ORIGINAL ARTICLE
Workplace Stress What Is the Role of Positive Mental Health? Kathryn M. Page, PsyD, Allison J. Milner, PhD, Angela Martin, PhD, Gavin Turrell, PhD, Billie Giles-Corti, PhD, and Anthony D. LaMontagne, SciD
Objective: To examine whether positive mental health (PMH)—a positively focused well-being construct—moderates the job stress–distress relationship. Methods: Longitudinal regression was used to test two waves of matched, population-level data from a sample of older, working Australian adults (n = 3291) to see whether PMH modified the relationship between work stress and later psychological distress. Results: Time 1 work stress was positively associated with distress at both time points. Positive mental health was negatively associated with work stress at both time points. Positive mental health modified the impact of work stress on psychological distress. This effect only occurred for those with the highest levels of PMH. Conclusions: Positive mental health may help protect workers from the effect of workplace stress but only in a small proportion of the population. Therefore, to improve workplace mental health, workplaces need to both prevent stress and promote PMH.
O
ne in five Americans has a mental disorder that impairs their social, interpersonal, and occupational functioning (1-year prevalence rate),1 with similar rates observed across the developed world.2 The significant economic cost associated with this burden of disease, and in particular to employers, has been well documented.3–5 Research into processes that may prevent or reduce the effect of workplace stress on employee mental health is therefore important for workplaces, as well as governments and occupational and public health bodies. Although the etiology of mental health problems is complex and includes a wide range of individual factors, work-related stress is a significant contributor. Studies have consistently shown that job stress, whether measured as overall perceived stress or reported exposure to psychosocial stressors, can impair mental health.5,6 Szeto and Dobson5 analyzed overall perceived levels of work stress in a nationally representative sample of more than 10,000 Canadian employees. Those who reported the highest level of stress were approximately three times more likely to have been treated for an emotional From the McCaughey VicHealth Centre for Community Well-being (Drs Page, Milner, Giles-Corti, and LaMontagne), Melbourne School of Population and Global Health, the University of Melbourne; the Tasmanian School of Business and Economics (Dr Martin), University of Tasmania, Hobart; School of Public Health and Social Work (Dr Turrell), Queensland University of Technology, Brisbane; and Population Health Strategic Research Centre (Drs Page and LaMontagne), School of Health and Social Development, Deakin University, Melbourne, Australia. Dr Turrell is supported by an Australian National Health and Medical Research Council (NHMRC) Senior Research Fellowship (#1003710). Dr LaMontagne is supported by Centre grant funding from the Victorian Health Promotion Foundation, Melbourne, Australia (#15732). Dr Giles-Corti is supported by an NHMRC Principal Research Fellowship (#1004900) and Centre grant funding from the Victorian Health Promotion Foundation, Melbourne, Australia (#15732). Drs Page and Milner are supported by an NHMRC Capacity Building grant (#546248). The authors declare no conflicts of interest. Address correspondence to: Kathryn M. Page, PsyD, Work, Health and Wellbeing Unit, Deakin Population Health Strategic Research Centre, Faculty of Health, Deakin University, Melbourne Burwood campus, 221 Burwood Hwy, Burwood, Victoria 3125, Australia ([email protected]). C 2014 by American College of Occupational and Environmental Copyright Medicine DOI: 10.1097/JOM.0000000000000230
814
or mental health problem in the past 12 months and approximately 2.4 times more likely to be diagnosed with a mood and/or anxiety disorder than their nonstressed counterparts. This cross-sectional finding is consistent with systematic reviews6,7 and longitudinal and prospective studies examining the link between stressors and mental health.8,9 For example, Stansfeld and Candy’s6 systematic review and meta-analysis found that low decision latitude, low social support, high psychological demands, job strain (high demand/low control), effort–reward imbalance (high effort/low reward), and job insecurity—all well-known psychosocial job stressors—were all predictive of common mental problems. Particularly strong associations were found between job strain and effort–reward imbalance and the onset of common mental health problems. These findings are consistent with the job demand–control and effort–reward imbalance models.10,11
IDENTIFYING FACTORS THAT PROTECT WORKER MENTAL HEALTH Identifying the factors that may protect workers from experiencing work-related mental health problems is an important area of ongoing research. To date, research has identified a number of work and individual-level moderators of the stressor–strain relationship including job control and supervisor support.12 Identifying such moderators is paramount to the buffering or interactive hypothesis in job stress research, which predicts that certain factors (typically job control or social support) can moderate the relationship between job stressors and adverse psychological outcomes.12,13 The emergence of positive psychology has opened the door to a wider array of factors that may moderate the relationship between work and mental (ill) health that have not yet been investigated within a job stress framework. Positive psychology is the study of “the conditions