from Field Officers if they had concerns about the wellbeing of specific workers or groups, especially following critical incidents. In each scenario, MATES offered expert advice on managing the situation effectively.

Field Officers also provided individualised support to the workforce as needed. Their consistent presence on-site, whether for talks, training sessions, or simply being available for informal chats, fostered familiarity and trust among workers. Through regular conversations, Field Officers assisted workers facing challenges at work or in their personal lives that impacted their mental health. Many workers proactively sought further support from Field Officers, who then connected them with case officers equipped to provide specialised assistance.

4. Results

The following sections present the survey results from both phases of the project, along with an additional comparison between 106 workers at Rookwood Weir (Time 4) and 1286 FIFO/DIDO workers from 11 similarly diverse sites who used the PAW-CON survey to evaluate psychosocial hazards between July 2023 and March 2024. This additional comparison is provided for greater insight into the success of the initiatives implemented at Rookwood Weir in line with our third hypothesis. For clarity, the relevant per-wave and benchmark sample sizes are restated within each Results subsection; analytic Ns for each model are reported in the corresponding table captions.

4.1. Evaluation of Psychosocial Hazards (Time 1 and Time 2)

4.1.1. Sample and Procedures

As described in Section 3.1, paper-based, anonymous surveys were administered during toolbox/pre-start meetings; Time 1—November 2021 (N = 138) and Time 2—August 2022 (N = 118). Due to the dynamic and changing nature of construction works at Rookwood Weir, participants at Time 1 and Time 2 were not the same. For similar reasons, it is also not possible to determine an exact response rate, although advice from the Rookwood Weir leadership team suggests it was above 50% of available workers on site during each data collection period.

4.1.2. Measures

People at Work-Construction (PAW-CON) Survey. Data on psychosocial hazards were gathered using the 36-item PAW-CON [6]. The PAW-CON is an empirically validated measure of the work environment underpinned by JD-R theory [10] and it has been used previously in research investigating psychosocial hazards within the building and construction industry [39,51]. The PAW-CON assesses 11 construction-relevant job demands and resources, with 3–4 items per domain:

• Role overload (4 items, e.g., I have unachievable deadlines).
• Role ambiguity (4 items, e.g., I know how to do the tasks required to get my job done).
• Role conflict (3 items, e.g., I receive incompatible requests from two or more people on this job).
• Job control (3 items, e.g., I have some say over the way I get the work done on this job).
• Coworker support (3 items, e.g., I get the help and support I need from my co-workers).
• Supervisor support (3 items, e.g., My direct supervisor on this job is willing to listen to my work-related problems).
• Supervisor task conflict (3 items, e.g., There are conflicts about ideas between me and my direct supervisor on this site).
• Supervisor relationship conflict (3 items, e.g., There are personality conflicts evident between me and my direct supervisor on this site).
• Praise and recognition (3 items, e.g., My direct supervisor on this job encourages me in my work with praise and thanks).
• Procedural justice (3 items, e.g., My direct supervisor on this job takes employee interests into account when making important decisions).
• Change consultation (4 items, e.g., I am clearly informed about the nature of the changes that take place in my workgroup).

All PAW-CON items were rated on a 7-point Likert scale (1 = Never, 7 = Always). For job resources, higher scores indicate greater availability of the resource; for job demands, higher scores indicate greater exposure to the demand. Role Ambiguity was reverse-coded so that higher scores reflect greater ambiguity. The PAW-CON survey is a practical tool that organisations can use to evaluate psychosocial hazards within their work environment. Each organisation partnering with MATES receives a detailed feedback report highlighting psychosocial hazards at a low (average score of 1.00–2.99 for job demands, 5.00–7.00 for job resources), moderate (3.00–4.99) or high (5.00–7.00 for job demands, 1.00–2.99 for job resources) potential risk. These classifications have been developed through ongoing industry consultation and the evaluation of PAW-CON responses from over 6000 workers in the Australian building and construction sector. Based on the findings, organisations are encouraged to prioritise their ongoing mental health and wellbeing initiatives towards those areas of greatest potential risk.

4.1.3. Statistical Analyses

To establish the prevalence of psychosocial hazards in the work environment at baseline (Time 1) and determine whether such hazards were increasing or decreasing prior to the implementation of control initiatives (Time 2), PAW-CON data from Time 1 and 2 were analysed using a one-way between-groups Multivariate Analysis of Variance (MANOVA), aiming to identify any significant differences in job demands and job resources within the general work environment over the nine-month period.

Table 1. Descriptive Statistics, Reliabilities and Intercorrelations between Study Variables (Full Sample—Time 1).

	M	SD	1	2	3	4	5	6	7	8	9	10	11
Role Overload	3.1	1.4	(0.88)
2. Role Ambiguity	2.1	1.1	0.311 **	(0.89)
3. Role Conflict	3.0	1.5	0.602 **	0.257 **	(0.90)
4. Supervisor Task Conflict	2.2	1.1	0.474 **	0.203 *	0.475 **	(0.90)
5. Supervisor Relationship Conflict	1.6	1.0	0.349 **	0.388 **	0.293 **	0.615 **	(0.93)
6. Job Control	4.1	1.6	−0.403 **	−0.488 **	−0.333 **	−0.176*	−0.282 **	(0.84)
7. Co-worker Support	5.9	1.2	−0.403 **	−0.489 **	−0.374 **	−0.343 **	−0.469 **	0.365 **	(0.89)
8. Supervisor Support	5.6	1.4	−0.347 **	−0.534 **	−0.336 **	−0.436 **	−0.543 **	0.427 **	0.646 **	(0.94)
9. Praise and Recognition	5.2	1.6	−0.383 **	−0.514 **	−0.347 **	−0.409 **	−0.528 **	0.463 **	0.570 **	0.796 **	(0.95)
10. Procedural Justice	5.5	1.4	−0.445 **	−0.566 **	−0.414 **	−0.503 **	−0.630 **	0.473 **	0.643 **	0.816 **	0.812 **	(0.90)
11. Change Consultation	4.6	1.6	−0.421 **	−0.499 **	−0.405 **	−0.265 **	−0.389 **	0.541 **	0.462 **	0.631 **	0.680 **	0.710 **	(0.92)

