Occupational stress, job characteristics, coping, and the mental health of nurses

Transcription

1 1 British Journal of Health Psychology (2011) C 2011 The British Psychological Society The British Psychological Society Occupational stress, job characteristics, coping, and the mental health of nurses G. Mark and A. P. Smith School of Psychology, Cardiff University, UK Objectives. This study investigated the relationships between job characteristics and coping in predicting levels of anxiety and depression in nurses. The study was based on current theories of occupational stress, and predictors included job demands, social support, decision authority and skill discretion control, effort, over-commitment, rewards, and ways of coping. It was predicted that job demands, extrinsic effort, over-commitment, and negative coping behaviours would be positively associated with depression and anxiety, and social support, rewards, decision authority, skill discretion control, and positive coping would be negatively associated with depression and anxiety. Methods. Participants were 870 nurses, who responded to a bulk mail sent randomly to 4,000 nurses from the south of England. Results. The results showed that job demands, extrinsic effort, and over-commitment were associated with higher levels of anxiety and depression. Social support, rewards, and skill discretion were negatively associated with mental health problems. Few interactions were found between the variables. Coping behaviours significantly added to the explanation of variance in anxiety and depression outcomes, over and above the use of demand control support, and effort reward factors alone. Conclusion. The results from the study demonstrated the importance of coping factors in work-stress research, in accordance with the multi-factorial premise of transactional stress models. It is argued that multi-factor research is needed to help develop effective organizational interventions. Stress in health professionals Research has shown that health professionals are a group at significant risk from the negative effects of stressful workplaces (Kirkcaldy & Martin, 2000; Tyler & Cushway, 1998). Nurses are particularly at risk from stress-related problems, with high rates of turnover, absenteeism, and burnout (Clegg, 2001; Kirkcaldy & Martin, 2000). Calnan, Wainwright, Forsythe, Wall, and Almond (2001) administered the GHQ (General Health Questionnaire) to health service staff and found that 27% of all hospital staff were classified as suffering stress and mental ill health, compared to between 14% and 18% of the general population. There is also evidence that nurses have higher than normal rates Correspondence should be addressed to Professor A. P. Smith, Centre for Occupational and Health Psychology, School of Psychology, Cardiff University, 63 Park Place, Cardiff, UK ( SmithAP@cardiff.ac.uk). DOI: /j x

2 2 G. Mark and A. P. Smith of physical illness, mortality, and psychiatric admissions (Kirkcaldy & Martin, 2000). Figures from 1979 to 1983 show that suicide rates for nurses were significantly higher than the national average, and life expectancy for nurses was approximately 72 years, only 1 year more than miners (Clegg, 2001). Sickness absence in the health service is about 4% (NHS Information Centre, 2011) and the Boorman review (Department of Health, 2009) estimated that the direct cost of staff sickness was 1.7 billion. An analysis by the Audit Commission (2011) found that nursing staff had above average rates of absenteeism (4.82%). Another cost closely associated with sickness absence but much harder to quantify is presenteeism. The Boorman review interim report found that 71% of nurses reported presenteeism compared to 45% of staff in an age matched sample in corporate services. Psychosocial stressors in nursing Nurses can be exposed on a daily basis to a large number of potent stressors, including conflict with physicians, discrimination, high workload, and dealing with death, patients, and their families (French, Lenton, Walters, & Eyles, 2000). McVicar (2003) states that many such situations encountered by nurses at work have a high cost in emotional labour. Lambert, Lambert, and Ito (2004) note that most research on nursing stress has taken place in the United Kingdom and United States, and showed that Japanese nurses can also suffer from negative mental health. Shift working and bullying are also common stressors in nursing, and both are thought to be related to numerous health problems (Ball, Pike, Cuff, Mellor-Clark, & Connell, 2002; Boggild & Knuttson, 1999). Demands control support and effort reward It is clear that some of the working conditions that characterize nursing may be implicated in stress-related issues. Two of the most influential theories commonly used in studying work-related stress are the demands control support model (Karasek & Theorell, 1990) and the effort reward imbalance (ERI) model (Siegrist, 1996). Both models have been found to predict many physical and psychological health outcomes, including mortality, heart disease, and depression (Van Der Doef & Maes, 1999) and have also been used in nursing populations (de Rijk, le Blanc, Schaufeli, & De Jonge, 1998; Weyers et al., 2006). The DCS (Demands-Control-Support) model predicts that those exposed to high levels of psychological demand, and low levels of social support, and job control (from the sub-factors of decision authority and skill discretion) are likely to suffer negative health outcomes. Karasek (1979) proposed an interaction between demands and control, so that when demands are high and control is low, a high-strain situation develops, and this is particularly likely to lead to negative health outcomes. High control (sometimes summed from skill discretion and decision authority, although also these variables can be considered individually) is also proposed to buffer the effect of high demands on health outcomes. This model may be very appropriate for nursing samples, because a lack of social support as well as excessive demands are common in nursing (Muncer, Taylor, Green, & McManus, 2001) and control may well vary by occupational grade. The ERI model (Siegrist, 1996) is also popular and influential in work-stress research. Based on the concept of reciprocity, the ERI model proposes that high levels of workrelated effort should be matched by high levels of reward (economic, recognition, promotion prospects, job security, etc). It is proposed that if efforts (external demands or internal motivations) are high, but rewards are low, then strain and negative health outcomes are likely to ensue. Like the DCS model, the relationship between efforts and

