**3. Results**

#### *3.1. Stress-Related Productivity Impairment Costs*

Of the 375 employees who reported experiencing stress at work, 23.4% (*n* = 82) reported at least one day of absence from work due to personal health problems in the past four weeks. Twenty-five employees (6.7%) also stated a high degree of presenteeism, reporting that they had trouble concentrating at work or doing their best due to personal health problems 'most of the time' or 'all of the time' in the past four weeks. Overall, stressed employees were associated with an average of 33.6% work impairment and \$45,240.70 (SD = 30,655.26) in productivity costs per employee.

Independent sample t-tests and a one-way ANOVA were performed to examine the differences in productivity costs depending on employee characteristics (see Tables 2 and 3). Significant differences were identified between roster type, contract, and residency status. As seen in Table 2, employees on permanent day shifts were more likely to be associated with higher productivity costs compared to employees on rotating/alternating shifts, Cohen's *D* = 0.41. Permanent employees were significantly more likely to report higher work impairment which was associated with higher productivity costs compared to contractors, Cohen's *D* = 0.42. Employees who resided in the mining towns were also more likely to be associated with higher productivity costs compared to FIFO/DIDO employees, Cohen's *D* = 0.35. Based on Cohen's *D* conventions, these differences were considered small to medium. Productivity costs did not significantly differ between males and females and among the different age groups.


**Table 3.** Group differences based on work impairment percentage and annual productivity cost.

Notes. *n* = 375; FIFO/DIDO = Fly-in, Fly-out/Drive-in, Drive-out; Differences in group sample size (*n*) is due to missing data on employee characteristics.

Productivity costs also significantly differed between self-reported levels of stress at work (see Table 3). A trend is observed in Table 3 with increased frequency of experiencing stress at work being associated with increased productivity costs. Employees who reported feeling stressed 'all of the time' in the previous four week period reported the highest productivity costs (*M* = \$75,337.16; SD = \$47,379.12). Post-hoc analysis using Games-Howell test were performed due to the unequal variances and sample sizes between the groups. Employees who reported experiencing stress 'all of the time' showed significantly higher costs compared to employees who reported only feeling stressed 'some of the time', *p* = 0.012 and the difference showed a large effect, Cohen's *D* = 1.02. Employees who reported feeling stressed 'most of the time' were also associated with significantly higher productivity costs compared to those who only experienced stress 'some of the time', *p* < 0.001, Cohen's *D* = 1.00 and 'a good bit of the time', *p* = 0.024, Cohen's *D* = 0.53. The effect sizes were large and medium, respectively. Employees who reported feeling stressed 'a good bit of the time' were also associated

with higher productivity costs compared to employees who only felt stressed 'some of the time' at work, *p* = 0.004, and the effect was small to medium size, Cohen's *D* = 0.43.

Differences in productivity costs depending on stage of change for stress management were examined. Table 4 reveals that, on average, the highest productivity impairment costs were associated with employees classified in the contemplation stage of change for stress management while employees in the pre-contemplation stage were associated with the lowest average productivity impairment cost. A one-way ANOVA was conducted to assess the differences in productivity costs based on employees' readiness to change their stress management behaviours. An overall significant difference was found, *F* (4, 307) = 6.78, *p* < 0.001. Post hoc analysis using Games-Howell test suggest that, contemplation employees were associated with significantly higher productivity costs compared to pre-contemplation employees, *p* < 0.001 and the difference had a medium to large effect, Cohen's *D* = 0.77. Employees in the action stage were also associated with significantly higher productivity costs compared to pre-contemplation employees, *p* = 0.015. However, the effect size was only small to medium, Cohen's *D* = 0.46. No other significant differences were found between the other groups.

**Table 4.** Work impairment percentage and annual productivity cost per person based on stage of change.


Notes. *n* = 375.

#### *3.2. Desire for Assistance with Stress Management*

Of the 375 employees who reported experiencing stress at work, only 28% (*n* = 105) reported wanting assistance to manage their stress. A hierarchical logistic regression was performed to predict the likelihood for wanting assistance with stress management. Participants' employee characteristics and reported levels of stress at work were included in Step 1 to partial out their effect in the regression model. The stage of change for stress management variable was included in Step 2. Step 1 was significant, χ2(8) = 39.76, *p* < 0.001, explaining 19.4% of the total variance (Nagelkerke *r*<sup>2</sup> = 0.19) and correctly classified 76.4% of the cases. The model's sensitivity was 37.7 and specificity was 91.4. As shown in Table 5, being female and having higher levels of stress were significant predictors of desiring stress management assistance. Female employees were twice as likely to want stress management assistance compared to male employees. Employees who reported experiencing stress more frequently were significantly more likely to report desiring stress management assistance compared to employees who reported experiencing stress less frequently. Age, shift work rotation, employment status, and current work arrangement were not significant predictors within the Step 1 of the model.