Note: N = 138; reliability coefficients (α) are in parentheses along the diagonal; * p < 0.05, ** p < 0.01. Psychosocial hazards are described as low (average score of 1.00–2.99 for job demands, 5.00–7.00 for job resources), moderate (3.00–4.99), or high (5.00–7.00 for job demands, 1.00–2.99 for job re-sources) potential risk based on an industry developed classification system.

and

Table 2. Descriptive Statistics, Reliabilities and Intercorrelations between Study Variables (Full Sample—Time 2).

	M	SD	1	2	3	4	5	6	7	8	9	10	11
Role Overload	2.9	1.3	(0.86)
2. Role Ambiguity	1.9	1.1	0.069	(0.90)
3. Role Conflict	2.8	1.3	0.609 **	0.281 **	(0.86)
4. Supervisor Task Conflict	2.3	1.2	0.478 **	0.158	0.530 **	(0.87)
5. Supervisor Relationship Conflict	1.6	1.0	0.371 **	0.224 *	0.336 **	0.732 **	(0.96)
6. Job Control	3.9	1.6	−0.102	−0.450 **	−0.228*	−0.182	−0.200 *	(0.88)
7. Co-worker Support	5.5	1.3	−0.181	−0.565 **	−0.305 **	−0.156	−0.206 *	0.368 **	(0.93)
8. Supervisor Support	5.3	1.6	−0.086	−0.544 **	−0.287 **	−0.282 **	−0.344 **	0.447 **	0.639 **	(0.98)
9. Praise and Recognition	4.7	1.8	−0.292 **	−0.461 **	−0.386 **	−0.375 **	−0.353 **	0.390 **	0.311 **	0.499 **	(0.96)
10. Procedural Justice	5.3	1.5	−0.351 **	−0.603 **	−0.461 **	−0.488 **	−0.496 **	0.406 **	0.436 **	0.625 **	0.765 **	(0.91)
11. Change Consultation	4.4	1.7	−0.325 **	−0.555 **	−0.492 **	−0.337 **	−0.272 **	0.573 **	0.489 **	0.574 **	0.611 **	0.660 **	(0.93)

Note: N = 118; reliability coefficients (α) are in parentheses along the diagonal; * p < 0.05, ** p < 0.01. Psychosocial hazards are described as low (average score of 1.00–2.99 for job demands, 5.00–7.00 for job resources), moderate (3.00–4.99), or high (5.00–7.00 for job demands, 1.00–2.99 for job re-sources) potential risk based on an industry developed classification system.

outline the mean, standard deviation, reliability, and intercorrelations across all job demands and resources for each sample.

4.1.4. Results from the Time 1 and 2 Surveys

Demographics. The overall sample of 256 workers was similar across the two data collection periods. The Time 1 cohort (138 workers) were predominately male (91.3%), had ten or more years of industry experience (64% of employees who reported) and had an average age of 40.5 years. The sample was primarily individuals employed by the principal contractor (58.1%). The most common employee role descriptions included tradesperson (23.6%), operator (18.1%), and labourer (15%). The direct supervisor was predominately the site manager (30.4%) or the foreman (27%). The Time 2 cohort (118 workers) were also predominately male (96.8%), had 10 or more years of industry experience (69.6% of workers who reported) and had an average age of 39.8 years. They were also predominately employed by the principal contractor (50.4%) as a tradesperson (25.7%), operator (19%) and labourer (15.2%). In contrast to the Time 1 sample, the most common direct supervisor of this group was the foreman (42.7%) followed by the site manager (25%).

Job Demands in the Work Environment. Initial insights from workers at baseline (Time 1) revealed that Role Overload and Role Conflict were, on average, areas of moderate risk according to the industry-established PAW-CON classification system noted earlier. Workers also reported that Role Ambiguity, Supervisor Task Conflict, and Supervisor Relationship Conflict were areas of low risk. Of particular concern during this pre-implementation period was the absence of any significant changes in worker experiences of reported job demands over the nine months between Time 1 and Time 2. While previously noted areas of moderate risk were now reported at, or slightly under, the threshold for moderate risk at Time 2, our analyses suggested that these changes, as well as areas previously noted as low risk, were not significantly different from the Time 1 results; F(5, 244) = 0.59, p = 0.70, Wilks’ Lambda = 0.99, partial eta squared = 0.01. These results (as outlined in

Table 3. Evaluation of Psychosocial Hazards—Time 1 vs. Time 2 (MANOVA).

