- Email: [email protected]
Job strain, burnout, and depressive symptoms: A prospective study among dentists
Journal of Affective Disorders 104 (2007) 103 – 110 www.elsevier.com/locate/jad
Research report
Job strain, burnout, and depressive symptoms: A prospective study among dentists Kirsi Ahola ⁎, Jari Hakanen Finnish Institute of Occupational Health, Topeliuksenkatu 41 a A, FI-00250 Helsinki, Finland Received 30 November 2006; received in revised form 13 March 2007; accepted 13 March 2007 Available online 19 April 2007
Abstract Background: Burnout has been presented as an antecedent of depression, but longitudinal data are lacking. We investigated whether burnout mediates the association between job strain and depressive symptoms. Methods: Two surveys were conducted. In 2003, 71% of Finnish dentists were reached, and the response rate of the 3-year followup was 84% (n = 2555). Burnout was measured with the Maslach Burnout Inventory and depressive symptoms with the Beck Depression Inventory. The sequences ‘job strain–burnout–depressive symptoms’ and ‘job strain–depressive symptoms–burnout’ were investigated with logistic regression analyses. Results: Of the burnout sufferers without depressive symptoms at baseline, 23% reported depressive symptoms at follow-up. The adjusted odds ratio of burnout for depressive symptoms was 2.6 (95% CI 2.0–3.5). The effect of job strain on depressive symptoms had an OR of 3.4 (95% CI 2.0–5.7), but it disappeared when adjusted for burnout. Of those who had depressive symptoms without burnout at baseline, 63% had burnout at follow-up. The adjusted odds ratio of depressive symptoms for burnout was 2.2 (95% CI 1.4–3.4). The effect of job strain on burnout had an OR of 27.9 (95% CI 6.5–120.2) for the men and 4.9 (95% CI 2.5–9.6) for the women. These effects remained significant after adjustment for depressive symptoms. Limitations: The study was conducted among one occupational group. Conclusions: There is a reciprocal relationship between burnout and depressive symptoms. Job strain predisposes to depression through burnout. In comparison, job strain predisposes to burnout directly and via depression. © 2007 Elsevier B.V. All rights reserved. Keywords: Job strain; Burnout; Depression; Dentists; Prospective study
1. Introduction In 1974, psychiatrist Herbert Freudenberger (1974) described a negative occupational phenomenon among dedicated volunteers working in a clinic for drug addicts: ‘Burnout’ meant that a staff member became ex⁎ Corresponding author. Tel.: +358 30 474 2492; fax: +358 30 474 2552. E-mail address: [email protected] (K. Ahola). 0165-0327/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.jad.2007.03.004
hausted from excessive demands on energy, strength or resources about a year after he or she began working. Among the physical and behavioural signs, the volunteer looked, acted and seemed depressed. In 1976, Christina Maslach (1976), a researcher in social psychology, wrote about how professionals in health and social services can lose all emotional feelings and concern for their clients after months of listening to their problems. On the basis of her interviews, burnout
104
K. Ahola, J. Hakanen / Journal of Affective Disorders 104 (2007) 103–110
occurred when helpers were unable to cope with continuous emotional stress. Burnout was found to correlate with mental illness. Since the development of the Maslach Burnout Inventory (MBI) (Maslach and Jackson, 1981; Maslach et al., 1996), occupational burnout has been extensively studied, and several formulations of the syndrome have been presented (Shirom, 2003). Even though no final consensus for the definition of burnout (Cox et al., 2005) or no binding diagnostic criteria for its assessment (Weber and Jaekel-Reinhard, 2000) have been established, occupational burnout is unanimously regarded as a consequence of chronic work-related stress (Maslach et al., 2001). Among researchers, there has been quite wide consensus of how to measure burnout, since over 90% of all burnout studies have employed the Maslach Burnout Inventory (Schaufeli and Enzmann, 1998). Burnout is quite common in developed countries (Schaufeli and Enzmann, 1998; Maslach et al., 2001; Ahola et al., 2005; Shirom, 2005). It has been shown to associate with physical illnesses (Honkonen et al., 2006) and depressive disorders (Ahola et al., 2005) and to predict medically certified sickness absences (ToppinenTanner et al., 2005). The job strain model (Karasek, 1979) is the most influential conceptualization of work stress in occupational health research. The dimensions of this model, psychological job demands and job control, and especially the combination of high demands and low control (called job strain) have predicted serious health consequences (Kivimäki et al., 2002; de Lange et al., 2003; Belkic et al., 2004; Kivimäki et al., 2006). High job demands and low control have also predicted psychiatric morbidity (Stansfeld et al., 1997, 1999; Niedhammer et al., 1998; Paterniti et al., 2002), but evidence on the mental health effects of job strain has not been consistent (Cropley et al., 1999; de Lange et al., 2003; Ylipaavalniemi et al., 2005). Only cross-sectional associations between job strain and burnout have been tested, and these studies have lent support to the association (van der Doef and Maes, 1999; Ahola et al., 2006). Burnout has been predicted by various psychosocial work characteristics (Kalimo et al., 2003; Borritz et al., 2005) and is thought to result especially from the combination of high demands and low resources at work (Schaufeli and Bakker, 2004). The strong association between burnout and depression (Glass and McKnight, 1996; Ahola et al., 2005) has raised questions about their conceptual overlap and redundancy. When the concepts of burnout and depression are examined on the basis of appearance, biomarkers, developmental process and statistical asso-
ciations, both similarities and differences arise (Golembiewski et al., 1992; Maslach and Schaufeli, 1993; Leiter and Durup, 1994; McKnight and Glass, 1995; Maier and Watkins, 1998; Schaufeli and Enzmann, 1998; Bakker et al., 2000; Brenninkmeyer et al., 2001; Iacovides et al., 1999, 2003; Toker et al., 2005, Middeldorp et al., 2006). On the basis of the evidence, it can be concluded, that even though burnout and depression are highly related and share common features, they are not completely redundant. One possible explanation for the partial overlap between burnout and depression is that burnout mediates the relationship between psychosocial work characteristics and depression. This explanation has been supported by cross-sectional data among general population (Ahola et al., 2006). In cross-sectional settings, burnout has also been shown indirectly to lead to depression (Leiter and Durup, 1994; Bakker et al., 2000). Furthermore, psychiatric symptoms have been shown to increase and mental health to deteriorate, as burnout advances (Golembiewski et al., 1992). In addition, the more severe burnout was, the closer it was qualitatively to depression according to the symptoms (Iacovides et al., 2003). However, to our knowledge, there are no prospective studies on the temporal relations between the psychosocial characteristics of work, burnout, and depression. The aim of our study was to investigate whether burnout mediated the relationship between job strain and depression in a 3-year follow-up study among Finnish dentists (Fig. 1). A fully mediated effect between job strain, burnout, and depression can be established if the following criteria are met (Kenny, 2005): (1) burnout predicts depression, (2) job strain predicts burnout, and (3) job strain predicts depression, but not after burnout is adjusted for. Similarly, we also investigated the possible reversed effect, i.e. whether depression mediates the relationship between job strain and burnout (research model B in Fig. 1). 2. Methods 2.1. Participants and procedure This study was part of a longitudinal research project that focused on psychosocial working conditions, work–family interface, well-being, and health in dentistry (Hakanen, 2004; Hakanen and Perhoniemi, 2006). In 2003, a questionnaire survey was aimed at every dentist who was a member of the Finnish Dental Association (FDA) (n = 4588). About 98% of the working-aged dentists employed in clinical work in
K. Ahola, J. Hakanen / Journal of Affective Disorders 104 (2007) 103–110
Finland belong to the FDA (FDA 2005). Altogether, 3255 (71%) dentists responded to the questionnaire at baseline, and 2555 of those identified 3 years later (n = 3035) responded to the follow-up in 2006 (84%). In 2003, the respondents represented all Finnish dentists in terms of age and sex (Hakanen et al., 2005). The participants of the follow-up study accounted for 57% of the dental profession in Finland. The respondents at baseline and follow-up did not differ from the nonrespondents at follow-up with respect to the level of burnout or depressive symptoms. 2.2. Measures Burnout was measured with the Maslach Burnout Inventory (MBI), which has high reliability and validity (Maslach et al., 1996). The MBI consists of 22 items in three subscales: emotional exhaustion (9 items), depersonalization (5 items), and personal accomplishment (8 items). The items were scored on a 7-point frequency rating scale ranging from 0 (never) to 6 (daily). High scores on emotional exhaustion and depersonalization and low scores on personal accomplishment are indicative of burnout. The items of personal accomplishment were reversed. We included persons with a maximum of two missing values on the emotional exhaustion scale, one missing value on the cynicism scale, and two missing values on the personal accomplishment scale. The missing values of a respondent were replaced by the mean of the existing values of the respondent on the dimension in question. The reliability (Cronbach's α) of the whole inventory was 0.89.