3 Stress and nurses 3 rewards is proposed to be an interaction, so that high levels of reward buffer high levels of effort (Peter & Siegrist, 1999). The ERI model further predicts that ERI at work will be experienced more frequently by those who are overly committed to their work. Intrinsic effort and over-commitment are often used as equivalent terms, although overcommitment is more likely to be used to represent a personality trait. This is the term that will be used in the current paper. The ERI model may be suited to studying work-related stress in nurses, as there is much evidence that nursing is a demanding occupation and thus requires effort, and levels of pay in newly qualified nurses may be lower than other high-stress occupational groups, such as teachers and police officers (Demerouti, Bakker, Nachreiner, & Schaufeli, 2000). Despite the popularity of the above two models, they are largely focused on job characteristics or environmental factors (Cox, Griffiths, & Rial-Gonzalez, 2000) and generally fail to take account of individual factors (over-commitment from ERI being an exception). These models cannot readily explain, for example, how different individuals exposed to the same levels of stressors, may suffer different health outcomes (Perrewe & Zellars, 1999). It should be noted that despite purporting to measure environmental job factors (e.g. workload, levels of social support etc.) these models use subjective individual ratings of job characteristics, rather than more objective measures (such as supervisor ratings or objective workload measures). Transactional theory and coping As well as the job characteristics of nursing, it is important to understand how differences between individuals may affect how they deal with stressors at work (Cox & Ferguson, 1991). Transactional theories of stress (e.g. Folkman & Lazarus, 1980), place emphasis not just on job characteristics, but also on subjective perceptions of stressors, and individual differences in ways of coping, viewing problems, past experience, personality-type etc. All these may be important in informing and affecting the workplace individual stress interaction (Cox & Ferguson, 1991; Moos & Holahan, 2003; Moos, Holahan, & Beutler, 2003; Penley, Tomaka, & Wiebe, 2002). In Folkman and Lazarus model (1980), individuals are proposed to appraise environmental stressors, including their level of potential threats and costs (primary appraisal), then to make potential plans to deal with stressors using known coping methods and past experience (secondary appraisal), and then to initiate coping. Coping has been described as any cognitive or behavioural efforts to manage, minimize, or tolerate events that individuals perceive as potentially threatening to their well-being (Folkman, Lazarus, Gruen, & DeLongis, 1986). Coping does not imply success in dealing with situations and coping responses to stressors can also be maladaptive. Methods may include problem solving, self-blame, escape/avoidance, wishful thinking, seeking advice and support, etc. (Folkman & Lazarus, 1980). Therefore coping occupies a dual role, both as a process following on from appraisal, but also as an individual difference variable, when people exhibit patterns of coping behaviours (carried out during the coping stage of transactional theory) that many suggest may be stable, or slowly changing over time (Folkman & Lazarus, 1980). For the purposes of this study (as in work by others including Folkman and Lazarus), coping is treated as an individual difference variable. Folkman et al. (1986) claim that problem-focused forms of coping (so called positive coping types) are likely to be associated with lower levels of negative health outcomes, and that coping of an emotional-focused (or negative) type, such as self-blame, wishful

4 4 G. Mark and A. P. Smith thinking, or escape/avoidance are likely to be associated with increased negative health. For example, Healy and McKay (2000) found that avoidance coping predicted poor mental health in nurses, and problem solving coping was positively related to satisfaction and health. Evidence is more mixed for the relationship between seeking advice and health outcomes. Rationale Nurses have been selected as the population for the current study due to the complex array of stressors that they face, and the high levels of negative mental and physical health they suffer from (Kirkcaldy & Martin, 2000). This is particularly important given the funding and staffing challenges facing the UK health service. The traditional models of demands control support, and ERI were tested simultaneously in this population, to see how much each contributes to the variance in mental health outcomes. Ways of coping were also investigated, to see how much additional variance this factor explains over the use of traditional job characteristics variables alone. Coping is also central in transactional stress models (Folkman & Lazarus, 1980). McVicar (2003) and Kirkcaldy and Martin (2000) suggest that there is a need for more understanding of how individual variation in perceptions and reactions to stressors in nurses affect health outcomes. The relationships between the above variables were investigated in a nursing population. It was hoped that assessing the relative importance of these factors, and finding any interactions between them in the prediction of anxiety and depression, will be useful for supporting the view that individual differences can add to the study of work-related ill health, as suggested by transactional models. It was also hoped that such information could help to provide empirical support for potential interventions to combat work stress. Hypotheses Hypothesis one predicted that positive coping behaviours (problem-focused coping) would be associated negatively with depression and anxiety in nurses, and negative coping behaviours (self-blame, wishful thinking, escape/avoidance) would be associated positively with anxiety and depression. No prediction was made about the relationship between seeking advice and outcomes, due to mixed evidence. Hypothesis two predicted that job demands would be positively associated with anxiety and depression in nurses, and skill discretion, decision authority, and social support would be negatively associated with depression and anxiety. A second part of the Hypothesis (2a) predicted that control variables and social support would significantly interact with the effect of demands in predicting anxiety and depression scores. Hypothesis three predicted that extrinsic effort and over-commitment would be associated positively with depression and anxiety in nurses, and intrinsic reward would be negatively associated with anxiety and depression. It was also predicted that rewards would significantly interact with the effect of over-commitment and extrinsic effort in predicting anxiety and depression scores (Hypothesis 3a). Hypothesis four predicted that there would be significant interactions between negative job characteristics (high job demands, extrinsic efforts) and positive coping behaviours (problem-focused coping) so that the latter would moderate the effects of negative job characteristics on mental health outcomes.