and processes that contribute to the flourishing or optimal functioning of people, groups, and institutions.”14 An application of positive psychology (also called positive organizational behavior15 ) in the work context includes identifying and enhancing worker strengths or positive capabilities to facilitate improved mental health outcomes. Here, worker strengths and positive capabilities refer to adaptable capacities and strengths such as psychological capital16,17 or positive psychological well-being.18 An increasing body of literature attests to the benefits of strengths or positive capabilities on both health- and work-related constructs19–22 and, in particular, how the presence of such factors, for example, optimism, positive emotion, and positive or supportive relationships, can protect and/or enhance mental and physical health.23,24 In this article, we look at whether positive mental health (PMH)—a broad construct that encompasses several positive and protective elements25 —can protect an employee from the ill-effects of stress on mental health, using two waves of population-level data. Positive mental health is a positively focused construct that captures the degree to which individuals are flourishing in life, that is, reporting high levels of positive emotional, social, and psychological well-being.23,25 Positive mental health should not be seen as the direct antithesis to psychological ill-being; in fact, the two are only moderately correlated and factor into two clear and distinct dimensions.26,27 Studies have also found that having high levels of JOEM r Volume 56, Number 8, August 2014
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
JOEM r Volume 56, Number 8, August 2014
PMH plus low levels of mental ill-health (sometimes called complete mental health) predicts work- and health-related outcomes over and above that explained by having low levels of mental illness alone.26 Therefore, studies examining the relationship between these variables, alongside work-related exposures, including workplace stress, may provide a new and potentially useful contribution to the occupational health literature. In this study, we look at the potential for PMH to moderate the relationship between workplace stress and conflict on later psychological distress. Specifically, we hypothesize that employees with higher levels of PMH will have lower levels of psychological distress associated with work stress than those with lower levels of PMH. That is, PMH will buffer or protect an individual from the negative impact of work stress.
METHODS This article used two waves of data from the Australian HABITAT (How Areas in Brisbane Influence Health and Activity) database. This database is a longitudinal, multilevel study of the health of mid-aged men and women in Brisbane City, Queensland, Australia.28 There have been four waves of data collected to date (2007, 2009, 2011, and 2013) although only the second and third waves are used in this study (hereto termed T1 and T2, respectively). The original survey was distributed in 2007 to 17,000 people in wave 1 and had a response rate of 65%. There was a response rate of 73% in wave 2 of the survey (2009) and a response rate of 68% in wave 3 (2011). HABITAT areas and participants were selected using a stratified two-stage design that involved identifying the primary area-level sampling unit (census collection districts or CCDs), selecting households with at least one resident aged 40 to 65 years from within CCDs, and then randomly selecting one person from each eligible household. First, the Australian Electoral Commission data were used to identify all households that had at least one person aged 40 to 65 years as at March 2007. Voting in Australia is compulsory for persons aged 18 years and over, so Australian Electoral Commission data provide near-complete coverage of the resident population. An average of 85 households per CCD was sampled using systematic without replacement probability proportional-to-size sampling, with size being defined as the number of households in each CCD with at least one person aged 40 to 65 years. The final stage of the sampling involved randomly selecting one person (of those aged 40 to 65 years) from each of the 17,000 households (85 × 200). For more information on the design, sampling, data collection, and analytical plan of HABITAT, please see Burton et al.28 Only participants who had reported data on the experience of conflict or stress at work in both T1 (2009) and T2 (2011), were employed at both time points, and had information on PMH and psychological distress, were included in the sample, resulting in 3291 participants. All individuals were employed (68.4% full-time, 20.5% part-time, and 11.2% casual) and 53.2% were female, aged on average 54.1 years (standard deviation, 6.25 years; range, 43 to 70 years.). The main outcome variable was nonspecific psychological distress as measured by the Kessler 6 (K6).29 The K6 is a widely used 6-item scale that screens for the presence of general symptomology of mental disorder. Items represent various behavioral, emotional, and psychophysiologic symptoms that are common to people with a variety of mental disorders, and therefore are indicative of mental ill-health.29 The K6 can discriminate with precision between cases and noncases of 12 Diagnostic Statistical ManualIV (version 4)/Composite International Diagnostic Interview mental disorders.29 We used a continuous scoring method. A high level of distress (probable serious mental illness) is indicated by a score of 13 or above and low distress (no probable serious mental illness) is indicated by a score of 0 to 12 (Australian Bureau of Statistics). The total K6 score, with each item being scored 0 to 4, is 24.29 The main exposure variable was the experience of serious work stress (termed “work stress” throughout). This was measured