105
In order to assess the level of burnout, we calculated a weighted sum score of the dimensional scores so that exhaustion, depersonalization and diminished personal accomplishment had different weights in the syndrome (Kalimo et al., 2003). This syndrome indicator was derived with the help of a discriminant function analysis in which various health-related indicators were used as dependent variables (Kalimo and Toppinen, 1997). Coefficients were formed by weighing each dimension so that the scores corresponded to the original response scale (0.4 × exhaustion + 0.3 × depersonalization + 0.3 × diminished personal accomplishment). Burnout was categorized as follows: no burnout (scores 0–1.49), mild burnout (scores 1.50– 3.49), and severe burnout (scores 3.50–6). This categorization meant that burnout was severe when symptoms were experienced approximately once a week or daily, they were mild when symptoms existed monthly, and there was no burnout when the symptoms were experienced only a few times a year or never (Kalimo et al., 2003). Burnout was dichotomized as no versus mild to severe. The short form of the Beck Depression Inventory (BDI) was used to assess depression (Beck and Beck, 1972). It consists of 13 items that are scored from 0 to 3 (α = 0.84). An acceptable answer was expected for at least 11 items. Missing values (2 at the most) were replaced by the mean of the existing values for that particular respondent. A sum score for the depressive symptoms was then calculated. Depressive symptoms were categorized as no (0–4 points), mild (5–7 points), moderate (8–15 points), and severe (16–39 points) (Beck and Beck, 1972). Depression was dichotomized as no versus mild, moderate or severe.
Fig. 1. The research models to test mediational relationships between job strain, burnout, and depression.
106
K. Ahola, J. Hakanen / Journal of Affective Disorders 104 (2007) 103–110
Job strain was measured with the Job Content Questionnaire (JCQ) (Karasek et al., 1998). A short scale of job demands comprised three items (α = 0.79; e.g. “My job requires working very fast”), and the scale for job control had nine items (α = 0.85) (e.g., “My job allows me to make a lot of decisions on my own”; “My job requires a high level of skills”). Responses were given on a 5-point scale ranging from 1 (strongly agree) to 5 (strongly disagree). We created an indicator for job strain as a continuous quotient of job demands distributed by job control (Landbergis et al., 1994). Information on age, sex, marital status, professional sector (private/public), type of employment (permanent/ fixed-term), job tenure, supervisory position (no/yes), and working hours were included in the questionnaire. Age was categorized as under 36, 36–45, 46–55, and over 55 years. Marital status was categorized as married or cohabiting, divorced or widowed, and unmarried. Job tenure was categorized as under 5, 5–9, 10–19, 20–29, and over 29 years. Working hours were categorized as full-time (at least 35 h a week), shortened (25–34 h a week), and part-time (no more than 24 h a week). 2.3. Statistical analysis The entire study sample was described with distributions of sociodemographic factors, work characteristics, burnout, and depression. Among those who were free of depressive symptoms at baseline, binary logistic regression analyses were used to explore the prospective associations concerning the research model A (Fig. 1): between burnout at baseline and depression at follow-up (path 1), between job strain at baseline and burnout at follow-up (path 2), and between job strain at baseline and depression at follow-up (path 3). The analyses were adjusted for sex, age and marital status. The final adjustment included also burnout at baseline as a continuous variable in the models with job strain. Interaction term was applied in the models to test whether job strain had an interaction effect with sex on depression. Concerning the research model B among those who were free of burnout at baseline (Fig. 1), the prospective associations between depression at baseline and burnout at follow-up (path 4), job strain at baseline and depression at follow-up (path 5), and job strain at baseline and burnout at follow-up (path 6) were also analysed with binary logistic regression models, which were adjusted for sex, age and marital status. The final adjustment included depression at baseline as a continuous variable in the models with job strain. Interaction term was applied to the models to test whether job strain had an interaction effect with sex on burnout.
3. Results The dentists were mostly women (74%), married or cohabiting (84%), and in permanent employment (94%). Table 1 shows also that a slight majority of participants worked full-time (63%) and in the public sector (62%). Of all participants, 51% were free of burnout and 72% were free of depressive symptoms at baseline (Table 2). 3.1. Research model A: job strain → burnout → depression The dentists who were free of depressive symptoms but reported symptoms of burnout at baseline, showed depressive symptoms at follow-up in 23% of the cases
Table 1 Baseline characteristics of the study population (n = 2555) Characteristic Sex Women Men Age 26–35 years 36–45 years 46–55 years 56–73 years Marital status Unmarried Married or cohabiting Divorced or widowed Missing data Professional sector Public Private Missing data Employment Permanent Fixed-term Missing data Job tenure 0–4 years 5–9 years 10–19 years 20–29 years 30 years or more Missing data Supervisory position Yes No Missing data Working hours Full-time Shortened Part-time Missing data
n (%) 1883 (74) 672 (26) 384 (15) 898 (35) 884 (35) 389 (15) 179 (7) 2143 (84) 220 (9) 13 (1) 1578 (62) 956 (37) 21 (1) 2392 (94) 141 (6) 22 (1) 144 (6) 304 (12) 805 (32) 860 (34) 431 (17) 11 (0) 677 (26) 1790 (70) 88 (3) 1606 (63) 542 (21) 386 (15) 21 (1)
K. Ahola, J. Hakanen / Journal of Affective Disorders 104 (2007) 103–110 Table 2 Job strain, burnout and depression at baseline and at follow-up (n = 2555)
Job strain, mean (S.D.) Missing data, n (%) Burnout, n (%) No Mild Severe Missing data Depression, n (%) No Mild Moderate Severe Missing data
Baseline
Follow-up
0.78 (0.28) 139 (5)
0.81 (0.30) 135 (5)
1313 (51) 1120 (44) 56 (2) 66 (3)
1173 (46) 1194 (47) 68 (3) 120 (5)
1840 (72) 350 (14) 296 (12) 28 (1) 41 (2)
1774 (69) 382 (15) 291 (11) 28 (1) 80 (3)
107
Table 4 Prospective associations of job strain at baseline for burnout at 3-year follow-up Model 1
Model 2
Sex
n/cases
OR (95% CI)
OR (95% CI)
Men Women
337/71 861/238
27.87 (6.46–120.2) 4.87 (2.46–9.64)
22.31 (5.08–98.07) 3.99 (1.99–7.99)
Model 1 adjusted for sex, age and marital status at baseline. Model 2 adjusted for sex, age, marital status and depression at baseline. OR indicates odds ratio. CI indicates confidence interval.
3.2. Research model B: job strain → depression → burnout
S.D. indicates standard deviation.
compared with 10% of those who did not report burnout symptoms at baseline. In this sub-population the adjusted probability of showing depressive symptoms at follow-up after having burnout at baseline was 2.6fold (95% CI 2.0–3.5) compared to not having burnout at baseline (path 1 in Fig. 1). Job strain predicted burnout among those free of depression at baseline (path 2 in Fig. 1). The adjusted probability of having burnout was 11.8-fold (95% CI 7.4– 18.7) for each one-point increase in job strain score in this group of dentists. When the level of burnout at baseline was also adjusted for in the final model, the probability of having burnout at follow-up decreased but remained statistically significant (OR 1.8, 95% CI 1.04–3.1). Table 3 shows that job strain predicted also depression (path 3 in Fig. 1). When the level of burnout at baseline was adjusted for in the final model, the effect of job strain on depression disappeared, thus showing that burnout fully mediated the impact of job strain on depressive symptoms. There were no significant interactions (p N 0.3) between sex and strain in the models for depressive symptoms.
Table 3 Prospective associations of job strain at baseline for new cases of depression at 3-year follow-up
Job strain
Total n/cases
Model 1
Model 2
OR (95% CI)
OR (95% CI)
1684/242
3.39 (2.03–5.66)
1.30 (0.73–2.30)
Model 1 adjusted for sex, age and marital status at baseline. Model 2 adjusted for sex, age, marital status and burnout at baseline. OR indicates odds ratio. CI indicates confidence interval.