Job Demand/Resource	Sample	N	Mean	SD	Mean Difference	Adjusted Cohen’s D
Role Overload	Time 1	135	3.06	1.40
	Time 2	115	2.96	1.30	−0.09	−0.07
Role Ambiguity	Time 1	135	2.11	1.12
	Time 2	115	1.94	1.08	−0.16	−0.15
Role Conflict	Time 1	135	2.96	1.51
	Time 2	115	2.80	1.34	−0.16	−0.11
Supervisor Task Conflict	Time 1	135	2.25	1.13
	Time 2	115	2.32	1.21	0.07	0.06
Supervisor Relationship Conflict	Time 1	135	1.59	0.96
	Time 2	115	1.62	0.95	0.03	0.03
Job Control	Time 1	134	4.05	1.58
	Time 2	115	3.92	1.58	−0.13	−0.08
Co-Worker Support	Time 1	134	5.83	1.18
	Time 2	115	5.51	1.30	−0.32 *	−0.26
Supervisor Support	Time 1	134	5.61	1.42
	Time 2	115	5.27	1.61	−0.35	−0.23
Praise and Recognition	Time 1	134	5.17	1.60
	Time 2	115	4.68	1.76	−0.50 *	−0.29
Procedural Justice	Time 1	134	5.44	1.40
	Time 2	115	5.29	1.48	−0.15	−0.11
Change Consultation	Time 1	134	4.56	1.58
	Time 2	115	4.44	1.71	−0.12	−0.07

Note: * p < 0.05; ** p < 0.01; Analyses used listwise deletion.

) suggest that Role Overload and Role Conflict were persistently at, or were approaching, moderate risk in the work environment and would likely benefit from targeted improvement strategies aimed at addressing their underlying factors.

Job Resources in the Work Environment. Workers at baseline (Time 1) highlighted job resources at low and moderate levels of potential risk. Specifically, results revealed that Job Control and Change Consultation were areas of moderate risk, while Coworker Support, Supervisor Support, Praise and Recognition and Procedural Justice were, on average, areas of low risk for workers. However, in contrast to worker experiences of job demands, results at Time 2 suggest significant differences in reported levels of job resources on site over time: F(6, 242) = 2.09, p = 0.05, Wilks’ Lambda = 0.95, partial eta squared = 0.05. Compared with Time 1, workers at Time 2 reported significant decreases in Coworker Support (M_Diff = −0.32, p = 0.04) and Praise and Recognition (M_Diff = −0.50, p = 0.02), whereas levels of Job Control, Supervisor Support, Procedural Justice, and Change Consultation trended lower but were not significantly different from baseline. These results (as shown in Table 3) suggest that Job Control and Change Consultation were persistently at moderate risk in the work environment, whereas Coworker Support and Praise and Recognition were areas likely to continue to deteriorate without targeted interventions developed to address the underlying factors.

4.2. Evaluation of Psychosocial Hazards (Time 3 and Time 4)

4.2.1. Sample and Procedures

As per Section 3.1, paper-based, anonymous surveys were administered during toolbox/pre-start meetings; Time 3—March 2023 (N = 176) and Time 4—August 2023 (N = 106). Recruitment of participants and data collection activities, including survey measures, were the same as Time 1 and 2. Due to the dynamic and changing nature of construction works at Rookwood Weir, participants at Time 3 and Time 4 were not the same.

4.2.2. Statistical Analyses

To establish the prevalence of psychosocial hazards in the work environment following the implementation of control initiatives and determine whether such hazards continued to be managed over time, PAW-CON data from Time 3 and Time 4 were analysed using a one-way between-groups Multivariate Analysis of Variance (MANOVA), aiming to identify any significant differences in job demands and job resources within the work environment in the post-intervention period.

Table 4. Descriptive Statistics, Reliabilities and Intercorrelations between Study Variables (Full Sample—Time 3).

	M	SD	1	2	3	4	5	6	7	8	9	10	11
Role Overload	2.9	1.2	(0.88)
2. Role Ambiguity	1.9	0.9	0.286 **	(0.91)
3. Role Conflict	2.5	1.3	0.561 **	0.253 **	(0.90)
4. Supervisor Task Conflict	2.1	1.1	0.420 **	0.245 **	0.562 **	(0.93)
5. Supervisor Relationship Conflict	1.6	1.0	0.285 **	0.234 **	0.455 **	0.760 **	(0.96)
6. Job Control	4.4	1.5	−0.205 **	−0.346 **	−0.228 **	−0.197 *	−0.211 **	(0.78)
7. Co-worker Support	5.7	1.3	−0.292 **	−0.314 **	−0.455 **	−0.308 **	−0.395 **	0.498 **	(0.95)
8. Supervisor Support	5.6	1.4	−0.318 **	−0.283 **	−0.397 **	−0.409 **	−0.457 **	0.578 **	0.618 **	(0.96)
9. Praise and Recognition	5.2	1.5	−0.289 **	−0.290 **	−0.410 **	−0.415 **	−0.435 **	0.543 **	0.550 **	0.669 **	(0.96)
10. Procedural Justice	5.6	1.2	−0.290 **	−0.319 **	−0.481 **	−0.474 **	−0.444 **	0.466 **	0.549 **	0.669 **	0.670 **	(0.89)
11. Change Consultation	5.0	1.4	−0.313 **	−0.357 **	−0.445 **	−0.340 **	−0.332 **	0.580 **	0.582 **	0.591 **	0.694 **	0.762 **	(0.90)

Note: N = 176; reliability coefficients (α) are in parentheses along the diagonal; * p < 0.05, ** p < 0.01. Psychosocial hazards are described as low (average score of 1.00–2.99 for job demands, 5.00–7.00 for job resources), moderate (3.00–4.99), or high (5.00–7.00 for job demands, 1.00–2.99 for job re-sources) potential risk based on an industry developed classification system.

and

Table 5. Descriptive Statistics, Reliabilities and Intercorrelations between Study Variables (Full Sample—Time 4).