Workplace Stress
through the question “in last 2 years have you experienced serious stress or conflict at work?” (0 = no; 1 = yes) at T1 and T2 to assess contemporaneous and historical effects on psychological distress. We also included several covariates, all measured at T1. The second key variable was PMH, measured by the Warwick Edinburgh Mental Well-being Scale.25 The Warwick Edinburgh Mental Well-being Scale measures positive feelings (eg, “I’ve been feeling relaxed”), positive functioning (eg, “I’ve been dealing with problems well”), and positive relationships (eg, “I’ve been feeling close to other people”). Confirmatory factor analysis showed all items loaded on a single factor with a Cronbach α of 0.91 in a population sample.25 The median score for the scale in a population sample is 51 with a total scale score of 70.25 Older adults typically have higher levels of PMH.25 Positive mental health was assessed as a possible effect modifier (moderator) of the relationship between exposure to serious work stress and psychological distress over time. Descriptive statistics were used to examine the characteristics of the sample by whether or not they have experienced work stress. We examined the relationship between continuous or multicategorical variables using pairwise regression. Correlations between dichotomous variables were examined using point biserial correlation. After this, multilevel ordinary least squared regression models were used to assess the relationship between work stress (at T1 and T2) and psychological distress (at T2). Multilevel models include both fixed and random parts to allow for circumstances when observations are clustered, such as in the present analysis when there are multiple time points for individuals who are nested within areas of residence (because of the nature of the HABITAT study design). Level 1 included the exposure and covariates, which varied by follow-up period. We also adjusted for time, as well as personal-level characteristics by including a dummy variable to control for time-invariant individuallevel factors. In the first model, we regressed exposure to work stress at T1 or T2 onto psychological distress at T2. We followed the same procedure in the second model but controlled for all other covariates. Third, an interaction term tested whether the relationship between work stress and psychological distress differed depending on the level of PMH. A likelihood ratio test was used to test the fit of models with interaction terms versus those without. We then calculated the predicted and graphed effects of work stress on psychological distress, controlling for all other covariates. All analyses were conducted using Stata version 12.1 (StataCorp, College Station, TX).20
RESULTS A description of the sample is presented in Table 1. The mean psychological distress score was 5.9 in those who experienced work stress and 3.7 among those who did not report work stress. The majority of the sample (approximately 70%) fell into the low level of psychological distress (scores up to 12). Those who had not experienced work stress had higher PMH scores than those who had. Those who had experienced work stress seemed to be of younger age than those who had not. Table 2 presents a correlation matrix of the main variables of interest. As shown, there was a correlation of 0.16 between T1 work stress and T2 psychological distress. Positive mental health, at both T1 and T2, was a strong predictor of psychological distress at T2 (r = −0.48, P < 0.001; and r = −0.60; P < 0.001, respectively). The relationship between experience of work stress at T1 and psychological distress at T2 (r = −0.16; P < 0.001) was less pronounced than the relationship between work stress at T2 and psychological distress at T2 (r = 0.21; P < 0.001). Similarly, PMH had a stronger contemporaneous association (ie, T2 and T2) with work stress (r = −0.61; P < 0.001) than T1 reports of PMH (r = −0.48; P < 0.001). Table 3 shows the results of bivariate and multivariate analyses for the main effects. Experience of work stress at T2 was associated with a small but significantly higher increase in psychological
C 2014 American College of Occupational and Environmental Medicine
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
815
JOEM r Volume 56, Number 8, August 2014
Page et al
TABLE 1. Sample Characteristics of Those With and Without Conflict/Stress at Work
T2 distress T1 PMH
TABLE 2. Correlation Matrix Between Stress/Conflict at Work (T1 and T2), Psychological Distress, and Positive Mental Health
Conflict/Stress at Work (T1) (n = 984)
No Conflict/Stress at Work (T1) (n = 2307)
Mean (SD)
Mean (SD)
5.98 (4.52) 48.60 (7.70) n (%)
3.72 (3.59) 51.80 (7.30) n (%)
Sex Male 459 (46.60) Female 525 (53.40) Age group (T1), yrs 40–44 148 (15.04) 45–49 297 (30.18) 50–54 277 (28.15) 55–59 173 (17.58) 60–65 85 (8.64) 66–70 4 (0.41) Long-term health condition (T1) No 393 (39.90) Yes 591 (60.10) Education (T1) High school 254 (25.81) Certificate or 280 (28.46) diploma Bachelors 326 (33.13) Postgraduate 124 (12.60) T1 household structure Alone 174 (17.77) Single 128 (13.07) Couple without 242 (24.72) dependents Couple with 435 (44.43) dependents
1. T2 distress 2. T1 conflict/stress (1 = no; 2 = yes) 3. T2 conflict/stress (1 = no; 2 = yes) 4. T2 PMH 5. T1 PMH
1104 (47.80) 1203 (52.20)
2
3
4
− 0.17* − 0.13*
0.64*
0.16* 0.21*
0.38*
− 0.60* − 0.48*
− 0.13* − 0.20*
*P < 0.001. T1 = T1, and T2 = T2. PMH, positive mental health.