*IJERPH* **2019** , *16*, 94

FIFO/DIDO = fly-in,

fly-out/DRIVE-in,

 drive-out; OR = odds ratio; CI = confidence intervals; † Reference.

When the stages of readiness to change were added in the model, the second block was significant, χ2(4) = 32.65, *p* < 0.001, explaining an additional 13.9% of the total variance. Overall, the model was significant, χ2(12) = 72.42, *p* < 0.001 and explained 33.3% of the total variance (Nagelkerke *r*<sup>2</sup> = 0.33) and accurately classified 78.2% of the cases. The model's sensitivity was 49.4 and specificity was 89.4. Pre-contemplation employees were the reference category for the readiness to change one's approach to stress management. As shown in Table 5, employees in all readiness stages were significantly more likely to desire stress management assistance than stressed employees who reported that their stress management habits do not need changing. However, it is important to note that the odds ratio's lower confidence intervals for employees in the action stage nearly encompasses 1, indicating that being in the action stage may not necessarily increase the employees' likelihood to desire assistance for their stress. The odds ratio of the action group in desiring assistance is also small (OR = 2.85) compared to employees in the pre-contemplation stage. Examination of the other groups revealed that employees in the contemplation stage showed the highest odds ratio, indicating that this group were 11.05 times more likely to want assistance for stress management compared to the pre-contemplation group. Furthermore, compared to the pre-contemplative employees, preparation employees were 7.13 times and maintenance employees were 5.22 times more likely to desire assistance for stress management.

#### **4. Discussion**

Within a mining workforce sample, this study identified employee characteristics and stress management stages of change associated with high stress-related productivity impairment costs and desire for assistance with stress management. Of the employees who reported experiencing stress at work, employee groups associated with significantly higher productivity impairment costs included: day shift workers; permanent contract employees; employees who reside within the mining towns; frequently stressed employees; employees intending to better manage their stress in the next six months; and employees who are actively managing their stress. Although previous research has reported that FIFO miners were at risk of stress due to their work arrangements [12,13] in the current study they had lower stress impairment costs than permanent day shift employees.

The lower productivity impairment costs associated with alternating rosters, contractors and FIFO/DIDO employment may be related to the roles and responsibilities associated with the different types of employees appointed to permanent day shifts versus alternating contracts. To protect participant anonymity the current study did not gather data regarding job position, however, the researchers are aware that within the current sample, employees working permanent day shifts included managers, professionals, administrative, and operational staff. Comparatively, employees with alternating rosters, FIFO/DIDO employment, and contractual work were more likely to be appointed to operational mining roles. A study by Ling et al. [5] showed that managers within the coal mining industry had higher average lost productivity time costs compared to machine operators and trade workers. Managers were also associated with higher psychological distress [2].

In this mining workforce sample, employees in the contemplation stage for stress management had the highest average annual stress-related productivity impairment cost (\$53,761 per employee). Although individuals who are classified as contemplative intend to improve their behaviour in the next six months, they are considered not ready to self-initiate immediate changes [18]. Research has shown in a national population based sample that a stage-matched stress management intervention was effective in achieving rapid progression through the stages of effective stress management strategy adoption [18]. Given that the contemplation group for stress management in this study had high impairment costs, savings could be achieved by assisting these employees to manage or remove stress that is impacting on their work productivity. Future research should be conducted to identify if implementation of a stage of change matched stress management intervention could achieve similar results in a mining workforce as achieved in the population based sample.

Consistent with previous research that found females were more likely to exhibit help-seeking behaviours and access professional stress management or mental health support services [21–23], the current study found that females were significantly more likely to report wanting assistance with stress management. Although males and females had similar average impairment costs, the higher proportion of males in the mining workforce and the lower likelihood of males to want assistance suggests that practitioners need to ensure that stress management promotion is appropriate for engaging males. Male employees appear to be unlikely to initiate help seeking behaviours despite experiencing stress at work. Given that this study found that higher frequency of experiencing stress at work was associated with higher impairment costs, it was encouraging to identify that higher stress frequency was also associated with greater desire for assistance with stress management. This suggests that workplace provided stress management assistance will likely appeal to high productivity impairment cost employees that were frequently experiencing stress at work.

Similarly, stage of change for adoption of effective stress management behaviours analyses revealed that employees in the contemplation stage had the both the highest cost impairment and highest likelihood to report wanting assistance. By contrast, employees in the pre-contemplation stage who believed their stress management habits did not need changing, were found to have both the lowest cost impairment and lowest likelihood to report wanting assistance. This again indicates that workplace provided stress management assistance will likely appeal to high productivity impairment cost employees that were intending to improve their stress management in the next six months.