	M	SD	1	2	3	4	5	6	7	8	9	10	11
Role Overload	2.4	1.2	(0.90)
2. Role Ambiguity	1.6	0.7	0.465 **	(0.90)
3. Role Conflict	2.2	1.2	0.643 **	0.410 **	(0.88)
4. Supervisor Task Conflict	1.9	1.1	0.533 **	0.319 **	0.596 **	(0.89)
5. Supervisor Relationship Conflict	1.4	0.8	0.482 **	0.369 **	0.453 **	0.755 **	(0.98)
6. Job Control	4.5	1.5	−00.124	−0.416 **	−0.216 *	−0.248 *	−0.284 **	(0.81)
7. Co-worker Support	5.9	1.2	−0.357 **	−0.470 **	−0.447 **	−0.484 **	−0.587 **	0.517 **	(0.93)
8. Supervisor Support	5.9	1.3	−0.411 **	−0.552 **	−0.453 **	−0.566 **	−0.566 **	0.447 **	0.720 **	(0.92)
9. Praise and Recognition	5.5	1.5	−0.419 **	−0.489 **	−0.398 **	−0.483 **	−0.523 **	0.460 **	0.578 **	0.773 **	(0.93)
10. Procedural Justice	5.8	1.2	−0.480 **	−0.473 **	−0.440 **	−0.560 **	−0.613 **	0.394 **	0.583 **	0.765 **	0.759 **	(0.91)
11. Change Consultation	5.4	1.4	−0.343 **	−0.473 **	−0.394 **	−0.462 **	−0.470 **	0.479 **	0.485 **	0.673 **	0.708 **	0.726 **	(0.89)

Note: N = 106; reliability coefficients (α) are in parentheses along the diagonal; * p < 0.05, ** p < 0.01. Psychosocial hazards are described as low (average score of 1.00–2.99 for job demands, 5.00–7.00 for job resources), moderate (3.00–4.99), or high (5.00–7.00 for job demands, 1.00–2.99 for job re-sources) potential risk based on an industry developed classification system.

outline the mean, standard deviation, reliability, and intercorrelations across all job demands and job resources for each sample.

4.2.3. Results from the Time 3 and 4 Surveys

Demographics. The overall sample of 282 workers was similar to the previous data collection periods (e.g., Time 1 and 2) and remained consistent across Time 3 and Time 4. The Time 3 cohort (176 workers) were predominately male (96.1%), had ten or more years of industry experience (69.5% of employees who reported) and had an average age of 41.3 years. The sample was primarily workers employed by a subcontractor (52%). The most common employee role descriptions included tradesperson (35.5%), operator (21.9%), and labourer (14.2%). The direct supervisor was predominately the leading hand (35.8%) or the foreman (29%). The Time 4 cohort (106 workers) were also predominately male (97.2%), had 10 or more years of industry experience (65.3% of workers who reported) and had an average age of 42.3 years. They were also employed in a similar range of positions, such as tradesperson (46.2%), operator (19.8) and labourer (14.2%). In contrast to the Time 3 sample, these respondents were predominately employed by the principal contractor (52.4%) and their direct supervisor was most commonly the site manager (38.7%) or the foreman (33%).

Job Demands in the Work Environment. Results following the implementation of mental health and wellbeing initiatives on site (Time 3) revealed that Role Overload, Role Ambiguity, Role Conflict, Supervisor Task Conflict, and Supervisor Relationship Conflict were, on average, reported by workers as low risk according to the industry-established PAW-CON classification system noted earlier. Furthermore, over the next six months, experiences of all job demands were trending, on average, lower at Time 4, approaching statistically significant levels of improvements; F(5, 264) = 2.10, p = 0.07, Wilks’ Lambda = 0.96, partial eta squared = 0.04. Importantly, compared with the data collected prior to the implementation of initiatives on site (Time 2), results suggest significant improvements in the work environment across most job demands at Time 4; F(5, 214) = 3.90, p ≤ 0.00, Wilks’ Lambda = 0.92, partial eta squared = 0.08. Specifically, reported levels of Role Overload (M_Diff = −0.55, p ≤ 0.01), Role Ambiguity (M_Diff = −0.39, p ≤ 0.01), Role Conflict (M_Diff = −0.62, p ≤ 0.00), and Supervisor Task Conflict (M_Diff = −0.44, p = 0.01), were all significantly lower at Time 4 compared to Time 2. Results for Supervisor Relationship Conflict (M_Diff = −0.23, p = 0.06) trended lower but were only marginally significantly different between the two time periods. Overall, these results suggest that the initiatives implemented on site were reasonably effective in reducing workplace stressors and eliminating areas of moderate risk, leading to a measurable improvement in employees’ perceptions of their work environment. Please refer to

Table 6. Evaluation of Psychosocial Hazards—Time 3 vs. Time 4 (MANOVA).

Job Demand/Resource	Sample	N	Mean	SD	Mean Difference	Adjusted Cohen’s D
Role Overload	Time 3	165	2.81	1.20
	Time 4	105	2.42	1.20	−0.39	−0.33
Role Ambiguity	Time 3	165	1.83	0.89
	Time 4	105	1.56	0.71	−0.27	−0.34
Role Conflict	Time 3	165	2.47	1.34
	Time 4	105	2.18	1.19	−0.29	−0.23
Supervisor Task Conflict	Time 3	165	2.10	1.10
	Time 4	105	1.88	1.05	−0.22	−0.20
Supervisor Relationship Conflict	Time 3	165	1.58	0.95
	Time 4	105	1.38	0.83	−0.20	−0.22
Job Control	Time 3	162	4.34	1.49
	Time 4	104	4.50	1.52	0.16	0.11
Co-Worker Support	Time 3	162	5.68	1.28
	Time 4	104	5.86	1.24	0.18	0.15
Supervisor Support	Time 3	162	5.61	1.45
	Time 4	104	5.90	1.29	0.30	0.22
Praise and Recognition	Time 3	162	5.15	1.51
	Time 4	104	5.50	1.55	0.36	0.23
Procedural Justice	Time 3	162	5.61	1.21
	Time 4	104	5.84	1.23	0.24	0.19
Change Consultation	Time 3	162	5.04	1.47
	Time 4	104	5.39	1.44	0.36	0.24

Note: * p < 0.05; ** p < 0.01; Analyses used listwise deletion.

and

Table 7. Evaluation of Psychosocial Hazards—Time 2 vs. Time 4 (MANOVA).