Among the dentists who were free of burnout at baseline, but reported depressive symptoms, burnout was present at follow-up in 63% of the cases compared with 20% among those who did not show symptoms of depression at baseline. In this sub-population the adjusted probability of having burnout at follow-up after having experienced depressive symptoms at baseline was 2.2-fold (95% OR 1.4–3.4) compared to the situation of not having experienced depressive symptoms at baseline (path 4 in Fig. 1). Job strain predicted depression among those free of burnout at baseline. The adjusted probability of having depression was 7.5-fold (95% CI 3.5–15.9) for each one-point increase in job strain score (path 5 in Fig. 1). When the level of depression at baseline was adjusted for in the final model, the probability of showing depression at follow-up decreased but remained statistically significant (OR 4.8, 95% CI 2.1–11.0). A significant interaction (p = 0.03) between sex and strain was found in the model of strain on burnout in this sub-group. Table 4 shows that job strain predicted burnout both among the men and among the women (path 6 in Fig. 1). After adjustment for the level of depression at baseline, these relationships decreased but remained statistically significant. This result indicates that job strain had a direct as well as an indirect effect through depressive symptoms on burnout. 4. Discussion To our knowledge, this is the first study to investigate the temporal relationship between job strain, occupational burnout and depressive symptoms in a prospective setting. We found that the effects between burnout and depression are reciprocal: Occupational burnout predicted new cases of depressive symptoms and depression
108
K. Ahola, J. Hakanen / Journal of Affective Disorders 104 (2007) 103–110
predicted new cases of burnout. In addition, burnout mediated fully the relationship between job strain and depression. This means that after controlling for the mediating effect of burnout, there was no direct relationship between job strain and the incidence of depression; job strain was related to burnout, which in turn was related to depression. Instead, depression mediated only partially the relationship between job strain and burnout indicating both direct and mediated relationship between job strain and burnout. On the basis of the present study, it seems possible that occupational burnout is a phase in the development of work-related depression. This view has earlier been supported by cross-sectional data on the general population (Ahola et al., 2006). The results of our study agree also with earlier studies showing burnout to mediate between psychosocial factors and health. In these studies, health was indicated by the duration of company-registered sickness absences (Bakker et al., 2003), psychosomatic health complaints (Schaufeli and Bakker, 2004), and selfrated health and work ability (Hakanen et al., 2006). In addition to burnout leading to depression, depression also predisposed to burnout. There are at least two main mechanisms to explain this reversed effect (de Lange et al., 2004). Firstly, those who are depressed probably have lower resources to meet the demands of their work, and this situation predisposes them to burnout. Secondly, it is possible that those who are depressed perceive and evaluate their situation at work more negatively than those who are free of depression. On the basis of the present study, the path of job strain predisposing to depression through experienced burnout appear stronger than the path of job strain predisposing to burnout via depression. The extent to which burnout differs from depressive symptomatology has been debated in the past (Taris, 2006). In studies conducted to rule out the possibility of redundancy, it has been recommended to include wellvalidated measures of burnout and depressive symptoms to ascertain that burnout and its correlates are not due to the effects of depression (Shirom, 2005). The present longitudinal study supports the distinction between burnout and depression phenomena. Furthermore, it contradicts the suggestion of burnout and depression developing from the same risk factors simultaneously, “in tandem”, as McKnight and Glass (1995) have suggested on the basis of their small sample (n = 100) of nurses. However, further research is needed to fully understand the cross-lagged reciprocal associations between burnout and depression. As a risk factor for burnout, job strain was emphasized especially among healthy men in our study. The impact of chronic stressors and psychosocial factors on health has
generally been stronger for women than for men (Denton et al., 2004). However, work-related factors such as high job strain and occupational burnout have associated with mental disorders especially among men (Ahola et al., 2005; Virtanen et al. 2007). This finding may reflect the significance work has in men's lives. At least two limitations of this study are noteworthy. Firstly, this study was based on survey data that present problems of self-report bias and common method variance (Lindell and Whitney, 2001). However, variance in self-report measures of job conditions has been largely attributed to variations in the objective work environment (Spector, 1992). Subjective evaluation of work characteristics has also been shown to be more strongly related to mental well-being than objective evaluation is (Stansfeld et al., 1999). Still, it is well-known that common method variance may artificially inflate associations through factors such as negative affectivity and social desirability (Parkes, 1990; Heinisch and Jex, 1997). Ahola et al. (2006) found that a relationship between high job strain and depression, as well as between burnout and depression, appeared regardless of the depression measure used, although the associations were stronger with self-reported variables than with the use of a structured psychiatric interview. Thus we argue that it is unlikely that common method variance has caused major confounding in our study, which employed a prospective design, included only “healthy workers” at baseline concerning the outcome variable, and controlled for the baseline values with respect to the mediating variable. Nevertheless, the impact of common method variance should be further examined in future research by combining other measures with self-report inventories in a prospective setting. Secondly, the study population comprised only one occupational group, the dentists. Our sample was large and representatively comprised the majority of working dentists in Finland. Dentists practice human service work and, as for socioeconomic status, belong to the group of upper white-collar workers. Therefore, the results must be generalized with caution outside this reference group. On the other hand, the results of our study agree with the findings of a previous crosssectional population-based study on the relationship between job strain, burnout and depression, which included 30- to 64-year-old employees from all types of occupations (Ahola et al., 2006). Because our study included dentists who were working and who responded to both questionnaires, it is possible that this procedure excluded the dentists with the worst situations and thus resulted in weakened associations. However, we were able to reach 71% of the whole profession in the first
K. Ahola, J. Hakanen / Journal of Affective Disorders 104 (2007) 103–110
phase of the study, and at follow-up the respondents and the non-respondents did not differ in relation to burnout or depressive symptoms. Sickness absence and disability pensions granted on the basis of depressive disorders have increased in Finland during the present millennium (Salminen, 2004). Among intensive human service work, like dentistry, the importance of mental well-being at work is evident. If human service professionals are to perform well and keep working until statutory retirement age, their working conditions should be continuously evaluated. In primary health care, and especially in occupational health services, burnout could be considered a serious alarm signal of an unfavourably developing working situation. Among employed clients, assessment procedures and early intervention practices concerning workrelated well-being should be further developed. In conclusion, job strain predisposes to depression through burnout. In comparison, job strain predisposes to burnout directly and via depression. Even though the relationship between burnout and depression is reciprocal, the path from burnout to depression appears to be stronger than the path from depression to burnout. 5. Contributors K. Ahola and J. Hakanen together planned the design, J. Hakanen gathered the data, K. Ahola analysed the data and wrote the first version of the manuscript, and K. Ahola and J. Hakanen together refined the manuscript. Acknowledgements This study was supported by the Finnish Work Environment Fund (project no. 105325) and the Finnish Dental Association. The Finnish Work Environment Fund and The Finnish Dental Association had no further role in the study design, the analysis and interpretation of the data, writing or submitting the manuscript. References Ahola, K., Honkonen, T., Isometsä, E., Kalimo, R., Nykyri, E., Aromaa, A., Lönnqvist, J., 2005. The relationship between jobrelated burnout and depressive disorders — results from the Finnish Health 2000 Study. J. Affect. Disord. 88, 55–62. Ahola, K., Honkonen, T., Kivimäki, M., Virtanen, M., Isometsä, E., Aromaa, A., Lönnqvist, J., 2006. Contribution of burnout to the association between job strain and depression: the Health 2000 Study. J. Occup. Environ. Med. 48, 1023–1030. Bakker, A.B., Schaufeli, W.B., Demerouti, E., Janssen, P.P., van der Hulst, R., Brouwer, J., 2000. Using equity theory to examine the difference between burnout and depression. Anxiety Stress Coping 13, 247–268.
109
Bakker, A.B., Demerouti, E., de Boer, E., Schaufeli, W.B., 2003. Job demand and job resources as predictors of absence duration and frequency. J. Vocat. Behav. 62, 341–356. Beck, A.T., Beck, R.W., 1972. Screening depressed patients in family practice. A rapid technique. Postgrad. Med. 52, 81–85. Belkic, K., Landsbergis, P.A., Schnall, P.L., Baker, D., 2004. Is job strain a major source of cardiovascular disease risk? Scand. J. Work Environ. Health 30, 85–128. Borritz, M., Bültmann, U., Rugulies, R., Christensen, K.B., Villadsen, E., Kristensen, T., 2005. Psychosocial work characteristics as predictors for burnout: findings from 3-year follow up of the PUMA study. J. Occup. Environ. Med. 47, 1015–1025. Brenninkmeyer, V., Van Yperen, N.W., Buunk, B.P., 2001. Burnout and depression are not identical twins: is decline of superiority a distinguishing feature? Pers. Individ. Differ. 30, 873–880. Cox, T., Tisserand, M., Taris, T., 2005. The conceptualization and measurement of burnout: questions and directions. Work Stress 19, 187–191. Cropley, M., Steptoe, A., Joekes, K., 1999. Job strain and psychiatric morbidity. Psychol. Med. 29, 1411–1416. de Lange, A.H., Taris, T.W., Houtman, I.L.D., Kompier, M.A.J., Bongers, P.M., 2003. “The very best of the millennium”: longitudinal research and the demand–control–(support) model. J. Occup. Health Psychol. 8, 282–305. de Lange, A.H., Taris, T.W., Kompier, M.A.J., Houtman, I.L.D., Bongers, P.M., 2004. The relationships between work characteristics and mental health: examining normal, reversed and reciprocal relationships in a 4-wave study. Work Stress 18, 149–166. FDA, 2005. Facts about the number of dentists [Lukumäärätietoja hammaslääkäreistä]. (http://www.hammasll.fi/). Downloaded 10 June 2005. Freudenberger, H.J., 1974. Staff burn-out. J. Soc. Issues 30, 159–165. Glass, D.C., McKnight, J.D., 1996. Perceived control, depressive symptomatology, and professional burnout: a review of the evidence. Psychol. Health 11, 23–48. Golembiewski, R.T., Munzenrider, R.F., Scherb, K., Billingsley, W., 1992. Burnout and psychiatric “cases”: early evidence of an association. Anxiety Stress Coping 5, 69–78. Hakanen, J.J., 2004. Hammaslääkäreiden työhyvinvointi Suomessa [Work-Related Well-Being Among Finnish Dentists]. Finnish Institute of Occupational Health and the Finnish Dental Association, Helsinki. Hakanen, J.J., Bakker, A.B., Demerouti, E., 2005. How dentists cope with their job demands and stay engaged: the moderating role of job resources. Eur. J. Oral Sci. 113, 479–487. Hakanen, J.J., Perhoniemi, R., 2006. Hammaslääkäreiden työhyvinvointi Suomessa. Kolmen vuoden seurantatutkimus [Work-Related Well-Being Among Finnish Dentists. Three-Year Follow-Up Study]. Finnish Institute of Occupational Health and the Finnish Dental Association, Helsinki. Hakanen, J.J., Bakker, A.B., Schaufeli, W., 2006. Burnout and engagement among teachers. J. Sch. Psychol. 43, 495–513. Heinisch, D.A., Jex, S.M., 1997. Negative affectivity and gender as moderators of the relationship between work-related stressors and depressed mood at work. Work Stress 11, 46–57. Honkonen, T., Ahola, K., Pertovaara, M., Isometsä, E., Kalimo, R., Nykyri, E., Aromaa, A., Lönnqvist, J., 2006. The association between burnout and physical illness in the general population — results from the Finnish Health 2000 Study. J. Psychosom. Res. 61, 59–66. Iacovides, A., Fountoulakis, K.N., Moysidou, Ch., Ierodiakonou, Ch., 1999. Burnout in nursing staff: is there a relationship between depression and burnout? Int. J. Psychiatry Med. 29, 421–433.