5 Stress and nurses 5 Hypothesis five predicted that coping, efforts, rewards, demands, control variables, and support would account for a significant amount of the variance in anxiety and depression scores in nurses. It was also predicted that ways of coping would significantly add to the explained variance in outcomes, over and above use of DCS and ERI alone (Hypothesis 5a). Method Participants The participants in this study were a sample of 870 nurses from all occupational grades and roles employed in the UK health service. Four thousand nurses were selected at random by the UK Royal College of Nursing, and these individuals were mailed a request for participants for a study into health and safety at work, with a focus on stress and work pressures. Eight hundred and seventy nurses responded (22% response rate). Those who responded were 790 women and 80 men (mean age = 44.84, SD = 8.8). Participants were treated in accordance with BPS ethical guidelines for treatment of participants (British Psychological Society, 2004), and ethical approval was provided by the Cardiff University School of Psychology ethics committee. Materials A 31-page questionnaire booklet was produced, containing an instruction page that informed participants as to the purposes of the study, their right to withdraw, and the anonymous treatment of data. The booklet also contained questions on demographic data, work type (shift/contract/permanent), and five main questionnaires. The 21-item version of the ERI Questionnaire (Siegrist, 1996) was as used in the Whitehall II Study (Kuper, Singh-Manoux, Siegrist, & Marmot, 2002). Three subscales were measured: over-commitment (internal motivations, e.g., over-commitment to work), extrinsic effort (external pressures), and internal reward (adequate rewards). Participants responded on a 4-point Likert scale indicating to what extent (if experienced) they found the suggested work situations distressing. Mean scores were converted to percentages for each sub-factor. This was done for other scales (as below) to standardize scores across different variables, and to make results easier to interpret. Cronbach s scores were calculated as.80 for the over-commitment subscale,.74 for extrinsic effort, and.84 for intrinsic reward. A 27-item version of the Job Content Questionnaire (JCQ; Karasek et al., 1988) was used. Four subscales were measured: job demands (workload, time pressure); decision authority (control over decisions); skill discretion (opportunity to use skills); and levels of social support. Participants responded how often they experienced the suggested situations at work on a 4-point Likert scale. Scores were converted to percentages, and Cronbach s scores were calculated as.85 for the social support subscale,.81 for decision authority,.68 for job demands, and.68 for skill discretion. The Ways of Coping Checklist (WCCL; Folkman and Lazarus, 1980) is a well-known 68-item scale used to assess coping behaviours, and this study used Vitaliano, Russo, Carr, Maiuro, and Becker s (1985) 42-item revised version of this scale. The 42 items are used to assess five factors, labelled: problem-focused coping ( =.84); seek advice ( =.82); self-blame ( =.88); wishful thinking ( =.84); and escape/avoidance ( =.76). Participants were asked to think of a recent stressful work experience and to indicate