250 (10.84) 595 (25.79) 552 (23.93) 512 (22.19) 347 (15.04) 51 (2.21)
Results of the interaction test were significant (x2 (1) = 146.6; P < 0.001), suggesting that the relationship between work stress and psychological distress depended on an individual’s level of PMH (Fig. 1). Individuals with higher PMH generally had the lowest levels of work-related psychological distress, whereas those with the lowest PMH had the highest levels of work-related distress. These results reinforce the main effect of conflict or stress at work on psychological distress. The interaction analysis shows that people who scored above about 58 on the PMH scale (the top 10%) and who experienced work stress did not experience significantly higher levels of distress as a result. This is demonstrated in the overlapping 95% CIs, which suggests that at very high levels of PMH, there is no difference between those who did and those who did not experience conflict at work and psychological distress. If there was a statistically significant difference between the groups, there would be no overlap between CIs between the groups. In saying this, Fig. 1 also demonstrates the increasing and incremental beneficial effects of PMH as there is a general reduction in distress as people score higher on the PMH measure.
1336 (57.90) 971 (42.10) 744 (32.25) 708 (30.69) 617 (26.74) 238 (10.32) 297 (12.97) 274 (11.97) 713 (31.14) 1006 (43.93)
DISCUSSION
PMH, positive mental health; SD, standard deviation.
distress at T2 in both models. In the bivariate model, T1 work stress was associated with a 1.70 increase in psychological distress (95% confidence interval [CI], 1.41 to 2.00; P < 0.001) whereas T2 distress was associated with a 2.27 increase (95% CI, 1.98 to 2.55; P < 0.001). The association between work stress reduced after adjusting for other relevant covariates in the multivariate model (right-hand columns). This shows that work stress measured at T1 was associated with a significant increase in T2 psychological distress (0.40; 95% CI, 0.14 to 0.66; P = 0.002), whereas T2 work stress had a stronger influence on T2 psychological distress (1.11; 95% CI, 0.86 to 1.37; P < 0.001). This suggests that work stress has a stronger contemporaneous effect on psychological distress. The association between PMH and psychological distress was significant in both bivariate and multivariate analyses. As can be seen in the multivariate analysis, T1 PMH was negatively associated with T2 distress (−0.07; 95% CI, −0.09 to 0.05; P < 0.001); T2 PMH had an even greater effect on T2 distress (−0.25; 95% CI, −0.26 to −0.23; P < 0.001). This also suggests a stronger contemporaneous effect of PMH. After adjusting for covariates, results show that age was negatively associated with psychological distress and having a long-term health condition was positively associated with distress (compared with no health condition). 816
1
In this study, we explore whether PMH might buffer workers from the adverse effects of workplace stress. Identifying the variables that might protect mental health aligns with an ongoing agenda in public health and occupational health that focuses on promoting PMH and other positive capacities18 and understanding how such capacities may be applied to protect or improve worker health.30,31 Our results show that the presence of positive psychological factors, such as those encapsulated by PMH, can protect or buffer an individual from suffering adverse mental health outcomes as a result of work stress. Our findings build on previous research finding that positive constructs can play a buffering role13,24 and add new insights by showing that these effects are also evident in a work context. This study also expands the remit of studies investigating the buffering or interactive hypothesis in stress research.12 High PMH may provide employees with an increased capacity to bounce back or recover from adversity.15,32 Furthermore, there were increasing incremental benefits of PMH in reducing psychological distress, particularly for individuals in the top 10% range of PMH—a score of 58 out of the 70. Notably, this score is higher than the population average of 51, reported by Tennant et al.25 in development of the scale. Experiencing serious work stress at T1 was strongly and negatively related to PMH at T2. This question needs further investigation as it may be that even individuals with the highest levels of PMH may become less resilient when exposed to long periods of work stress. When taken together, our results suggest that PMH—although a potentially important individual-level factor—does not negate the need to prevent
C 2014 American College of Occupational and Environmental Medicine
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
JOEM r Volume 56, Number 8, August 2014
Workplace Stress
TABLE 3. Effect of Stress/Conflict at Work (T1 and T2) on Psychological Distress (T2): Bivariate and Multivariate Analysis, Multilevel Regression Model Using the HABITAT Data Set Bivariate
T1 stress/conflict No Yes T2 stress/conflict No Yes T1 PMH
Multivariate
Coefficient (95% CI)
P
Coefficient (95% CI)
P
1 = reference 1.70 (1.41 to 2.00)
1 = reference
Workplace Stress What Is the Role of Positive Mental Health? Kathryn M. Page, PsyD, Allison J. Milner, PhD, Angela Martin, PhD, Gavin Turrell, PhD, Billie Giles-Corti, PhD, and Anthony D. LaMontagne, SciD
Objective: To examine whether positive mental health (PMH)—a positively focused well-being construct—moderates the job stress–distress relationship. Methods: Longitudinal regression was used to test two waves of matched, population-level data from a sample of older, working Australian adults (n = 3291) to see whether PMH modified the relationship between work stress and later psychological distress. Results: Time 1 work stress was positively associated with distress at both time points. Positive mental health was negatively associated with work stress at both time points. Positive mental health modified the impact of work stress on psychological distress. This effect only occurred for those with the highest levels of PMH. Conclusions: Positive mental health may help protect workers from the effect of workplace stress but only in a small proportion of the population. Therefore, to improve workplace mental health, workplaces need to both prevent stress and promote PMH.