Job Demand/Resource	Sample	N	Mean	SD	Mean Difference	Adjusted Cohen’s D
Role Overload	Time 2	115	2.96	1.30
	Time 4	105	2.42	1.20	−0.55 **	−0.44
Role Ambiguity	Time 2	115	1.94	1.08
	Time 4	105	1.56	0.71	−0.39 **	−0.43
Role Conflict	Time 2	115	2.80	1.34
	Time 4	105	2.18	1.19	−0.62 **	−0.49
Supervisor Task Conflict	Time 2	115	2.32	1.21
	Time 4	105	1.88	1.05	−0.44 **	−0.39
Supervisor Relationship Conflict	Time 2	115	1.62	0.95
	Time 4	105	1.38	0.83	−0.23	−0.26
Job Control	Time 2	115	3.92	1.58
	Time 4	104	4.50	1.52	0.58 **	0.38
Co-Worker Support	Time 2	115	5.51	1.30
	Time 4	104	5.86	1.24	0.36 *	0.28
Supervisor Support	Time 2	115	5.27	1.61
	Time 4	104	5.90	1.29	0.63 **	0.43
Praise and Recognition	Time 2	115	4.68	1.76
	Time 4	104	5.50	1.55	0.83 **	0.50
Procedural Justice	Time 2	115	5.29	1.48
	Time 4	104	5.84	1.23	0.55 **	0.41
Change Consultation	Time 2	115	4.44	1.71
	Time 4	104	5.39	1.44	0.95 **	0.60

Note: * p < 0.05; ** p < 0.01; Analyses used listwise deletion.

for further details on the above results.

Job Resources in the Work Environment. Workers at Time 3 continued to highlight Job Control as an area of moderate risk, while reporting Coworker Support, Supervisor Support, Praise and Recognition, Procedural Justice and Change Consultation as areas of low risk. Compared to Time 3, experiences of all job resources were trending, on average, higher at Time 4, although no significant differences were reported between the two time periods; F(6, 259) = 0.77, p = 0.60, Wilks’ Lambda = 0.98, partial eta squared = 0.02. Importantly, when compared with the data collected prior to the implementation of initiatives on site (Time 2), results suggest significant improvements in the work environment across all job resources at Time 4; F(6, 212) = 3.58, p ≤ 0.01, Wilks’ Lambda = 0.91, partial eta squared = 0.09. Specifically, reported levels of Job Control (M_Diff = 0.58., p = 0.01), Coworker Support (M_Diff = 0.36, p = 0.04), Supervisor Support (M_Diff = 0.63, p ≤ 0.01), Praise and Recognition (M_Diff = 0.83, p ≤ 0.00), Procedural Justice (M_Diff = 0.55, p ≤ 0.01) and Change Consultation (M_Diff = 0.95, p ≤ 0.00), were all significantly higher at Time 4 compared to Time 2. Overall, these results suggest that the initiatives implemented on site were reasonably effective in improving workplace resources and eliminating areas of moderate risk, leading to a measurable improvement in employees’ perceptions of their work environment. Please refer to Table 6 and Table 7 for further details on the above results.

4.3. Comparing Rookwood Weir and Similarly Diverse FIFO/DIDO Sites

4.3.1. Sample and Procedures

To provide additional insight into the success of mental health and wellbeing initiatives implemented at Rookwood Weir, 106 workers who participated in the research at Time 4 (August 2023) were compared with 1286 construction workers from 11 similarly diverse FIFO/DIDO sites who used the PAW-CON survey to evaluate psychosocial hazards between July 2023 and March 2024, but did not use the survey findings to implement a holistic hazard management intervention. All comparative sites were located in regional Queensland and Northern Territory, Australia. Similarly to the data collected at Rookwood Weir, recruitment was conducted by MATES field officers, with invitations provided during toolbox/pre-start briefings and program awareness sessions, either before the workday or at times scheduled by the principal contractor. Participation was voluntary and, after providing consent, participants were provided with a paper questionnaire to complete. Owing to the dynamic nature of project work, an exact response rate could not be calculated; however, daily site reports across the 11 FIFO/DIDO sites indicated approximate response rates ranging from 50% to 100%. Comparisons with the benchmark cohort are provided to contextualise Rookwood’s Time-4 outcomes using the same instrument and procedures, and are not intended to serve as a controlled test of the holistic program.

4.3.2. Measures

People at Work-Construction Survey. Data on psychosocial hazards were gathered using the same 36-item PAW-CON survey [6] used to collect data at Rookwood Weir.

General Mental Health. General mental health was measured with the 12-item version of the General Health Questionnaire GHQ [52]. The GHQ is a commonly used measure for detecting minor psychiatric disorders and psychological distress among participants in community and organisational settings [53,54]. In line with previous research evaluating psychosocial hazards and general mental health in the construction industry [6], the standard response scale of the GHQ was adapted to be scored on a 7-point scale, ranging from 1 (Never) to 7 (Always), which is consistent with the response scale of the PAW-CON. All positively worded items were subsequently reverse-coded, such that higher average scores across the 12-items indicated increased psychological distress and likelihood of poor mental health. Finally, to identify psychological distress, survey responses were transformed prior to analyses, whereby “Never” = 0, “Rarely/Once in a While” = 1; “Some of the Time/Fairly Often” = 2; “Often/Always” = 3. Once transformed, a participant’s sum score of the GHQ was computed out of a maximum of 36, where a score exceeding 12 was used as a threshold for identifying cases of “psychological distress” vs. “typical” general mental health. This threshold is commonly used within the literature for detecting minor psychiatric disorders and instances of poor mental health among participants [54,55].