110
K. Ahola, J. Hakanen / Journal of Affective Disorders 104 (2007) 103–110
Iacovides, A., Fountoulakis, K.N., Kaprinis, St., Kaprinis, G., 2003. The relationship between job stress, burnout and clinical depression. J. Affect. Disord. 75, 209–221. Kalimo, R., Toppinen, S., 1997. Työuupumus Suomen työikäisellä väestöllä [Burnout Among Finnish Working Population]. Työterveyslaitos, Helsinki. Kalimo, R., Pahkin, K., Mutanen, P., Toppinen-Tanner, S., 2003. Staying well or burning out at work: work characteristics and personal resources as long-term predictors. Work Stress 17, 109–122. Karasek, R.A., 1979. Job demands, job decision latitude and mental strain: implications for job redesign. Adm. Sci. Q. 24, 285–308. Karasek, R.A., Brisson, C., Kawakami, N., Houtman, I., Bongers, P., Amick, B., 1998. The job content questionnaire (JCQ): an instrument for intentionally comparative assessments of psychosocial work characteristics. J. Occup. Health Psychol. 3, 322–355. Kenny, D.A., 2005. Mediation. In: Everitt, B.S., Howell, D.C. (Eds.), Encyclopedia of Statistics in Behavioral Science, vol. 3. John Wiley, Chichester, pp. 1194–1198. Kivimäki, M., Leino-Arjas, P., Luukkonen, R., Riihimäki, H., Vahtera, J., Kirjonen, J., 2002. Work stress and risk of cardiovascular mortality: prospective cohort study of industrial employees. Br. Med. J. 325, 857. Kivimäki, M., Virtanen, M., Elovainio, M., Kouvonen, A., Väänänen, A., Vahtera, J., 2006. Work stress in the etiology of coronary heart disease: systematic review and meta-analysis of prospective cohort studies. Scand. J. Work Environ. Health 32, 431–442. Landbergis, P.A., Schnall, P.L., Warren, K., Pickering, T.G., Schwatrz, J.E., 1994. Association between ambulatory blood pressure and alternative formulations of job strain. Scand. J. Work Environ. Health 20, 349–363. Leiter, M.P., Durup, J., 1994. The discriminant validity of burnout and depression: a confirmatory factor analytic study. Anxiety Stress Coping 7, 357–373. Lindell, M.K., Whitney, D.J., 2001. Accounting for common method variance in cross-sectional research designs. J. Appl. Psychol. 86, 114–121. Maier, S.F., Watkins, L.R., 1998. Cytokines for psychologists: implications of bidirectional immune-to-brain communication for understanding behavior, mood, and cognition. Psychol. Rev. 105, 83–107. Maslach, C., 1976. Burned-out. Human Behav. 5, 16–22. Maslach, C., Jackson, S.E., 1981. The measurement of experienced burnout. J. Occup. Behav. 2, 99–113. Maslach, C., Schaufeli, W.B., 1993. Historical and conceptual development of burnout. In: Schaufeli, W.B., Maslach, C., Marek, T. (Eds.), Professional Burnout: Recent Developments in Theory and Research. Taylor & Francis, London, pp. 1–16. Maslach, C., Jackson, S.E., Leiter, M.P., 1996. Maslach Burnout Inventory Manual, 3rd ed. Consulting Psychologists Press, Palo Alto, CA. Maslach, C., Schaufeli, W.B., Leiter, M.P., 2001. Job burnout. Annu. Rev. Psychol. 52, 397–422. McKnight, J.D., Glass, D., 1995. Perceptions of control, burnout, and depressive symptomatology: a replication and extension. J. Consult. Clin. Psychol. 63, 490–494. Middeldorp, C.M., Cath, D.C., Boosma, D.I., 2006. A twin-family study of the association between employment, burnout and anxious depression. J. Affect. Disord. 90, 163–169.
Niedhammer, I., Goldberg, M., Leclerc, A., Bugel, I., David, S., 1998. Psychosocial factors at work and subsequent depressive symptoms in the Gazel cohort. Scand. J. Work, Environ. Health 24, 197–205. Parkes, K.R., 1990. Coping, negative affectivity, and the work environment: additive and interactive predictors of mental health. J. Appl. Psychol. 75, 399–409. Paterniti, S., Niedhammer, I., Lang, T., Consoli, S.M., 2002. Psychosocial factors at work, personality traits and depressive symptoms. Longitudinal results from the GAZEL study. Br. J. Psychiatry 181, 111–117. Salminen, J.K., 2004. Mielenterveyden häiriöt ja työkyvyttömyys [mental disorders and work disability]. Suom. Lääkäril. 59, 4745–4748. Schaufeli, W.B., Bakker, A.B., 2004. Job demands, job resources, and their relationship with burnout and engagement: a multi-sample study. J. Organ. Behav. 25, 293–315. Schaufeli, W., Enzmann, D., 1998. The Burnout Companion to Study and Practice: A Critical Analysis. Taylor & Francis, London. Shirom, A., 2003. Job-related burnout: a review. In: Quick, J.C., Tetrick, L.E. (Eds.), Handbook of Occupational Health Psychology. American Psychological Association, Washington, DC, pp. 245–264. Shirom, A., 2005. Reflections on the study of burnout. Work Stress 19, 263–270. Spector, P.R., 1992. A consideration of the validity and meaning of self-report measures of job conditions. Int. Rev. Ind. Organ. Psychol. 69, 293–310. Stansfeld, S.A., Fuhrer, R., Head, J., Ferrie, J., Shipley, M., 1997. Work and psychiatric disorder in the Whitehall II study. J. Psychosom. Res. 43, 73–81. Stansfeld, S.A., Fuhrer, R., Shipley, M.J., Marmot, M.G., 1999. Work characteristics predict psychiatric disorder: prospective results from the Whitehall II study. Occup. Environ. Med. 56, 302–307. Taris, T.W., 2006. Bricks without clay: on urban myths in occupational health psychology. Work Stress 20, 99–104. Toker, S., Shirom, A., Shapira, I., Berliner, S., Melamed, S., 2005. The association between burnout, depression, anxiety, and inflammation biomarkers: C-reactive protein and fibrinogen in men and women. J. Occup. Health Psychol. 10, 344–362. Toppinen-Tanner, S., Ojajärvi, A., Väänänen, A., Kalimo, R., Jäppinen, P., 2005. Burnout as a predictors of medically certified sick-leave absences and their diagnosed causes. Behav. Med. 31, 18–27. van der Doef, M., Maes, S., 1999. The job demand–control(–support) model and psychological well-being: a review of 20 years of empirical research. Work Stress 13, 87–114. Virtanen, M., Honkonen, T., Kivimäki, M., Ahola, K., Vahtera, J., Aromaa, A., Lönnqvist, J., 2007. Work stress, mental health and antidepressant medication: findings from the Health 2000 Study. J. Affect. Disord. 98, 189–197. Weber, A., Jaekel-Reinhard, A., 2000. Burnout syndrome: a disease of modern societies. Occup. Med. 50, 512–517. Ylipaavalniemi, J., Kivimäki, M., Elovainio, M., Virtanen, M., Keltikangas-Järvinen, L., Vahtera, J., 2005. Psychosocial work characteristics and incidence of newly diagnosed depression: a prospective cohort study of three different models. Soc. Sci. Med. 61, 111–122.