O
ne in five Americans has a mental disorder that impairs their social, interpersonal, and occupational functioning (1-year prevalence rate),1 with similar rates observed across the developed world.2 The significant economic cost associated with this burden of disease, and in particular to employers, has been well documented.3–5 Research into processes that may prevent or reduce the effect of workplace stress on employee mental health is therefore important for workplaces, as well as governments and occupational and public health bodies. Although the etiology of mental health problems is complex and includes a wide range of individual factors, work-related stress is a significant contributor. Studies have consistently shown that job stress, whether measured as overall perceived stress or reported exposure to psychosocial stressors, can impair mental health.5,6 Szeto and Dobson5 analyzed overall perceived levels of work stress in a nationally representative sample of more than 10,000 Canadian employees. Those who reported the highest level of stress were approximately three times more likely to have been treated for an emotional From the McCaughey VicHealth Centre for Community Well-being (Drs Page, Milner, Giles-Corti, and LaMontagne), Melbourne School of Population and Global Health, the University of Melbourne; the Tasmanian School of Business and Economics (Dr Martin), University of Tasmania, Hobart; School of Public Health and Social Work (Dr Turrell), Queensland University of Technology, Brisbane; and Population Health Strategic Research Centre (Drs Page and LaMontagne), School of Health and Social Development, Deakin University, Melbourne, Australia. Dr Turrell is supported by an Australian National Health and Medical Research Council (NHMRC) Senior Research Fellowship (#1003710). Dr LaMontagne is supported by Centre grant funding from the Victorian Health Promotion Foundation, Melbourne, Australia (#15732). Dr Giles-Corti is supported by an NHMRC Principal Research Fellowship (#1004900) and Centre grant funding from the Victorian Health Promotion Foundation, Melbourne, Australia (#15732). Drs Page and Milner are supported by an NHMRC Capacity Building grant (#546248). The authors declare no conflicts of interest. Address correspondence to: Kathryn M. Page, PsyD, Work, Health and Wellbeing Unit, Deakin Population Health Strategic Research Centre, Faculty of Health, Deakin University, Melbourne Burwood campus, 221 Burwood Hwy, Burwood, Victoria 3125, Australia ([email protected]). C 2014 by American College of Occupational and Environmental Copyright Medicine DOI: 10.1097/JOM.0000000000000230
814
or mental health problem in the past 12 months and approximately 2.4 times more likely to be diagnosed with a mood and/or anxiety disorder than their nonstressed counterparts. This cross-sectional finding is consistent with systematic reviews6,7 and longitudinal and prospective studies examining the link between stressors and mental health.8,9 For example, Stansfeld and Candy’s6 systematic review and meta-analysis found that low decision latitude, low social support, high psychological demands, job strain (high demand/low control), effort–reward imbalance (high effort/low reward), and job insecurity—all well-known psychosocial job stressors—were all predictive of common mental problems. Particularly strong associations were found between job strain and effort–reward imbalance and the onset of common mental health problems. These findings are consistent with the job demand–control and effort–reward imbalance models.10,11
IDENTIFYING FACTORS THAT PROTECT WORKER MENTAL HEALTH Identifying the factors that may protect workers from experiencing work-related mental health problems is an important area of ongoing research. To date, research has identified a number of work and individual-level moderators of the stressor–strain relationship including job control and supervisor support.12 Identifying such moderators is paramount to the buffering or interactive hypothesis in job stress research, which predicts that certain factors (typically job control or social support) can moderate the relationship between job stressors and adverse psychological outcomes.12,13 The emergence of positive psychology has opened the door to a wider array of factors that may moderate the relationship between work and mental (ill) health that have not yet been investigated within a job stress framework. Positive psychology is the study of “the conditions and processes that contribute to the flourishing or optimal functioning of people, groups, and institutions.”14 An application of positive psychology (also called positive organizational behavior15 ) in the work context includes identifying and enhancing worker strengths or positive capabilities to facilitate improved mental health outcomes. Here, worker strengths and positive capabilities refer to adaptable capacities and strengths such as psychological capital16,17 or positive psychological well-being.18 An increasing body of literature attests to the benefits of strengths or positive capabilities on both health- and work-related constructs19–22 and, in particular, how the presence of such factors, for example, optimism, positive emotion, and positive or supportive relationships, can