Past and Future Suicide Ideation. Past and future suicide ideation were measured using two questions adapted from the Suicide Behaviors Questionnaire-Revised (SBQ-R) [56]. Participants were asked to evaluate feelings of past suicide ideation by responding to the question, “How often have you thought about taking your own life in the past 6 months?” The language in this question was adapted to suit the Australian building and construction context by using the wording “taking your own life” instead of “killing yourself”, which is used in the SBQ-R. Feelings of future suicide ideation were evaluated with the question, “How likely is it that you will attempt suicide someday?” Both questions were reported using the same response scales as noted in the SBQ-R (e.g., Never-Very Often; Never-Very Likely). To ease interpretation of results and help identify at-risk individuals, responses were transformed prior to analyses into two states, whereby past suicide ideation was recoded so that “Never/Rarely” = “Rarely” and “Sometimes/Often/Very Often” = “Often”. Conversely, future suicide ideation was recoded so that “Never/No Change at All/Rather Unlikely/Unlikely” = “Unlikely”, and “Likely/Rather Likely/Very Likely” = “Likely”.

4.3.3. Statistical Analyses

To establish workplace mental health trends at Rookwood Weir following the implementation of initiatives and compare them against the broader Australian FIFO/DIDO construction work environment, the data were analysed using a one-way between-groups Multivariate Analysis of Variance (MANOVA) to identify any significant differences in job demands, job resources, mental health, and suicide ideation among the two groups.

Table 8. Descriptive Statistics, Reliabilities and Intercorrelations between Study Variables (Full Sample—Time 4 and Comparable FIFO/DIDO Sites).

	M	SD	1	2	3	4	5	6	7	8	9	10	11	12	13	14
Role Overload	2.6	1.3	(0.87)
2. Role Ambiguity	1.8	0.8	0.276 **	(0.87)
3. Role Conflict	2.5	1.3	0.638 **	0.306 **	(0.90)
4. Supervisor Task Conflict	2.1	1.2	0.497 **	0.308 **	0.563 **	(0.92)
5. Supervisor Relationship Conflict	1.5	1.0	0.371 **	0.290 **	0.409 **	0.669 **	(0.94)
6. Job Control	4.2	1.4	−0.144 **	−0.304 **	−0.090 **	−0.138 **	−0.191 **	(0.81)
7. Co-worker Support	5.7	1.2	−0.363 **	−0.393 **	−0.347 **	−0.350 **	−0.357 **	0.378 **	(0.93)
8. Supervisor Support	5.6	1.5	−0.380 **	−0.433 **	−0.355 **	−0.494 **	−0.557 **	0.396 **	0.579 **	(0.95)
9. Praise and Recognition	5.0	1.6	−0.346 **	−0.400 **	−0.296 **	−0.451 **	−0.460 **	0.424 **	0.500 **	0.688 **	(0.94)
10. Procedural Justice	5.6	1.3	−0.394 **	−0.453 **	−0.396 **	−0.537 **	−0.560 **	0.343 **	0.498 **	0.734 **	0.713 **	(0.90)
11. Change Consultation	5.0	1.5	−0.290 **	−0.467 **	−0.308 **	−0.341 **	−0.354 **	0.431 **	0.446 **	0.583 **	0.618 **	0.669 **	(0.91)
12. Psychological Distress	2.6	0.9	0.408 **	0.398 **	0.388 **	0.359 **	0.381 **	−0.352 **	−0.426 **	−0.430 **	−0.472 **	−0.453 **	−0.479 **	(0.90)
13. Past Suicidal Ideation	0.3	0.7	0.103 **	0.125 **	0.132 **	0.146 **	0.188 **	−0.125 **	−0.142 **	−0.189 **	−0.195 **	−0.142 **	−0.133 **	0.398 **
14. Future Suicidal Ideation	0.5	1.0	0.103 **	0.139 **	0.124 **	0.130 **	0.175 **	−0.083 **	−0.137 **	−0.152 **	−0.195 **	−0.169 **	−0.181 **	0.350 **	0.480 **

Note: N = 1392; reliability coefficients (α) are in parentheses along the diagonal; * p < 0.05, ** p < 0.01.

outlines the mean, standard deviation, reliability, and intercorrelations across all job demands and job resources for the full sample.

4.3.4. Results from the Additional Sites and Rookwood Weir

Demographics. The 1392 workers sampled were similar across the two groups. The demographics of the Rookwood Weir sample (Time 4—106 workers) is reported previously above. The non-Rookwood FIFO/DIDO sample (1286 workers) were predominately male (95.4%), had ten or more years of industry experience (64.5% of employees who reported) and had an average age of 38 years. The sample was primarily workers employed by a subcontractor (67.3%). Employee role descriptions included a range of positions, including tradesperson (41.2%), labourer (13.9%), leading hand/foreman (10.6%) and operators (10.2%). The direct supervisor was predominately the foreman (36.5%) and the leading hand (32.7%).