Research report
Job strain, burnout, and depressive symptoms: A prospective study among dentists Kirsi Ahola ⁎, Jari Hakanen Finnish Institute of Occupational Health, Topeliuksenkatu 41 a A, FI-00250 Helsinki, Finland Received 30 November 2006; received in revised form 13 March 2007; accepted 13 March 2007 Available online 19 April 2007
Abstract Background: Burnout has been presented as an antecedent of depression, but longitudinal data are lacking. We investigated whether burnout mediates the association between job strain and depressive symptoms. Methods: Two surveys were conducted. In 2003, 71% of Finnish dentists were reached, and the response rate of the 3-year followup was 84% (n = 2555). Burnout was measured with the Maslach Burnout Inventory and depressive symptoms with the Beck Depression Inventory. The sequences ‘job strain–burnout–depressive symptoms’ and ‘job strain–depressive symptoms–burnout’ were investigated with logistic regression analyses. Results: Of the burnout sufferers without depressive symptoms at baseline, 23% reported depressive symptoms at follow-up. The adjusted odds ratio of burnout for depressive symptoms was 2.6 (95% CI 2.0–3.5). The effect of job strain on depressive symptoms had an OR of 3.4 (95% CI 2.0–5.7), but it disappeared when adjusted for burnout. Of those who had depressive symptoms without burnout at baseline, 63% had burnout at follow-up. The adjusted odds ratio of depressive symptoms for burnout was 2.2 (95% CI 1.4–3.4). The effect of job strain on burnout had an OR of 27.9 (95% CI 6.5–120.2) for the men and 4.9 (95% CI 2.5–9.6) for the women. These effects remained significant after adjustment for depressive symptoms. Limitations: The study was conducted among one occupational group. Conclusions: There is a reciprocal relationship between burnout and depressive symptoms. Job strain predisposes to depression through burnout. In comparison, job strain predisposes to burnout directly and via depression. © 2007 Elsevier B.V. All rights reserved. Keywords: Job strain; Burnout; Depression; Dentists; Prospective study
1. Introduction In 1974, psychiatrist Herbert Freudenberger (1974) described a negative occupational phenomenon among dedicated volunteers working in a clinic for drug addicts: ‘Burnout’ meant that a staff member became ex⁎ Corresponding author. Tel.: +358 30 474 2492; fax: +358 30 474 2552. E-mail address: [email protected] (K. Ahola). 0165-0327/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.jad.2007.03.004
hausted from excessive demands on energy, strength or resources about a year after he or she began working. Among the physical and behavioural signs, the volunteer looked, acted and seemed depressed. In 1976, Christina Maslach (1976), a researcher in social psychology, wrote about how professionals in health and social services can lose all emotional feelings and concern for their clients after months of listening to their problems. On the basis of her interviews, burnout
104
K. Ahola, J. Hakanen / Journal of Affective Disorders 104 (2007) 103–110
occurred when helpers were unable to cope with continuous emotional stress. Burnout was found to correlate with mental illness. Since the development of the Maslach Burnout Inventory (MBI) (Maslach and Jackson, 1981; Maslach et al., 1996), occupational burnout has been extensively studied, and several formulations of the syndrome have been presented (Shirom, 2003). Even though no final consensus for the definition of burnout (Cox et al., 2005) or no binding diagnostic criteria for its assessment (Weber and Jaekel-Reinhard, 2000) have been established, occupational burnout is unanimously regarded as a consequence of chronic work-related stress (Maslach et al., 2001). Among researchers, there has been quite wide consensus of how to measure burnout, since over 90% of all burnout studies have employed the Maslach Burnout Inventory (Schaufeli and Enzmann, 1998). Burnout is quite common in developed countries (Schaufeli and Enzmann, 1998; Maslach et al., 2001; Ahola et al., 2005; Shirom, 2005). It has been shown to associate with physical illnesses (Honkonen et al., 2006) and depressive disorders (Ahola et al., 2005) and to predict medically certified sickness absences (ToppinenTanner et al., 2005). The job strain model (Karasek, 1979) is the most influential conceptualization of work stress in occupational health research. The dimensions of this model, psychological job demands and job control, and especially the combination of high demands and low control (called job strain) have predicted serious health consequences (Kivimäki et al., 2002; de Lange et al., 2003; Belkic et al., 2004; Kivimäki et al., 2006). High job demands and low control have also predicted psychiatric morbidity (Stansfeld et al., 1997, 1999; Niedhammer et al., 1998; Paterniti et al., 2002), but evidence on the mental health effects of job strain has not been consistent (Cropley et al., 1999; de Lange et al., 2003; Ylipaavalniemi et al., 2005). Only cross-sectional associations between job strain and burnout have been tested, and these studies have lent support to the association (van der Doef and Maes, 1999; Ahola et al., 2006). Burnout has been predicted by various psychosocial work characteristics (Kalimo et al., 2003; Borritz et al., 2005) and is thought to result especially from the combination of high demands and low resources at work (Schaufeli and Bakker, 2004). The strong association between burnout and depression (Glass and McKnight, 1996; Ahola et al., 2005) has raised questions about their conceptual overlap and redundancy. When the concepts of burnout and depression are examined on the basis of appearance, biomarkers, developmental process and statistical asso-
ciations, both similarities and differences arise (Golembiewski et al., 1992; Maslach and Schaufeli, 1993; Leiter and Durup, 1994; McKnight and Glass, 1995; Maier and Watkins, 1998; Schaufeli and Enzmann, 1998; Bakker et al., 2000; Brenninkmeyer et al., 2001; Iacovides et al., 1999, 2003; Toker et al., 2005, Middeldorp et al., 2006). On the basis of the evidence, it can be concluded, that even though burnout and depression are highly related and share common features, they are not completely redundant. One possible explanation for the partial overlap between burnout and depression is that burnout mediates the relationship between psychosocial work characteristics and depression. This explanation has been supported by cross-sectional data among general population (Ahola et al., 2006). In cross-sectional settings, burnout has also been shown indirectly to lead to depression (Leiter and Durup, 1994; Bakker et al., 2000). Furthermore, psychiatric symptoms have been shown to increase and mental health to deteriorate, as burnout advances (Golembiewski et al., 1992). In addition, the more severe burnout was, the closer it was qualitatively to depression according to the symptoms (Iacovides et al., 2003). However, to our knowledge, there are no prospective studies on the temporal relations between the psychosocial characteristics of work, burnout, and depression. The aim of our study was to investigate whether burnout mediated the relationship between job strain and depression in a 3-year follow-up study among Finnish dentists (Fig. 1). A fully mediated effect between job strain, burnout, and depression can be established if the following criteria are met (Kenny, 2005): (1) burnout predicts depression, (2) job strain predicts burnout, and (3) job strain predicts depression, but not after burnout is adjusted for. Similarly, we also investigated the possible reversed effect, i.e. whether depression mediates the relationship between job strain and burnout (research model B in Fig. 1). 2. Methods 2.1. Participants and procedure This study was part of a longitudinal research project that focused on psychosocial working conditions, work–family interface, well-being, and health in dentistry (Hakanen, 2004; Hakanen and Perhoniemi, 2006). In 2003, a questionnaire survey was aimed at every dentist who was a member of the Finnish Dental Association (FDA) (n = 4588). About 98% of the working-aged dentists employed in clinical work in
K. Ahola, J. Hakanen / Journal of Affective Disorders 104 (2007) 103–110
Finland belong to the FDA (FDA 2005). Altogether, 3255 (71%) dentists responded to the questionnaire at baseline, and 2555 of those identified 3 years later (n = 3035) responded to the follow-up in 2006 (84%). In 2003, the respondents represented all Finnish dentists in terms of age and sex (Hakanen et al., 2005). The participants of the follow-up study accounted for 57% of the dental profession in Finland. The respondents at baseline and follow-up did not differ from the nonrespondents at follow-up with respect to the level of burnout or depressive symptoms. 2.2. Measures Burnout was measured with the Maslach Burnout Inventory (MBI), which has high reliability and validity (Maslach et al., 1996). The MBI consists of 22 items in three subscales: emotional exhaustion (9 items), depersonalization (5 items), and personal accomplishment (8 items). The items were scored on a 7-point frequency rating scale ranging from 0 (never) to 6 (daily). High scores on emotional exhaustion and depersonalization and low scores on personal accomplishment are indicative of burnout. The items of personal accomplishment were reversed. We included persons with a maximum of two missing values on the emotional exhaustion scale, one missing value on the cynicism scale, and two missing values on the personal accomplishment scale. The missing values of a respondent were replaced by the mean of the existing values of the respondent on the dimension in question. The reliability (Cronbach's α) of the whole inventory was 0.89.