protect and/or enhance mental and physical health.23,24 In this article, we look at whether positive mental health (PMH)—a broad construct that encompasses several positive and protective elements25 —can protect an employee from the ill-effects of stress on mental health, using two waves of population-level data. Positive mental health is a positively focused construct that captures the degree to which individuals are flourishing in life, that is, reporting high levels of positive emotional, social, and psychological well-being.23,25 Positive mental health should not be seen as the direct antithesis to psychological ill-being; in fact, the two are only moderately correlated and factor into two clear and distinct dimensions.26,27 Studies have also found that having high levels of JOEM r Volume 56, Number 8, August 2014
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
JOEM r Volume 56, Number 8, August 2014
PMH plus low levels of mental ill-health (sometimes called complete mental health) predicts work- and health-related outcomes over and above that explained by having low levels of mental illness alone.26 Therefore, studies examining the relationship between these variables, alongside work-related exposures, including workplace stress, may provide a new and potentially useful contribution to the occupational health literature. In this study, we look at the potential for PMH to moderate the relationship between workplace stress and conflict on later psychological distress. Specifically, we hypothesize that employees with higher levels of PMH will have lower levels of psychological distress associated with work stress than those with lower levels of PMH. That is, PMH will buffer or protect an individual from the negative impact of work stress.
METHODS This article used two waves of data from the Australian HABITAT (How Areas in Brisbane Influence Health and Activity) database. This database is a longitudinal, multilevel study of the health of mid-aged men and women in Brisbane City, Queensland, Australia.28 There have been four waves of data collected to date (2007, 2009, 2011, and 2013) although only the second and third waves are used in this study (hereto termed T1 and T2, respectively). The original survey was distributed in 2007 to 17,000 people in wave 1 and had a response rate of 65%. There was a response rate of 73% in wave 2 of the survey (2009) and a response rate of 68% in wave 3 (2011). HABITAT areas and participants were selected using a stratified two-stage design that involved identifying the primary area-level sampling unit (census collection districts or CCDs), selecting households with at least one resident aged 40 to 65 years from within CCDs, and then randomly selecting one person from each eligible household. First, the Australian Electoral Commission data were used to identify all households that had at least one person aged 40 to 65 years as at March 2007. Voting in Australia is compulsory for persons aged 18 years and over, so Australian Electoral Commission data provide near-complete coverage of the resident population. An average of 85 households per CCD was sampled using systematic without replacement probability proportional-to-size sampling, with size being defined as the number of households in each CCD with at least one person aged 40 to 65 years. The final stage of the sampling involved randomly selecting one person (of those aged 40 to 65 years) from each of the 17,000 households (85 × 200). For more information on the design, sampling, data collection, and analytical plan of HABITAT, please see Burton et al.28 Only participants who had reported data on the experience of conflict or stress at work in both T1 (2009) and T2 (2011), were employed at both time points, and had information on PMH and psychological distress, were included in the sample, resulting in 3291 participants. All individuals were employed (68.4% full-time, 20.5% part-time, and 11.2% casual) and 53.2% were female, aged on average 54.1 years (standard deviation, 6.25 years; range, 43 to 70 years.). The main outcome variable was nonspecific psychological distress as measured by the Kessler 6 (K6).29 The K6 is a widely used 6-item scale that screens for the presence of general symptomology of mental disorder. Items represent various behavioral, emotional, and psychophysiologic symptoms that are common to people with a variety of mental disorders, and therefore are indicative of mental ill-health.29 The K6 can discriminate with precision between cases and noncases of 12 Diagnostic Statistical ManualIV (version 4)/Composite International Diagnostic Interview mental disorders.29 We used a continuous scoring method. A high level of distress (probable serious mental illness) is indicated by a score of 13 or above and low distress (no probable serious mental illness) is indicated by a score of 0 to 12 (Australian Bureau of Statistics). The total K6 score, with each item being scored 0 to 4, is 24.29 The main exposure variable was the experience of serious work stress (termed “work stress” throughout). This was measured
Workplace Stress