Differences between Rookwood Weir and Comparable FIFO/DIDO Sites. The results suggest that workers at Rookwood Weir experienced, on average, lower job demands, higher job resources, and reduced psychological distress compared to workers in the broader Australian FIFO/DIDO construction work environment; F(14, 1073) = 1.675, p = 0.05, Wilks’ Lambda = 0.98, partial eta squared = 0.02. When reviewing the results for job demands, Rookwood Weir workers, on average, reported significantly lower levels of Role Ambiguity (M_Diff = −0.24, p = 0.02) and Role Conflict (M_Diff = −0.31, p = 0.05). In terms of job resources, Rookwood Weir workers reported significantly higher levels of Job Control (M_Diff = 0.47, p ≤ 0.01), Supervisor Support (M_Diff = 0.38, p = 0.03), Praise and Recognition (M_Diff = 0.65, p ≤ 0.00), Procedural Justice (M_Diff = 0.31, p = 0.05), and Change Consultation (M_Diff = 0.52, p ≤ 0.01). However, average levels of Role Overload, Supervisor Task Conflict, Supervisor Relationship Conflict, and Coworker Support were not significantly different between the two groups. In terms of mental health and suicide ideation, the results suggest that Rookwood Weir workers, on average, reported significantly lower instances of psychological distress (M_Diff = −0.12, p ≤ 0.04) than workers at comparable FIFO/DIDO sites; past and future suicidal ideation did not differ significantly between groups. Please refer to

Table 9. Evaluation of Psychosocial Hazards and Mental Health—Comparable FIFO/DIDO Sites vs. Rookwood Weir—Time 4 (MANOVA).

Variable	Sample	N	Mean	SD	Mean Difference	Adjusted Cohen’s D
Role Overload	Comparable FIFO/DIDO Sites	1010	2.59	1.30
	Rookwood Weir—Time 4	78	2.48	1.24	−0.12	−0.09
Role Ambiguity	Comparable FIFO/DIDO Sites	1010	1.82	0.83
	Rookwood Weir—Time 4	78	1.59	0.72	−0.24 *	−0.30
Role Conflict	Comparable FIFO/DIDO Sites	1010	2.51	1.34
	Rookwood Weir—Time 4	78	2.20	1.20	−0.31 *	−0.24
Supervisor Task Conflict	Comparable FIFO/DIDO Sites	1010	2.13	1.14
	Rookwood Weir—Time 4	78	1.90	0.96	−0.24	−0.23
Supervisor Relationship Conflict	Comparable FIFO/DIDO Sites	1010	1.56	0.99
	Rookwood Weir—Time 4	78	1.41	0.81	−0.15	−0.16
Job Control	Comparable FIFO/DIDO Sites	1010	4.21	1.43
	Rookwood Weir—Time 4	78	4.67	1.47	0.47 **	0.32
Co-Worker Support	Comparable FIFO/DIDO Sites	1010	5.75	1.20
	Rookwood Weir—Time 4	78	5.90	1.15	0.15	0.13
Supervisor Support	Comparable FIFO/DIDO Sites	1010	5.53	1.51
	Rookwood Weir—Time 4	78	5.91	1.18	0.38 *	0.28
Praise and Recognition	Comparable FIFO/DIDO Sites	1010	4.99	1.60
	Rookwood Weir—Time 4	78	5.65	1.41	0.65 **	0.43
Procedural Justice	Comparable FIFO/DIDO Sites	1010	5.56	1.34
	Rookwood Weir—Time 4	78	5.88	1.20	0.31 *	0.24
Change Consultation	Comparable FIFO/DIDO Sites	1010	4.93	1.49
	Rookwood Weir—Time 4	78	5.45	1.36	0.52 **	0.37
Psychological Distress	Comparable FIFO/DIDO Sites	1010	0.50	0.50
	Rookwood Weir—Time 4	78	0.38	0.49	−0.12 *	−0.24
Past Suicidal Ideation	Comparable FIFO/DIDO Sites	1010	0.08	0.28
	Rookwood Weir—Time 4	78	0.08	0.27	−0.01	−0.03
Future Suicidal Ideation	Comparable FIFO/DIDO Sites	1010	0.02	0.12
	Rookwood Weir—Time 4	78	0.00	0.00	−0.02	−0.18

Note: * p < 0.05; ** p < 0.01; Analyses used listwise deletion.

for further details on the above results.

5. Discussion

This case study evaluated the effectiveness of a structured psychosocial risk management initiative delivered at a large regional construction site in partnership with MATES in Construction. Using the JD-R framework, the initiative aimed to reduce psychosocial hazards, improve job resources, and ultimately enhance mental health and wellbeing outcomes for workers. The intervention was assessed over multiple timepoints and benchmarked against a broader sample of FIFO/DIDO construction sites, enabling a baseline to compare the psychological wellbeing of the case workforce.

This case presented a workplace exposed to psychosocial hazards typical of the construction industry, including long hours, manual labour, temporary project-based work and associated job uncertainty, and a predominantly male workforce (with associated cultural norms) [2,16,17]. Additionally, the remote location and rostered work arrangements (i.e., FIFO/DIDO work with long periods away from home) further exacerbates these challenges by contributing to psychological isolation [6]. Adverse weather events resulted in compressed timeframes, introduction of night shift and increased job pressure. From Time 1 to Time 2, the results indicated workers were experiencing role overload and role conflict; poor job control and change consultation; and deterioration of coworker support and praise and recognition. Without targeted intervention, the escalating demands of the project would likely begin to outweigh the diminishing resources available to the workforce.

The findings support H1 and H2 and partially support H3. In line with Hypothesis 1, there was a significant reduction in job demands—particularly role overload, role ambiguity, role conflict and supervisor task conflict—at the end of the project (Time 4) compared to the pre-intervention period (Time 2). This is a promising outcome, given that chronic exposure to high job demands has been consistently linked to mental ill health, emotional exhaustion, and burnout in the construction sector [2,40]. These results suggest the targeted interventions were effective in altering workplace conditions that would otherwise place strain on employees, offering practical insight into how demands can be managed even within high-pressure environments like construction.