105
In order to assess the level of burnout, we calculated a weighted sum score of the dimensional scores so that exhaustion, depersonalization and diminished personal accomplishment had different weights in the syndrome (Kalimo et al., 2003). This syndrome indicator was derived with the help of a discriminant function analysis in which various health-related indicators were used as dependent variables (Kalimo and Toppinen, 1997). Coefficients were formed by weighing each dimension so that the scores corresponded to the original response scale (0.4 × exhaustion + 0.3 × depersonalization + 0.3 × diminished personal accomplishment). Burnout was categorized as follows: no burnout (scores 0–1.49), mild burnout (scores 1.50– 3.49), and severe burnout (scores 3.50–6). This categorization meant that burnout was severe when symptoms were experienced approximately once a week or daily, they were mild when symptoms existed monthly, and there was no burnout when the symptoms were experienced only a few times a year or never (Kalimo et al., 2003). Burnout was dichotomized as no versus mild to severe. The short form of the Beck Depression Inventory (BDI) was used to assess depression (Beck and Beck, 1972). It consists of 13 items that are scored from 0 to 3 (α = 0.84). An acceptable answer was expected for at least 11 items. Missing values (2 at the most) were replaced by the mean of the existing values for that particular respondent. A sum score for the depressive symptoms was then calculated. Depressive symptoms were categorized as no (0–4 points), mild (5–7 points), moderate (8–15 points), and severe (16–39 points) (Beck and Beck, 1972). Depression was dichotomized as no versus mild, moderate or severe.
Fig. 1. The research models to test mediational relationships between job strain, burnout, and depression.
106
K. Ahola, J. Hakanen / Journal of Affective Disorders 104 (2007) 103–110
Job strain was measured with the Job Content Questionnaire (JCQ) (Karasek et al., 1998). A short scale of job demands comprised three items (α = 0.79; e.g. “My job requires working very fast”), and the scale for job control had nine items (α = 0.85) (e.g., “My job allows me to make a lot of decisions on my own”; “My job requires a high level of skills”). Responses were given on a 5-point scale ranging from 1 (strongly agree) to 5 (strongly disagree). We created an indicator for job strain as a continuous quotient of job demands distributed by job control (Landbergis et al., 1994). Information on age, sex, marital status, professional sector (private/public), type of employment (permanent/ fixed-term), job tenure, supervisory position (no/yes), and working hours were included in the questionnaire. Age was categorized as under 36, 36–45, 46–55, and over 55 years. Marital status was categorized as married or cohabiting, divorced or widowed, and unmarried. Job tenure was categorized as under 5, 5–9, 10–19, 20–29, and over 29 years. Working hours were categorized as full-time (at least 35 h a week), shortened (25–34 h a week), and part-time (no more than 24 h a week). 2.3. Statistical analysis The entire study sample was described with distributions of sociodemographic factors, work characteristics, burnout, and depression. Among those who were free of depressive symptoms at baseline, binary logistic regression analyses were used to explore the prospective associations concerning the research model A (Fig. 1): between burnout at baseline and depression at follow-up (path 1), between job strain at baseline and burnout at follow-up (path 2), and between job strain at baseline and depression at follow-up (path 3). The analyses were adjusted for sex, age and marital status. The final adjustment included also burnout at baseline as a continuous variable in the models with job strain. Interaction term was applied in the models to test whether job strain had an interaction effect with sex on depression. Concerning the research model B among those who were free of burnout at baseline (Fig. 1), the prospective associations between depression at baseline and burnout at follow-up (path 4), job strain at baseline and depression at follow-up (path 5), and job strain at baseline and burnout at follow-up (path 6) were also analysed with binary logistic regression models, which were adjusted for sex, age and marital status. The final adjustment included depression at baseline as a continuous variable in the models with job strain. Interaction term was applied to the models to test whether job strain had an interaction effect with sex on burnout.
3. Results The dentists were mostly women (74%), married or cohabiting (84%), and in permanent employment (94%). Table 1 shows also that a slight majority of participants worked full-time (63%) and in the public sector (62%). Of all participants, 51% were free of burnout and 72% were free of depressive symptoms at baseline (Table 2). 3.1. Research model A: job strain → burnout → depression The dentists who were free of depressive symptoms but reported symptoms of burnout at baseline, showed depressive symptoms at follow-up in 23% of the cases
Table 1 Baseline characteristics of the study population (n = 2555) Characteristic Sex Women Men Age 26–35 years 36–45 years 46–55 years 56–73 years Marital status Unmarried Married or cohabiting Divorced or widowed Missing data Professional sector Public Private Missing data Employment Permanent Fixed-term Missing data Job tenure 0–4 years 5–9 years 10–19 years 20–29 years 30 years or more Missing data Supervisory position Yes No Missing data Working hours Full-time Shortened Part-time Missing data
n (%) 1883 (74) 672 (26) 384 (15) 898 (35) 884 (35) 389 (15) 179 (7) 2143 (84) 220 (9) 13 (1) 1578 (62) 956 (37) 21 (1) 2392 (94) 141 (6) 22 (1) 144 (6) 304 (12) 805 (32) 860 (34) 431 (17) 11 (0) 677 (26) 1790 (70) 88 (3) 1606 (63) 542 (21) 386 (15) 21 (1)
K. Ahola, J. Hakanen / Journal of Affective Disorders 104 (2007) 103–110 Table 2 Job strain, burnout and depression at baseline and at follow-up (n = 2555)
Job strain, mean (S.D.) Missing data, n (%) Burnout, n (%) No Mild Severe Missing data Depression, n (%) No Mild Moderate Severe Missing data
Baseline
Follow-up
0.78 (0.28) 139 (5)
0.81 (0.30) 135 (5)
1313 (51) 1120 (44) 56 (2) 66 (3)
1173 (46) 1194 (47) 68 (3) 120 (5)
1840 (72) 350 (14) 296 (12) 28 (1) 41 (2)
1774 (69) 382 (15) 291 (11) 28 (1) 80 (3)
107
Table 4 Prospective associations of job strain at baseline for burnout at 3-year follow-up Model 1
Model 2
Sex
n/cases
OR (95% CI)
OR (95% CI)
Men Women
337/71 861/238
27.87 (6.46–120.2) 4.87 (2.46–9.64)
22.31 (5.08–98.07) 3.99 (1.99–7.99)
Model 1 adjusted for sex, age and marital status at baseline. Model 2 adjusted for sex, age, marital status and depression at baseline. OR indicates odds ratio. CI indicates confidence interval.
3.2. Research model B: job strain → depression → burnout
S.D. indicates standard deviation.
compared with 10% of those who did not report burnout symptoms at baseline. In this sub-population the adjusted probability of showing depressive symptoms at follow-up after having burnout at baseline was 2.6fold (95% CI 2.0–3.5) compared to not having burnout at baseline (path 1 in Fig. 1). Job strain predicted burnout among those free of depression at baseline (path 2 in Fig. 1). The adjusted probability of having burnout was 11.8-fold (95% CI 7.4– 18.7) for each one-point increase in job strain score in this group of dentists. When the level of burnout at baseline was also adjusted for in the final model, the probability of having burnout at follow-up decreased but remained statistically significant (OR 1.8, 95% CI 1.04–3.1). Table 3 shows that job strain predicted also depression (path 3 in Fig. 1). When the level of burnout at baseline was adjusted for in the final model, the effect of job strain on depression disappeared, thus showing that burnout fully mediated the impact of job strain on depressive symptoms. There were no significant interactions (p N 0.3) between sex and strain in the models for depressive symptoms.
Table 3 Prospective associations of job strain at baseline for new cases of depression at 3-year follow-up
Job strain
Total n/cases
Model 1
Model 2
OR (95% CI)
OR (95% CI)
1684/242
3.39 (2.03–5.66)
1.30 (0.73–2.30)
Model 1 adjusted for sex, age and marital status at baseline. Model 2 adjusted for sex, age, marital status and burnout at baseline. OR indicates odds ratio. CI indicates confidence interval.