through the question “in last 2 years have you experienced serious stress or conflict at work?” (0 = no; 1 = yes) at T1 and T2 to assess contemporaneous and historical effects on psychological distress. We also included several covariates, all measured at T1. The second key variable was PMH, measured by the Warwick Edinburgh Mental Well-being Scale.25 The Warwick Edinburgh Mental Well-being Scale measures positive feelings (eg, “I’ve been feeling relaxed”), positive functioning (eg, “I’ve been dealing with problems well”), and positive relationships (eg, “I’ve been feeling close to other people”). Confirmatory factor analysis showed all items loaded on a single factor with a Cronbach α of 0.91 in a population sample.25 The median score for the scale in a population sample is 51 with a total scale score of 70.25 Older adults typically have higher levels of PMH.25 Positive mental health was assessed as a possible effect modifier (moderator) of the relationship between exposure to serious work stress and psychological distress over time. Descriptive statistics were used to examine the characteristics of the sample by whether or not they have experienced work stress. We examined the relationship between continuous or multicategorical variables using pairwise regression. Correlations between dichotomous variables were examined using point biserial correlation. After this, multilevel ordinary least squared regression models were used to assess the relationship between work stress (at T1 and T2) and psychological distress (at T2). Multilevel models include both fixed and random parts to allow for circumstances when observations are clustered, such as in the present analysis when there are multiple time points for individuals who are nested within areas of residence (because of the nature of the HABITAT study design). Level 1 included the exposure and covariates, which varied by follow-up period. We also adjusted for time, as well as personal-level characteristics by including a dummy variable to control for time-invariant individuallevel factors. In the first model, we regressed exposure to work stress at T1 or T2 onto psychological distress at T2. We followed the same procedure in the second model but controlled for all other covariates. Third, an interaction term tested whether the relationship between work stress and psychological distress differed depending on the level of PMH. A likelihood ratio test was used to test the fit of models with interaction terms versus those without. We then calculated the predicted and graphed effects of work stress on psychological distress, controlling for all other covariates. All analyses were conducted using Stata version 12.1 (StataCorp, College Station, TX).20
RESULTS A description of the sample is presented in Table 1. The mean psychological distress score was 5.9 in those who experienced work stress and 3.7 among those who did not report work stress. The majority of the sample (approximately 70%) fell into the low level of psychological distress (scores up to 12). Those who had not experienced work stress had higher PMH scores than those who had. Those who had experienced work stress seemed to be of younger age than those who had not. Table 2 presents a correlation matrix of the main variables of interest. As shown, there was a correlation of 0.16 between T1 work stress and T2 psychological distress. Positive mental health, at both T1 and T2, was a strong predictor of psychological distress at T2 (r = −0.48, P < 0.001; and r = −0.60; P < 0.001, respectively). The relationship between experience of work stress at T1 and psychological distress at T2 (r = −0.16; P < 0.001) was less pronounced than the relationship between work stress at T2 and psychological distress at T2 (r = 0.21; P < 0.001). Similarly, PMH had a stronger contemporaneous association (ie, T2 and T2) with work stress (r = −0.61; P < 0.001) than T1 reports of PMH (r = −0.48; P < 0.001). Table 3 shows the results of bivariate and multivariate analyses for the main effects. Experience of work stress at T2 was associated with a small but significantly higher increase in psychological
C 2014 American College of Occupational and Environmental Medicine
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
815
JOEM r Volume 56, Number 8, August 2014
Page et al
TABLE 1. Sample Characteristics of Those With and Without Conflict/Stress at Work
T2 distress T1 PMH
TABLE 2. Correlation Matrix Between Stress/Conflict at Work (T1 and T2), Psychological Distress, and Positive Mental Health
Conflict/Stress at Work (T1) (n = 984)
No Conflict/Stress at Work (T1) (n = 2307)
Mean (SD)
Mean (SD)
5.98 (4.52) 48.60 (7.70) n (%)
3.72 (3.59) 51.80 (7.30) n (%)
Sex Male 459 (46.60) Female 525 (53.40) Age group (T1), yrs 40–44 148 (15.04) 45–49 297 (30.18) 50–54 277 (28.15) 55–59 173 (17.58) 60–65 85 (8.64) 66–70 4 (0.41) Long-term health condition (T1) No 393 (39.90) Yes 591 (60.10) Education (T1) High school 254 (25.81) Certificate or 280 (28.46) diploma Bachelors 326 (33.13) Postgraduate 124 (12.60) T1 household structure Alone 174 (17.77) Single 128 (13.07) Couple without 242 (24.72) dependents Couple with 435 (44.43) dependents
1. T2 distress 2. T1 conflict/stress (1 = no; 2 = yes) 3. T2 conflict/stress (1 = no; 2 = yes) 4. T2 PMH 5. T1 PMH
1104 (47.80) 1203 (52.20)
2
3
4
− 0.17* − 0.13*
0.64*
0.16* 0.21*
0.38*
− 0.60* − 0.48*
− 0.13* − 0.20*
*P < 0.001. T1 = T1, and T2 = T2. PMH, positive mental health.