6 6 G. Mark and A. P. Smith on a 4-point Likert scale how often they used each of the suggested behaviours. Final scores were converted to percentages of maximum scores. The Hospital Anxiety and Depression Scale (HADS; Zigmond & Snaith, 1983) is a 14-item scale that aims to measure self-reported anxiety and depression using two subscales. Participants respond on a 4-point Likert scale how often they have felt or experienced the suggested anxious or depressed feelings or situations in the past week. Reliability scores were calculated as.84 for anxiety, and.78 for depression. Total scores were calculated of 21 for each subscale, with 11 or more considered as a potential clinical case, in line with recommendations by Zigmond and Snaith (1983). Procedure Those who responded with interest to the request for participants were sent a questionnaire package containing the measures as described above, along with a freepost return envelope. Instructions were given where appropriate as specified by (or adapted from as relevant) the original questionnaire authors. Completion time was estimated at min. It was emphasized that results were anonymous, and that any concerns about mental health that came about as a result of completing the questionnaire, should be raised with the participant s doctor. Contact details were given for the research team, in case participants had any other queries or requests for information about the research. Analysis Results were analysed to test the stated hypotheses, with a variety of statistical techniques using the computer statistics package SPSS 13. Descriptive statistics were first used to compare participants for anxiety and depression scores against the clinical cut-off scores for the HADS, and Pearson correlations were used to examine the relationships between anxiety and depression, and the predictor variables (coping, attributions, and job characteristics). A range of multiple regressions was also carried out to investigate the effects of the various potential predictors of depression and anxiety. Multiple regressions were used so that variables could be entered simultaneously, including those from different theories and models, to see their relative predictive power (using standardized beta weights). Regressions were carried out in order of increasingly complexity to test the stated hypotheses. The regressions were as follows: coping variables (from WCCL) against anxiety; coping variables against depression; JCQ variables against anxiety plus interaction effects (four regressions, one main effects, and three with one interaction term); JCQ variables against depression plus interactions (four regressions as above); ERI variables (intrinsic and extrinsic efforts, intrinsic rewards) against anxiety plus interactions (three regressions, one main effects, two with one interaction term); ERI variables against depression with interactions (three regressions as above); all JCQ and ERI variables entered simultaneously against anxiety; all JCQ and ERI variables against depression. Finally, all predictor variables (IVs Independent Variables) from the WCCL (coping) JCQ (demands and controls) and ERI (efforts and rewards) were entered against anxiety and depression, using hierarchical multiple regression, with JCQ variables in the first block using entry method, ERI variables in the second block using entry method, and WCCL variables in the third block using entry method. Additionally, a further set of regressions were conducted to test interactions between positive coping behaviours

7 Stress and nurses 7 Table 1. Levels of clinical anxiety and depression in nurses, and correlations of coping and job characteristics against anxiety and depression Percentage with Percentage with Percentage claim illness clinical anxiety clinical depression caused/worsened by scores scores work stress Nurses 26.3% 5.9% 44.8% HAD-A HAD-D Problem-focused coping Self-blame Wishful thinking Seek advice Escape/avoidance Job demands Social support Skill discretion Decision authority Extrinsic effort Over-commitment Intrinsic reward Note. p.05; p.01. and negative job characteristics in the prediction of anxiety and depression, however no significant interactions were found and results are not reported. It can be seen that the regressions above increased in complexity, moving from using predictor variables from single to multiple theoretical models. This was carried out so that earlier models could provide context to later ones, and to show that individual coping variables could be significant predictors even when more traditional job-characteristic variables were present. Some of the models include significant interaction effects, however those where significant interaction effects were not found (e.g., job demands by social support in anxiety, or any JCQ variables in predicting depression) are not included. Results Table 1 shows that for scores on the HADS, 26.3% of sampled nurses scored at clinical levels for anxiety (11 or more; Zigmond & Snaith, 1983). For depression scores, 5.9% of nurses scored above the clinical cut-off point. Additionally 44.8% of nurses indicated that they believed that they had suffered an illness in the past year that had been caused or made worse by stress at work. Table 1 also shows Pearson correlations between all independent variables and anxiety and depression. The table shows that negative coping characteristics, such as selfblame, escape/avoidance, and wishful thinking, show significant positive correlations with anxiety and depression, with correlations of between.28 and.48. Problem-focused coping has a small but significant negative correlation with depression. The table shows that negative job characteristics such as job demands and extrinsic effort correlate positively with anxiety and depression, as does over-commitment, whereas positive job characteristics such as skill discretion, decision authority, intrinsic

8 8 G. Mark and A. P. Smith Table 2. Regressions of coping against anxiety and depression Anxiety weight error weight Significance (Constant) Problem-focused coping Self-blame Seeks advice Wishful thinking Escape/avoidance Model: R =.492, R 2 =.242 F: Depression weight error weight Significance (Constant) Problem -focused coping Self-blame Seek advice Wishful thinking Escape/avoidance Model: R =.457, R 2 =.209 F: reward, and social support show significant negative correlations of between.26 and.57 with anxiety and depression. Table 2 shows two regressions where all coping variables (problem-focused coping, seeking advice, self-blame, wishful thinking, escape/avoidance) were regressed against anxiety and depression. Variables show similar associations with anxiety and depression as those in the correlations. Self-blame and escape/avoidance predict increased levels of anxiety. Problem-focused coping was not associated with significantly lower anxiety scores (however this relationship becomes significant if seeking advice is removed, suggesting there may be some collinearity between these variables). The above variables accounted for 24.2% of the variance in anxiety scores, and self-blame was the most important factor by standardized beta weight, followed by escape/avoidance and problem-focused coping. For the depression regression, self-blame and escape/avoidance had significant positive associations with depression, and problem-focused coping and seeking advice had significant negative associations with depression scores. Self-blame and escape/avoidance were again the most important predictors by standardized beta weight. These factors accounted for 20.9% of the variance in depression scores. Table 3 shows regressions of the JCQ variables of job demands, control (skill discretion and decision authority), and social support, against anxiety and depression. For anxiety, an initial regression showed that job demands were associated positively with anxiety, and social support and skill discretion negatively with anxiety. Three further anxiety regressions were then conducted, with these main effects, and entering interactions between job demands and decision authority, demands and social support, and demands and skill discretion. The interaction between decision authority and job demands was found to be significant. This relationship is shown in Figure 1, and shows that at low levels of job demands, anxiety levels are similar regardless of level of decision