Among the dentists who were free of burnout at baseline, but reported depressive symptoms, burnout was present at follow-up in 63% of the cases compared with 20% among those who did not show symptoms of depression at baseline. In this sub-population the adjusted probability of having burnout at follow-up after having experienced depressive symptoms at baseline was 2.2-fold (95% OR 1.4–3.4) compared to the situation of not having experienced depressive symptoms at baseline (path 4 in Fig. 1). Job strain predicted depression among those free of burnout at baseline. The adjusted probability of having depression was 7.5-fold (95% CI 3.5–15.9) for each one-point increase in job strain score (path 5 in Fig. 1). When the level of depression at baseline was adjusted for in the final model, the probability of showing depression at follow-up decreased but remained statistically significant (OR 4.8, 95% CI 2.1–11.0). A significant interaction (p = 0.03) between sex and strain was found in the model of strain on burnout in this sub-group. Table 4 shows that job strain predicted burnout both among the men and among the women (path 6 in Fig. 1). After adjustment for the level of depression at baseline, these relationships decreased but remained statistically significant. This result indicates that job strain had a direct as well as an indirect effect through depressive symptoms on burnout. 4. Discussion To our knowledge, this is the first study to investigate the temporal relationship between job strain, occupational burnout and depressive symptoms in a prospective setting. We found that the effects between burnout and depression are reciprocal: Occupational burnout predicted new cases of depressive symptoms and depression
108
K. Ahola, J. Hakanen / Journal of Affective Disorders 104 (2007) 103–110
predicted new cases of burnout. In addition, burnout mediated fully the relationship between job strain and depression. This means that after controlling for the mediating effect of burnout, there was no direct relationship between job strain and the incidence of depression; job strain was related to burnout, which in turn was related to depression. Instead, depression mediated only partially the relationship between job strain and burnout indicating both direct and mediated relationship between job strain and burnout. On the basis of the present study, it seems possible that occupational burnout is a phase in the development of work-related depression. This view has earlier been supported by cross-sectional data on the general population (Ahola et al., 2006). The results of our study agree also with earlier studies showing burnout to mediate between psychosocial factors and health. In these studies, health was indicated by the duration of company-registered sickness absences (Bakker et al., 2003), psychosomatic health complaints (Schaufeli and Bakker, 2004), and selfrated health and work ability (Hakanen et al., 2006). In addition to burnout leading to depression, depression also predisposed to burnout. There are at least two main mechanisms to explain this reversed effect (de Lange et al., 2004). Firstly, those who are depressed probably have lower resources to meet the demands of their work, and this situation predisposes them to burnout. Secondly, it is possible that those who are depressed perceive and evaluate their situation at work more negatively than those who are free of depression. On the basis of the present study, the path of job strain predisposing to depression through experienced burnout appear stronger than the path of job strain predisposing to burnout via depression. The extent to which burnout differs from depressive symptomatology has been debated in the past (Taris, 2006). In studies conducted to rule out the possibility of redundancy, it has been recommended to include wellvalidated measures of burnout and depressive symptoms to ascertain that burnout and its correlates are not due to the effects of depression (Shirom, 2005). The present longitudinal study supports the distinction between burnout and depression phenomena. Furthermore, it contradicts the suggestion of burnout and depression developing from the same risk factors simultaneously, “in tandem”, as McKnight and Glass (1995) have suggested on the basis of their small sample (n = 100) of nurses. However, further research is needed to fully understand the cross-lagged reciprocal associations between burnout and depression. As a risk factor for burnout, job strain was emphasized especially among healthy men in our study. The impact of chronic stressors and psychosocial factors on health has
generally been stronger for women than for men (Denton et al., 2004). However, work-related factors such as high job strain and occupational burnout have associated with mental disorders especially among men (Ahola et al., 2005; Virtanen et al. 2007). This finding may reflect the significance work has in men's lives. At least two limitations of this study are noteworthy. Firstly, this study was based on survey data that present problems of self-report bias and common method variance (Lindell and Whitney, 2001). However, variance in self-report measures of job conditions has been largely attributed to variations in the objective work environment (Spector, 1992). Subjective evaluation of work characteristics has also been shown to be more strongly related to mental well-being than objective evaluation is (Stansfeld et al., 1999). Still, it is well-known that common method variance may artificially inflate associations through factors such as negative affectivity and social desirability (Parkes, 1990; Heinisch and Jex, 1997). Ahola et al. (2006) found that a relationship between high job strain and depression, as well as between burnout and depression, appeared regardless of the depression measure used, although the associations were stronger with self-reported variables than with the use of a structured psychiatric interview. Thus we argue that it is unlikely that common method variance has caused major confounding in our study, which employed a prospective design, included only “healthy workers” at baseline concerning the outcome variable, and controlled for the baseline values with respect to the mediating variable. Nevertheless, the impact of common method variance should be further examined in future research by combining other measures with self-report inventories in a prospective setting. Secondly, the study population comprised only one occupational group, the dentists. Our sample was large and representatively comprised the majority of working dentists in Finland. Dentists practice human service work and, as for socioeconomic status, belong to the group of upper white-collar workers. Therefore, the results must be generalized with caution outside this reference group. On the other hand, the results of our study agree with the findings of a previous crosssectional population-based study on the relationship between job strain, burnout and depression, which included 30- to 64-year-old employees from all types of occupations (Ahola et al., 2006). Because our study included dentists who were working and who responded to both questionnaires, it is possible that this procedure excluded the dentists with the worst situations and thus resulted in weakened associations. However, we were able to reach 71% of the whole profession in the first
K. Ahola, J. Hakanen / Journal of Affective Disorders 104 (2007) 103–110
phase of the study, and at follow-up the respondents and the non-respondents did not differ in relation to burnout or depressive symptoms. Sickness absence and disability pensions granted on the basis of depressive disorders have increased in Finland during the present millennium (Salminen, 2004). Among intensive human service work, like dentistry, the importance of mental well-being at work is evident. If human service professionals are to perform well and keep working until statutory retirement age, their working conditions should be continuously evaluated. In primary health care, and especially in occupational health services, burnout could be considered a serious alarm signal of an unfavourably developing working situation. Among employed clients, assessment procedures and early intervention practices concerning workrelated well-being should be further developed. In conclusion, job strain predisposes to depression through burnout. In comparison, job strain predisposes to burnout directly and via depression. Even though the relationship between burnout and depression is reciprocal, the path from burnout to depression appears to be stronger than the path from depression to burnout. 5. Contributors K. Ahola and J. Hakanen together planned the design, J. Hakanen gathered the data, K. Ahola analysed the data and wrote the first version of the manuscript, and K. Ahola and J. Hakanen together refined the manuscript. Acknowledgements This study was supported by the Finnish Work Environment Fund (project no. 105325) and the Finnish Dental Association. The Finnish Work Environment Fund and The Finnish Dental Association had no further role in the study design, the analysis and interpretation of the data, writing or submitting the manuscript. References Ahola, K., Honkonen, T., Isometsä, E., Kalimo, R., Nykyri, E., Aromaa, A., Lönnqvist, J., 2005. The relationship between jobrelated burnout and depressive disorders — results from the Finnish Health 2000 Study. J. Affect. Disord. 88, 55–62. Ahola, K., Honkonen, T., Kivimäki, M., Virtanen, M., Isometsä, E., Aromaa, A., Lönnqvist, J., 2006. Contribution of burnout to the association between job strain and depression: the Health 2000 Study. J. Occup. Environ. Med. 48, 1023–1030. Bakker, A.B., Schaufeli, W.B., Demerouti, E., Janssen, P.P., van der Hulst, R., Brouwer, J., 2000. Using equity theory to examine the difference between burnout and depression. Anxiety Stress Coping 13, 247–268.
109
Bakker, A.B., Demerouti, E., de Boer, E., Schaufeli, W.B., 2003. Job demand and job resources as predictors of absence duration and frequency. J. Vocat. Behav. 62, 341–356. Beck, A.T., Beck, R.W., 1972. Screening depressed patients in family practice. A rapid technique. Postgrad. Med. 52, 81–85. Belkic, K., Landsbergis, P.A., Schnall, P.L., Baker, D., 2004. Is job strain a major source of cardiovascular disease risk? Scand. J. Work Environ. Health 30, 85–128. Borritz, M., Bültmann, U., Rugulies, R., Christensen, K.B., Villadsen, E., Kristensen, T., 2005. Psychosocial work characteristics as predictors for burnout: findings from 3-year follow up of the PUMA study. J. Occup. Environ. Med. 47, 1015–1025. Brenninkmeyer, V., Van Yperen, N.W., Buunk, B.P., 2001. Burnout and depression are not identical twins: is decline of superiority a distinguishing feature? Pers. Individ. Differ. 30, 873–880. Cox, T., Tisserand, M., Taris, T., 2005. The conceptualization and measurement of burnout: questions and directions. Work Stress 19, 187–191. Cropley, M., Steptoe, A., Joekes, K., 1999. Job strain and psychiatric morbidity. Psychol. Med. 29, 1411–1416. de Lange, A.H., Taris, T.W., Houtman, I.L.D., Kompier, M.A.J., Bongers, P.M., 2003. “The very best of the millennium”: longitudinal research and the demand–control–(support) model. J. Occup. Health Psychol. 8, 282–305. de Lange, A.H., Taris, T.W., Kompier, M.A.J., Houtman, I.L.D., Bongers, P.M., 2004. The relationships between work characteristics and mental health: examining normal, reversed and reciprocal relationships in a 4-wave study. Work Stress 18, 149–166. FDA, 2005. Facts about the number of dentists [Lukumäärätietoja hammaslääkäreistä]. (http://www.hammasll.fi/). Downloaded 10 June 2005. Freudenberger, H.J., 1974. Staff burn-out. J. Soc. Issues 30, 159–165. Glass, D.C., McKnight, J.D., 1996. Perceived control, depressive symptomatology, and professional burnout: a review of the evidence. Psychol. Health 11, 23–48. Golembiewski, R.T., Munzenrider, R.F., Scherb, K., Billingsley, W., 1992. Burnout and psychiatric “cases”: early evidence of an association. Anxiety Stress Coping 5, 69–78. Hakanen, J.J., 2004. Hammaslääkäreiden työhyvinvointi Suomessa [Work-Related Well-Being Among Finnish Dentists]. Finnish Institute of Occupational Health and the Finnish Dental Association, Helsinki. Hakanen, J.J., Bakker, A.B., Demerouti, E., 2005. How dentists cope with their job demands and stay engaged: the moderating role of job resources. Eur. J. Oral Sci. 113, 479–487. Hakanen, J.J., Perhoniemi, R., 2006. Hammaslääkäreiden työhyvinvointi Suomessa. Kolmen vuoden seurantatutkimus [Work-Related Well-Being Among Finnish Dentists. Three-Year Follow-Up Study]. Finnish Institute of Occupational Health and the Finnish Dental Association, Helsinki. Hakanen, J.J., Bakker, A.B., Schaufeli, W., 2006. Burnout and engagement among teachers. J. Sch. Psychol. 43, 495–513. Heinisch, D.A., Jex, S.M., 1997. Negative affectivity and gender as moderators of the relationship between work-related stressors and depressed mood at work. Work Stress 11, 46–57. Honkonen, T., Ahola, K., Pertovaara, M., Isometsä, E., Kalimo, R., Nykyri, E., Aromaa, A., Lönnqvist, J., 2006. The association between burnout and physical illness in the general population — results from the Finnish Health 2000 Study. J. Psychosom. Res. 61, 59–66. Iacovides, A., Fountoulakis, K.N., Moysidou, Ch., Ierodiakonou, Ch., 1999. Burnout in nursing staff: is there a relationship between depression and burnout? Int. J. Psychiatry Med. 29, 421–433.