250 (10.84) 595 (25.79) 552 (23.93) 512 (22.19) 347 (15.04) 51 (2.21)
Results of the interaction test were significant (x2 (1) = 146.6; P < 0.001), suggesting that the relationship between work stress and psychological distress depended on an individual’s level of PMH (Fig. 1). Individuals with higher PMH generally had the lowest levels of work-related psychological distress, whereas those with the lowest PMH had the highest levels of work-related distress. These results reinforce the main effect of conflict or stress at work on psychological distress. The interaction analysis shows that people who scored above about 58 on the PMH scale (the top 10%) and who experienced work stress did not experience significantly higher levels of distress as a result. This is demonstrated in the overlapping 95% CIs, which suggests that at very high levels of PMH, there is no difference between those who did and those who did not experience conflict at work and psychological distress. If there was a statistically significant difference between the groups, there would be no overlap between CIs between the groups. In saying this, Fig. 1 also demonstrates the increasing and incremental beneficial effects of PMH as there is a general reduction in distress as people score higher on the PMH measure.
1336 (57.90) 971 (42.10) 744 (32.25) 708 (30.69) 617 (26.74) 238 (10.32) 297 (12.97) 274 (11.97) 713 (31.14) 1006 (43.93)
DISCUSSION
PMH, positive mental health; SD, standard deviation.
distress at T2 in both models. In the bivariate model, T1 work stress was associated with a 1.70 increase in psychological distress (95% confidence interval [CI], 1.41 to 2.00; P < 0.001) whereas T2 distress was associated with a 2.27 increase (95% CI, 1.98 to 2.55; P < 0.001). The association between work stress reduced after adjusting for other relevant covariates in the multivariate model (right-hand columns). This shows that work stress measured at T1 was associated with a significant increase in T2 psychological distress (0.40; 95% CI, 0.14 to 0.66; P = 0.002), whereas T2 work stress had a stronger influence on T2 psychological distress (1.11; 95% CI, 0.86 to 1.37; P < 0.001). This suggests that work stress has a stronger contemporaneous effect on psychological distress. The association between PMH and psychological distress was significant in both bivariate and multivariate analyses. As can be seen in the multivariate analysis, T1 PMH was negatively associated with T2 distress (−0.07; 95% CI, −0.09 to 0.05; P < 0.001); T2 PMH had an even greater effect on T2 distress (−0.25; 95% CI, −0.26 to −0.23; P < 0.001). This also suggests a stronger contemporaneous effect of PMH. After adjusting for covariates, results show that age was negatively associated with psychological distress and having a long-term health condition was positively associated with distress (compared with no health condition). 816
1
In this study, we explore whether PMH might buffer workers from the adverse effects of workplace stress. Identifying the variables that might protect mental health aligns with an ongoing agenda in public health and occupational health that focuses on promoting PMH and other positive capacities18 and understanding how such capacities may be applied to protect or improve worker health.30,31 Our results show that the presence of positive psychological factors, such as those encapsulated by PMH, can protect or buffer an individual from suffering adverse mental health outcomes as a result of work stress. Our findings build on previous research finding that positive constructs can play a buffering role13,24 and add new insights by showing that these effects are also evident in a work context. This study also expands the remit of studies investigating the buffering or interactive hypothesis in stress research.12 High PMH may provide employees with an increased capacity to bounce back or recover from adversity.15,32 Furthermore, there were increasing incremental benefits of PMH in reducing psychological distress, particularly for individuals in the top 10% range of PMH—a score of 58 out of the 70. Notably, this score is higher than the population average of 51, reported by Tennant et al.25 in development of the scale. Experiencing serious work stress at T1 was strongly and negatively related to PMH at T2. This question needs further investigation as it may be that even individuals with the highest levels of PMH may become less resilient when exposed to long periods of work stress. When taken together, our results suggest that PMH—although a potentially important individual-level factor—does not negate the need to prevent
C 2014 American College of Occupational and Environmental Medicine
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
JOEM r Volume 56, Number 8, August 2014
Workplace Stress
TABLE 3. Effect of Stress/Conflict at Work (T1 and T2) on Psychological Distress (T2): Bivariate and Multivariate Analysis, Multilevel Regression Model Using the HABITAT Data Set Bivariate
T1 stress/conflict No Yes T2 stress/conflict No Yes T1 PMH
Multivariate
Coefficient (95% CI)
P
Coefficient (95% CI)
P
1 = reference 1.70 (1.41 to 2.00)
1 = reference