9 Stress and nurses 9 Table 3. Regressions of job demands, control, and social support against anxiety and depression Anxiety weight error weight Significance (Constant) Job demands Social support Skill discretion Decision authority Job demands decision authority Model: R =.462, R 2 =.214 F: Depression weight Error weight Significance (Constant) Job demands Social support Skill discretion Decision authority Model: R =.473, R 2 =.224 F: authority. However, at high demands, anxiety scores are significantly higher in those with low decision authority. For the depression regression, job demands had a significant positive relationship with depression, and social support and skill discretion associated negatively with depression scores. No interactions were found for depression. For anxiety, job demands were the most important predictor by standardized beta weight, followed by social support and skill discretion. However for depression, social support was the most important predictor, with job demands and skill discretion of Anxiety Low Decision Authority High Decision Authority Low Job Demands High Job Demands Figure 1. Interaction of job demands and decision authority in predicting anxiety.

10 10 G. Mark and A. P. Smith Table 4. Regressions of intrinsic reward, extrinsic effort and over-commitment, against anxiety and depression Anxiety weight error weight Significance (Constant) Intrinsic reward Extrinsic effort Over-commitment Over-commitment intrinsic reward Model: R =.624, R 2 =.390 F: Depression weight error weight Significance (Constant) Intrinsic reward Extrinsic effort Over-commitment Model: R =.582, R 2 =.338 F: equal importance. The above variables accounted for 21.4% of the variance in anxiety, and 22.4% of the variance in depression scores. Table 4 shows regressions of intrinsic reward, extrinsic effort and over-commitment, against anxiety and depression. Intrinsic reward was negatively associated with anxiety and depression, extrinsic effort positively with anxiety and depression, and overcommitment was positively associated with depression. A significant interaction was found between over-commitment and intrinsic reward in predicting anxiety. Using the same method as described for JCQ variables and anxiety, the main-effect regression was calculated, and then two further regressions were carried out entering interaction effects (intrinsic and extrinsic efforts by rewards). Figure Anxiety 9 8 Low Intrinsic Reward High Intrinsic Reward Low over-commitment High over-commitment Figure 2. Interaction of over-commitment and intrinsic reward in predicting anxiety.

11 Stress and nurses 11 Table 5. Regressions of demands, control, social support, extrinsic effort, over-commitment and rewards against anxiety and depression Anxiety weight error weight Significance (Constant) Job demands Social support Skill discretion Decision authority Intrinsic reward Extrinsic effort Over-commitment Model: R =.643, R 2 =.413 F: Depression weight error weight Significance (Constant) Job demands Social support Skill discretion Decision Authority Intrinsic reward Extrinsic effort Over-commitment Model: R =.625, R 2 =.390 F: shows those individuals who exhibit low levels of over-commitment are less anxious when they feel more rewarded (thus respond well to rewards, and are more anxious when rewards are low). However for those who are over-committed, anxiety is much higher, and levels of reward make no difference to their anxiety levels. For anxiety, intrinsic reward was the most important variable by standardized beta weight (although in non-interaction regressions, over-commitment was the most important), and in depression, over-commitment was the most important. The above variables accounted for 39% of the variance in anxiety and 33.8% of the variance in depression scores (compared to 21.4% of anxiety and 22.4% in depression with demands, decision authority, skill discretion, and social support). Two more major sets of regressions were conducted. The first set (shown in Table 5) used entry method and entered all sub-factors from the DCS model together with those from the ERI model against anxiety and depression. Over-commitment and extrinsic effort were significantly positively associated with anxiety and depression scores. Job demands were significantly positively associated with anxiety, and social support, skill discretion, and intrinsic reward were all significantly negatively associated with anxiety and depression scores. For both anxiety and depression, over-commitment was the most important predictor by standardized beta weight, followed by skill discretion in anxiety, and social support and skill discretion in depression. No significant interactions were found. These factors accounted for 41.3% of the variance in overall anxiety scores, and 39% of the variance in depression scores. The unique variance added over demands, decision