110
K. Ahola, J. Hakanen / Journal of Affective Disorders 104 (2007) 103–110
Iacovides, A., Fountoulakis, K.N., Kaprinis, St., Kaprinis, G., 2003. The relationship between job stress, burnout and clinical depression. J. Affect. Disord. 75, 209–221. Kalimo, R., Toppinen, S., 1997. Työuupumus Suomen työikäisellä väestöllä [Burnout Among Finnish Working Population]. Työterveyslaitos, Helsinki. Kalimo, R., Pahkin, K., Mutanen, P., Toppinen-Tanner, S., 2003. Staying well or burning out at work: work characteristics and personal resources as long-term predictors. Work Stress 17, 109–122. Karasek, R.A., 1979. Job demands, job decision latitude and mental strain: implications for job redesign. Adm. Sci. Q. 24, 285–308. Karasek, R.A., Brisson, C., Kawakami, N., Houtman, I., Bongers, P., Amick, B., 1998. The job content questionnaire (JCQ): an instrument for intentionally comparative assessments of psychosocial work characteristics. J. Occup. Health Psychol. 3, 322–355. Kenny, D.A., 2005. Mediation. In: Everitt, B.S., Howell, D.C. (Eds.), Encyclopedia of Statistics in Behavioral Science, vol. 3. John Wiley, Chichester, pp. 1194–1198. Kivimäki, M., Leino-Arjas, P., Luukkonen, R., Riihimäki, H., Vahtera, J., Kirjonen, J., 2002. Work stress and risk of cardiovascular mortality: prospective cohort study of industrial employees. Br. Med. J. 325, 857. Kivimäki, M., Virtanen, M., Elovainio, M., Kouvonen, A., Väänänen, A., Vahtera, J., 2006. Work stress in the etiology of coronary heart disease: systematic review and meta-analysis of prospective cohort studies. Scand. J. Work Environ. Health 32, 431–442. Landbergis, P.A., Schnall, P.L., Warren, K., Pickering, T.G., Schwatrz, J.E., 1994. Association between ambulatory blood pressure and alternative formulations of job strain. Scand. J. Work Environ. Health 20, 349–363. Leiter, M.P., Durup, J., 1994. The discriminant validity of burnout and depression: a confirmatory factor analytic study. Anxiety Stress Coping 7, 357–373. Lindell, M.K., Whitney, D.J., 2001. Accounting for common method variance in cross-sectional research designs. J. Appl. Psychol. 86, 114–121. Maier, S.F., Watkins, L.R., 1998. Cytokines for psychologists: implications of bidirectional immune-to-brain communication for understanding behavior, mood, and cognition. Psychol. Rev. 105, 83–107. Maslach, C., 1976. Burned-out. Human Behav. 5, 16–22. Maslach, C., Jackson, S.E., 1981. The measurement of experienced burnout. J. Occup. Behav. 2, 99–113. Maslach, C., Schaufeli, W.B., 1993. Historical and conceptual development of burnout. In: Schaufeli, W.B., Maslach, C., Marek, T. (Eds.), Professional Burnout: Recent Developments in Theory and Research. Taylor & Francis, London, pp. 1–16. Maslach, C., Jackson, S.E., Leiter, M.P., 1996. Maslach Burnout Inventory Manual, 3rd ed. Consulting Psychologists Press, Palo Alto, CA. Maslach, C., Schaufeli, W.B., Leiter, M.P., 2001. Job burnout. Annu. Rev. Psychol. 52, 397–422. McKnight, J.D., Glass, D., 1995. Perceptions of control, burnout, and depressive symptomatology: a replication and extension. J. Consult. Clin. Psychol. 63, 490–494. Middeldorp, C.M., Cath, D.C., Boosma, D.I., 2006. A twin-family study of the association between employment, burnout and anxious depression. J. Affect. Disord. 90, 163–169.
Niedhammer, I., Goldberg, M., Leclerc, A., Bugel, I., David, S., 1998. Psychosocial factors at work and subsequent depressive symptoms in the Gazel cohort. Scand. J. Work, Environ. Health 24, 197–205. Parkes, K.R., 1990. Coping, negative affectivity, and the work environment: additive and interactive predictors of mental health. J. Appl. Psychol. 75, 399–409. Paterniti, S., Niedhammer, I., Lang, T., Consoli, S.M., 2002. Psychosocial factors at work, personality traits and depressive symptoms. Longitudinal results from the GAZEL study. Br. J. Psychiatry 181, 111–117. Salminen, J.K., 2004. Mielenterveyden häiriöt ja työkyvyttömyys [mental disorders and work disability]. Suom. Lääkäril. 59, 4745–4748. Schaufeli, W.B., Bakker, A.B., 2004. Job demands, job resources, and their relationship with burnout and engagement: a multi-sample study. J. Organ. Behav. 25, 293–315. Schaufeli, W., Enzmann, D., 1998. The Burnout Companion to Study and Practice: A Critical Analysis. Taylor & Francis, London. Shirom, A., 2003. Job-related burnout: a review. In: Quick, J.C., Tetrick, L.E. (Eds.), Handbook of Occupational Health Psychology. American Psychological Association, Washington, DC, pp. 245–264. Shirom, A., 2005. Reflections on the study of burnout. Work Stress 19, 263–270. Spector, P.R., 1992. A consideration of the validity and meaning of self-report measures of job conditions. Int. Rev. Ind. Organ. Psychol. 69, 293–310. Stansfeld, S.A., Fuhrer, R., Head, J., Ferrie, J., Shipley, M., 1997. Work and psychiatric disorder in the Whitehall II study. J. Psychosom. Res. 43, 73–81. Stansfeld, S.A., Fuhrer, R., Shipley, M.J., Marmot, M.G., 1999. Work characteristics predict psychiatric disorder: prospective results from the Whitehall II study. Occup. Environ. Med. 56, 302–307. Taris, T.W., 2006. Bricks without clay: on urban myths in occupational health psychology. Work Stress 20, 99–104. Toker, S., Shirom, A., Shapira, I., Berliner, S., Melamed, S., 2005. The association between burnout, depression, anxiety, and inflammation biomarkers: C-reactive protein and fibrinogen in men and women. J. Occup. Health Psychol. 10, 344–362. Toppinen-Tanner, S., Ojajärvi, A., Väänänen, A., Kalimo, R., Jäppinen, P., 2005. Burnout as a predictors of medically certified sick-leave absences and their diagnosed causes. Behav. Med. 31, 18–27. van der Doef, M., Maes, S., 1999. The job demand–control(–support) model and psychological well-being: a review of 20 years of empirical research. Work Stress 13, 87–114. Virtanen, M., Honkonen, T., Kivimäki, M., Ahola, K., Vahtera, J., Aromaa, A., Lönnqvist, J., 2007. Work stress, mental health and antidepressant medication: findings from the Health 2000 Study. J. Affect. Disord. 98, 189–197. Weber, A., Jaekel-Reinhard, A., 2000. Burnout syndrome: a disease of modern societies. Occup. Med. 50, 512–517. Ylipaavalniemi, J., Kivimäki, M., Elovainio, M., Virtanen, M., Keltikangas-Järvinen, L., Vahtera, J., 2005. Psychosocial work characteristics and incidence of newly diagnosed depression: a prospective cohort study of three different models. Soc. Sci. Med. 61, 111–122.