Consistent with Hypothesis 2, job resources also improved significantly over the time periods. Compared to the pre-intervention period (Time 2), notable gains were reported in job control, co-worker support, supervisor support, praise and recognition, procedural justice, and change consultation at the end of the project (Time 4). As highlighted by Lesener et al. [11], these types of job resources buffer the negative effects of high demands and promote engagement, resilience, and wellbeing. Improved perceptions of fairness and participation in change processes are also recognised as protective against psychological strain [37]. These findings align with broader calls for integrated, multi-level interventions that actively promote mental health, in addition to minimising harm [18,36].

Hypothesis 3 was partially supported: psychological distress was lower at the case site, whereas suicidal ideation did not differ from benchmarks. The research did not identify a significant difference in suicidal ideation between workers at Rookwood and the benchmark sites. This result may be influenced by the psychometric properties of the measure used. Suicidal ideation is a low-base-rate outcome with limited statistical power at site level over short windows, and it may be less sensitive than general distress to workplace condition changes. Targeted suicide-prevention capability and help-seeking pathways likely require longer follow-up to translate into detectable differences in ideation.

Supporting Hypothesis 3, workers reported lower levels of psychological distress at Rookwood compared to similar work sites at the end of the intervention. Although these variables were not tracked over the course of the project, these improvements are still significant given the elevated rates of mental ill-health in the construction sector [2,18]. The results are consistent with research showing that reducing job strain and improving psychosocial conditions can have a positive impact on psychological outcomes [13,16]. They highlight the value of industry-specific, evidence-based programs like MATES in Construction, which provide both individual-level support and organisational-level tools to address systemic drivers of poor mental health.

The role of leadership emerged as a central factor in the initiative’s success. Site leaders not only ensured full participation in training and survey activities but also demonstrated sustained commitment to the integration of mental health into routine site operations. Their proactive engagement with MATES and use of downtime for wellbeing-related initiatives helped normalise discussions of mental health and built a strong culture of support. This aligns with previous research demonstrating that leadership buy-in—both at senior and frontline levels—is a critical enabler of workplace mental health interventions [43].

While these results are encouraging, several limitations should be acknowledged. The dynamic nature of the construction workforce limited the ability to track matched participants across all timepoints, constraining causal inference. Furthermore, because surveys were anonymous and total on-shift employee numbers were unavailable, we could not investigate within-individual differences or determine exact response rates; consequently, our results reflect unadjusted comparisons of independent cross-sectional snapshots and residual confounding cannot be excluded. Several observed differences were also small in magnitude, and some p-values were close to the α = 0.05 threshold; while we treat p = 0.05 as statistically significant, these estimates should be interpreted cautiously and verified in future studies. Reliance on self-report measures introduces potential for social desirability or response bias, although the anonymity of the surveys likely helped mitigate this. The case study design also limits generalisability to other sites and is not a controlled multi-site trial, though benchmarking against 11 comparable FIFO/DIDO construction sites adds robustness to the findings. Additionally, suicidal ideation was measured once at the case site using brief items, which may limit sensitivity to change.

Future research would benefit from longer-term follow-up to assess the sustainability of change beyond the life of the project and from matched or panel designs and multi-site controlled evaluations to isolate mechanisms and strengthen generalisability. The use of mixed-methods approaches—including qualitative interviews with leaders, field officers and workers—could further illuminate how change occurs and assess leadership behaviours. Studies should also incorporate objective indicators—injury and near-miss data, absenteeism and turnover—and undertake economic evaluation. Finally, targeted analyses of high-risk subgroups (for example, apprentices and younger workers) and testing of specific controls during schedule compression or night work are warranted, alongside further evaluation of integrated, holistic approaches that embed psychosocial risk management into core safety and operational frameworks, particularly in male-dominated, high-risk industries like construction where mental health remains a critical concern [35].

6. Conclusions

This case study evaluated a structured holistic psychosocial hazard management intervention on a large regional construction project. The PAW-CON survey was used across four waves and outcomes benchmarked against 11 comparable FIFO/DIDO sites. Several job demands were significantly reduced (role overload, role ambiguity, role conflict, and supervisor task conflict), while key job resources significantly increased (job control, supervisor support, co-worker support, praise and recognition, procedural justice and change consultation). At the project’s end, workers at the case site reported lower psychological distress than workers at comparable sites, although suicidal ideation did not differ. Taken together, the findings indicate that a targeted, multilevel program implemented with consistent leadership support can shift both risks and resources in ways consistent with improved mental health.

Workplace characteristics likely to have driven these effects were visible leadership commitment and an industry partnership model. The site treated psychosocial risk management as core business rather than an add-on, mandated participation during routine processes, used downtime to deliver training and surveys, and normalised help-seeking through MATES field presence and connector training. These features coincided with reduced role conflict and ambiguity and increased change consultation, procedural justice, recognition and supervisor support, indicating a more enabling social and procedural climate.

Overall, this study adds to the evidence that structured, evidence-based interventions can deliver statistically significant improvements to psychosocial conditions and worker mental health. It also offers a practical example of how an integrated psychosocial hazard reduction program can operate in project-based construction. The JD-R framework and PAW-CON provided a robust basis for identifying key risks, guiding action and evaluating positive change, and the partnership with MATES, underpinned by committed leadership, was pivotal to implementation. The findings reinforce the importance of multi-level strategies that engage internal and external stakeholders, address systemic risks while providing individual support, and point to a clear pathway for strengthening psychological safety in construction.