12 12 G. Mark and A. P. Smith Table 6. Regressions of ways of coping, demands, control, support, extrinsic effort, over-commitment and rewards, against anxiety and depression Anxiety weight error weight Significance (Constant) Job demands Social support Skill discretion Decision authority Intrinsic reward Extrinsic effort Over-commitment Problem-focused coping Seek advice Self-blame Wishful thinking Escape/avoidance Model: R =.694, R 2 =.482 F: Depression weight error weight Significance (Constant) Job demands Social support Skill discretion Decision authority Intrinsic reward Extrinsic effort Over-commitment Problem-focused coping Seek advice Self blame Wishful thinking Escape/avoidance Model: R =.659, R 2 =.435 F: authority and skill discretion, and social support, by including efforts and rewards, is 18.9% in anxiety (or almost double the total amount of variance explained) and 16.6% in depression (a three-quarter increase in variance explained). The final regressions (Table 6) were hierarchical multiple regressions against anxiety and depression, where job demands, social support, and skill discretion and authority were entered by entry method in block one, intrinsic reward, extrinsic effort, and overcommitment were entered by entry method in block two, and problem focused coping, seeking advice, self-blame, wishful thinking, and escape/avoidance, were entered in block three. The direction of association between significant variables and outcomes are the same for almost all variables as those in previous regressions: problem-focused coping, social support, skill discretion, and intrinsic reward were significantly negatively associated with anxiety and depression scores, and self-blame and over-commitment were both significantly positively associated with anxiety and depression. Additionally,

13 Stress and nurses 13 escape/avoidance and extrinsic effort associated positively with depression scores, and seeking advice and job demands associated negatively with anxiety scores. Only decision authority, extrinsic effort, wishful thinking, and escape/avoidance, were non-significant predictors of anxiety, and only job demands, decision authority, wishful thinking, and seeking advice, were non-significant predictors of depression. By standardized beta weight, over-commitment was the most important predictor of anxiety and depression, followed by self-blame for anxiety and social support in depression. All other variables were of similar importance for anxiety and depression by standardized beta. The above variables accounted for 48% of the variance in the anxiety score, and 43.5% of the variance in the depression score. It can be seen by comparing the results from Table 6 with the regressions in Table 5, that the unique overall variance explained over just the DCS and ERI variables by ways of coping was 6.7% in anxiety and 4.5% overall in depression (about one-sixth more variance explained in anxiety, and one-eighth in depression). Discussion Descriptive statistics showed that overall 27.3% of nurses were above the clinical cut-off for self-reported levels of anxiety or depression as measured by the HAD. This compares to the 27% of hospital staff found to be suffering stress and mental ill health, by Calnan et al. (2001) and the 14 18% they found in the general population. It was also shown that 45% of nurses claimed that stress at work had caused, or made an existing illness worse. Mark and Smith (2011) found that 40% of a sample of 307 university staff, and 26% of a sample of 120 members of the general population, claimed that stress at work had affected health negatively. Thus, a greater percentage of the sample of nurses believed they had suffered stress-related illness compared to these two groups. Hypothesis one predicted that positive coping would be negatively associated with anxiety and depression and negative coping behaviours positively with anxiety and depression in nurses. The reported correlations show majority of the predictions of the experimental hypothesis were supported. These findings confirm those of many researchers, particularly Folkman et al. (1986). The data showed that negative coping behaviours were generally more important by standardized beta weight in the regressions than positive coping behaviours, suggesting that an absence of negative coping behaviours may in fact be more strongly associated with positive mental health outcomes than the presence of positive coping behaviours. Hypothesis two predicted that job demands would be positively associated with anxiety and depression, and control variables and social support would be negatively associated with outcomes. Also, significant interactions were predicted (Hypothesis 2a) between demands and control variables, and demands and support. Job demands correlated significantly with anxiety in almost all regression models. Job demands were also positively associated with depression in the DCS-only main-effects model. Decision authority did not emerge as a significant main-effects predictor in any regression, but did have a significant interaction relationship with job demands in the prediction of anxiety. This could suggest that skill discretion (chance to choose own skills) has an independent relationship to mental health outcomes, but decision authority (control over workplace events) is only related to outcomes through its relationship to job demands. These data support the first part of the hypothesis and the interaction of DA (Decision Authority) supports the second with regards to anxiety (as suggested by Karasek s work, 1979). Demands control support accounted for small to medium percentages of the variance in

14 14 G. Mark and A. P. Smith anxiety and depression, with job demands and social support being the most important factors by standardized beta weight. Hypothesis three predicted that extrinsic effort and over-commitment would be positively associated with depression and anxiety, and intrinsic reward would be negatively associated with anxiety and depression. ERI accounted for more of the variance overall in anxiety and depression than DCS. The regressions also showed that by standardized beta weight, rewards, extrinsic effort and over-commitment appeared to be as important as, or more important than DCS variables in the prediction of anxiety and depression. Table 5 shows that over-commitment was the most important predictor by beta weight for anxiety and depression. An interaction between overcommitment and intrinsic reward was significant in the anxiety analysis (Hypothesis 3a). However, as can be seen in Figure 2, this is unlikely to be a buffering effect, as the anxiety scores at high levels of reward and commitment are the same, and rewards were only associated with lower anxiety when over-commitment was low. This suggests that in normal circumstances, high rewards are associated with improved mental health, but when over-commitment is high, levels of reward are irrelevant or perhaps overwhelmed. Hypothesis four predicted that there would be significant interactions between negative job characteristics (job demands and extrinsic effort), and positive coping behaviours, so that problem-focused coping would moderate the effects of negative job characteristics on mental health outcomes. However, no significant interactions were found between these variables in the prediction of anxiety or depression. Hypothesis five predicted that coping, efforts, rewards, and demand control support would account for a significant amount of the variance in anxiety and depression, and that ways of coping would significantly add to the explained variance in outcomes over and above use of DCS and ERI alone. The regressions show that variables from all of the experimental constructs (coping, DCS, and ERI) were represented in the final regression equations for anxiety and depression. The variables shown in Table 6 accounted for 48% of the variance in anxiety scores, and 43.5% in depression scores, the highest for any regressions. The inclusion of coping in the final regression models supports the assertion that individual difference variables can significantly contribute, and account for different percentages of the variance in depression and anxiety, over DCS and ERI factors alone. In summary, all but one of the hypotheses in this paper were fully or partially supported. The data from this sample of nurses support previous findings reported by Karasek (1979), Siegrist (1996), Folkman and Lazarus (1980) and others. The various regressions showed that demand controls support, and efforts and rewards both contributed separately to the overall regression models. While over-commitment appears to be the most significant predictor by standardized beta weight, the other sub-factors of the DCS and ERI models appear to be of importance. It is evident that both of these models make distinct contributions to anxiety and depression, and studies on nursing samples that use one construct and exclude the other, may miss out explaining important parts of the variance in outcomes. In addition, while the more traditional models of DCS and ERI accounted for the majority of variance, there is clearly a significant contribution to be made by coping behaviours. These results support the DRIVE model (Demands Resources and Individual Effects) as outlined in Mark and Smith (2008). This research also suggests that a primary focus on individual and social support factors, and a secondary focus on demand-type factors may be pertinent in occupations where the reduction of demands is not a realistic option (such as for

15 Stress and nurses 15 nurses, fire fighters, doctors, police, etc.). Such recommendations cannot be made by research that does not include coping or individual difference factors. Conclusions The results of this paper show that there are robust associations between ways of coping, job demands, levels of control, social support, extrinsic effort, over-commitment, rewards, and anxiety and depression in this sample of nurses. The data support much past research and the simultaneous use of multiple theoretical constructs from popular stress models, and shows that coping adds something new to the existing body of workplace stress research. The results showed that no one group of factors emerged overall as being the most important in accounting for variance in anxiety and depression, and DCS, ERI, and coping, each added uniquely to the study of anxiety and depression in nurses. These relationships, as well as the role of other individual differences in the stress process, should be explored in future research. The workplace is a complex environment, and stress at work is a complex process. The fact that different individuals can respond to the same stressors in different ways, shows that an understanding of how different individual difference factors and job characteristics compare, interact, and influence one another is very important. The result that nearly 45% of nurses believed that work stress had directly influenced their health, shows the importance of studying stress in this population, and it is believed that the best way to help both employees and employers alike, is through multi-factor stress research, based on transactional stress models. References Audit Commission (2011). Managing sickness absence in the NHS. Health Briefing. Ball, J., Pike, G., Cuff, C., Mellor-Clark, J., & Connell, J. (2002). RCN Working Well Survey. Retrieved from: well survey inside1/pdf Boggild, H., & Knutsson, A. (1999). Shift work, risk factors and cardiovascular disease. Scandinavian Journal of Work, Environment & Health, 25(2), British Psychological Society (2004). Guidelines for minimum standards of ethical approval in psychological research. Leicester, UK: Author. Retreived from data/ assets/pdf_file/0014/122360/bps-guidelines.pdf Calnan, M., Wainwright, D., Forsythe, M., Wall, B., & Almond, S. (2001). Mental health and stress in the workplace: The case of general practice in the UK. Social Science & Medicine, 52, Clegg, A. (2001). Occupational stress in Nursing: A review of the literature. Journal of Nursing Management, 9, Cox, T., & Ferguson, E. (1991). Individual differences, stress and coping. In C. L. Cooper & R. Payne (Eds.), Personality and stress: Individual differences in the stress process. (pp. 7 30). Chichester: John Wiley and Sons. Cox, T., Griffiths, A., & Rial-Gonzalez, E. (2000). Research on work related stress. European Agency for Health and Safety at Work. Luxembourg: Office for Official Publications of the European Communities. Demerouti, E., Bakker, A. B., Nachreiner, F., & Schaufeli, W. B. (2000). A model of burnout and life satisfaction among nurses. Journal of Advanced Nursing, 32, Department of Health (2009). NHS health and well-being. Final Report, November de Rijk, A. E., Le Blanc, P. M., Schaufeli, W. B., & De Jonge, J. (1998). Active coping and need for control as moderators of the job demand-control model: Effects on burnout. Journal of Occupational and Organizational Psychology, 71, 1 18.

16 16 G. Mark and A. P. Smith Folkman, S., & Lazarus, R. S. (1980). An analysis of coping in a middle-aged community sample. Journal of Health and Social Behaviour, 21, Folkman, S., Lazarus, R.S., Gruen, R.J., & DeLongis, A. (1986). Appraisal, coping, health status, & psychological symptoms. Journal of Personality and Social Psychology, 50(3), French, S. E., Lenton, R., Walters, V., & Eyles, J. (2000). An empirical evaluation of an expanded nursing stress scale. Journal of Nursing Measurement, 8, Healy, C. M., & Mckay, M. F. (2000). Nursing Stress: The effects of coping strategies and job satisfaction in a sample of Australian nurses. Journal of Advanced Nursing, 31(3), Karasek, R. (1979). Job demands, job decision latitude and mental strain: Implications for job redesign. Administrative Science Quarterly, 24, Karasek, R., Brisson, C., Kawakami, N., Houtman, I., Bongers, P., & Amick, B. (1988). The Job Content Questionnaire (JCQ): An instrument for internationally comparative assessments of psychosocial job characteristics. Journal of Occupational Health Psychology, 3(4), Karasek, R., & Theorell, T. (1990). Healthy work: Stress, productivity and the reconstruction of working life. New York: Basic Books. Kirkcaldy, B. D., & Martin, T. (2000). Job stress and satisfaction among nurses: Individual differences. Stress Medicine, 16, Kuper, H., Singh-Manoux, A., Siegrist, J., & Marmot, M. (2002). When reciprocity fails: Effort-reward imbalance in relation to coronary heart disease and health functioning within the Whitehall II Study. Occupational and Environmental Medicine, 59, Lambert, V. A., Lambert, C. E., & Ito, M. (2004). Workplace stressors, ways of coping and demographic characteristics as predictors of physical and mental health of Japanese hospital nurses. International Journal of Nursing Studies, 41, Mark, G. M., & Smith, A. P. (2008). Stress models: A review and suggested new direction. In J. Houdmont & S. Leka (Eds.), EA-OHP series (Vol. 3, pp ). Nottingham, UK: Nottingham University Press. Mark, G. M. & Smith, A.P. (2011). Effects of occupational stress, job characteristics, coping, and attributional style on the mental health and job satisfaction of university employees. Anxiety, Stress and Coping. 25, McVicar, A. (2003). Workplace stress in nursing: A literature review. Journal of Advanced Nursing, 44(6), Moos, R. H., & Holahan, C. J. (2003). Dispositional and contextual perspectives on coping: Towards an integrative framework. Journal of Clinical Psychology, 59(12), Moos, R. H., Holahan, B. J., & Beutler, L. E. (2003). Dispositional and contextual perspectives on coping. Journal of Clinical Psychology, 59(12), Muncer, S., Taylor, S., Green, D. W., & McManus, I. C. (2001). Nurses representations of the perceived causes of work-related stress: A network drawing approach. Work & Stress, 15(1), NHS Information Centre for Health and Social Care (2011). Sickness absence rates in the NHS: July-September Retreived from collections/workforce/sickness-absence/sickness-absence-rates-in-the-nhs-july september Penley, J. A., Tomaka, J., & Wiebe, J. S. (2002). The association of coping to physical and psychological health outcomes: A meta-analytic review. Journal of Behavioural Medicine, 25(6), Perrewe, P. L., & Zellars, K. L. (1999). An examination of attributions and emotions in the transactional approach to the organizational stress process. Journal of Organizational Behavior, 20, Peter, R., & Siegrist, J. (1999). Chronic psychosocial stress at work and cardiovascular disease: The role of effort-reward imbalance. International Journal of Law & Psychiatry, 22, Siegrist, J. (1996). Adverse health effects of high-effort/low-reward conditions. Journal of Occupational Health Psychology, 1,

17 Stress and nurses 17 Tyler, P., & Cushway, D. (1998). Stress & well-being in health-care staff: The role of negative affectivity, and perceptions of job demand and discretion. Stress Medicine, 14, Van Der Doef, M., & Maes, S. (1999). The job-demand (-support) model and psychological wellbeing: A review of 20 years of empirical research. Work & Stress, 13(2), Vitaliano, P. P., Russo, J., Carr, J. E., Maiuro, R. D., & Becker, J. (1985). The ways of coping checklist psychometric properties. Multivariate Behavioral Research, 20, Weyers, S., Peter, R. M. A., Boggild, H., Jeppesen, H., & Siegrist, J. (2006). Psychosocial work stress is associated with poor self-rated health in Danish nurses: A test of the effort reward imbalance model. Scandinavian Journal of Caring Sciences, 20, Zigmond, A. S., & Snaith, R. P. (1983). The Hospital Anxiety and Depression Scale. Acta Psychiatrica Scandinavia, 67, Received 28 April 2011; revised version received 9 August 2011