Stress and well-being at work: a literature review

The study presents empirical evidence of the personal and contextual variables that protect against the negative effects of stress on workers’ well-being and health. A systematic review of the literature of the last eleven years (2006 to 2016) was carried out in the Ebsco, LILACS, SciELO, Google Scholar, and PubMed databases, and in six Annual Reviews journals. Fifty national and international articles related to stress and well-being were analyzed. The results show that work-related factors, personal resources such as resilience, self-efficacy, emotional competencies, psychological detachment, work-life interface, and group-level factors have effects on well-being. The perceived social support of peers and managers, alongside autonomy at work, attenuate the negative impact of stress on well-being. It ends with suggestions for future studies to test relationships between variables that have not yet been explored and to deepen the understanding of the relationships between stress and well-being at work.

introduction Stress is one of the concepts most studied and mentioned in psychology, since it is manifested in everyday life as a factor that can put the psychological and physical health of people at risk [1][2][3][4] . The three main perspectives for approaches to stress are: i) one that emphasizes the physiological responses and reactions of the individual to the stressors; ii) one that seeks to investigate the stress factors in the environment that provoke damage to the well-being and health of individuals; and iii) the psychological/cognitive one that focuses on the relationship between the individual and the environment, that is, how the person evaluates and perceives the stressor event 5 . This last approach is cited by most of the articles studied [6][7][8][9] .
According to Lazarus 10 , the way we evaluate an event determines how we react emotionally. Thus, work-related stress is influenced by the perception that the subject has of the demands in the work environment and his/her capacity to face them. In other words, in order for the work-related stress process to occur, workers have to evaluate the situation and the demands of the environment as stressors, and see themselves as lacking the resources to deal with them, generating reactions with negative effects on their well-being. Considering the complexity of the stress phenomenon, permeated by a series of psychological, social, and biological processes that involve human-environmental interaction, recent studies inspired by the Positive Psychology movement have been trying to understand how these processes can interfere in the health and well-being of people [11][12][13] .
The movement known as Positive Psychology 14 emphasizes the positive aspects of human experience, and seeks to understand the qualities that help individuals develop their potential, maintain their physical and psychological health, and their personal well-being. This movement has also spread among the scholars of organizations, giving rise to a variety of studies developed in the work context, emphasizing the role of traits, positive affective states, behaviors, and virtues that predispose individuals to experience well-being at work (WBW) 15 . Work-related well-being can be said to encompass a diversity of experiences that include positive affective states (e.g., enthusiasm), low levels of negative affective states (e.g., anxiety), good psychosomatic health and cognitive states such as aspirations and judgments about job satisfaction 16 .
In addition to stress factors stemming from the workplace, personal resources such as self-esteem and emotional stability, and the workhome interface seem to influence the fluctuations in indicators of well-being 17 . The experience of stress at work has been consistently associated with negative results for employees, measured by a range of well-being indicators, both psychological and physiological [18][19][20] . However, in confronting events perceived as stressors, these individuals react differently. Although many studies conclude that there is a negative association between stress and well-being at work 1,[18][19][20][21][22] , there has been a growing number of studies that seek to explain under what conditions stressors may not adversely affect WBW 17,[23][24][25] . Various studies highlight the important role of resilience and the strategies used to control and deal with one's emotions (emotional self-regulation), and to cope with stressful situations, which can be effective in reducing stress and improving well-being in various contexts [26][27][28][29][30] .
The systematic review presented in this article sought to explore and understand the relationships that exist between stress and well-being at work, attempting to find evidence of which personal and/or environmental variables have protective effects for the well-being of workers, with positive impacts on their physical, mental, and social health. The study contributes to an overall view of the state of the art in this field of research, a critical assessment of the results obtained, as well as generating inputs to support proposals for improving well-being at work (WBW) and reducing stress in organizations. It also discusses the relevant theoretical and practical implications based on the results found and suggests directions for future research in the field.

Data collection and analysis procedure
The first step of searching for articles was carried out during June and July 2016, in the following databases: Ebsco, Google Scholar, Lilacs (Latin American and Caribbean Literature on Health Sciences), SciELO (Scientific Electronic Library Online), and PubMed (National Library of Medicine). A search was also done on six Annual Reviews journals (Annual Review of Environment and Resources, Law and Social Science, Neuroscience, Organizational Psychology, Psychology, Public Health). The databases were chosen because they cover a large and diversified volume of scientific publications representative of the diverse fields of knowledge, since stress is a well-studied construct. The Annual Reviews are publications dedicated to literature review articles, which would help identify the evolution of the findings on stress, well-being, and work.
The key words used for the article search were: "stress", "wellbeing", "job", and "work" in English, aiming to cover the national [Brazilian] and international literature. The database searches were configured to retrieve the words in the "title", "summary", and "keywords" fields. The Boolean operator "and" was used to request the selected words to appear in those fields. It was determined that the databases should present the articles published from 2006 to 2016, covering the last 11 years, a period with an intensification of studies on well-being in the context of work, driven by the Positive Psychology movement. In the PubMed database, specifically, we selected only review and meta-analysis articles in order to filter the initial search, which resulted in 3,996 titles. Through these procedures, a total of 675 references were obtained.
In the second step, the titles and abstracts of these works were read, excluding those that did not correspond to the general objective of this review, as well as theses, books, and editorials, with 134 thus remaining. In the third step, during the process of a more detailed reading of these articles, 81 of them were discarded, considering the following exclusion criteria: (a) type of study: theoretical articles; (b) focus of the study: those dealing only with burnout syndrome (antecedents, predictors, and prevalence), with the reverse effect between stressor and well-being/ tension, those to validate instruments or evaluate intervention programs for stress reduction; (c) measurement: articles in which well-being was measured only by job satisfaction; (d) sample type: consisting of university students, elderly, or young people not included in the labor market. In the final step, 50 articles were selected that met the following inclusion criteria: (a) type of study: empirical scientific review or meta-analysis articles; (b) year of publication: between 2006 and 2016; (c) focus of the study: dealing with the relations between stressors/stress and well-being at work/of workers, including moderating and mediating variables or not; (d) measurement: assessing or measuring stress and/or stress factors at work, and positive and negative aspects of well-being/health (physical, psychological, social, work-related, among others); (e) sample type: workers from various categories; (f) publication languages: Portuguese, English, or Spanish. Figure 1 illustrates these procedures performed to select the material.
For the systematization and analysis of the material, all the articles that met the inclusion criteria were read in full, and their basic information summarized, such as authorship, year of publication, methodological outline, study focus, study variables (independent variable -IV, dependent variable -DV, moderating and mediating variables), and conclusions (correlations between variables, main predictors, etc.). Predictive variables are those capable of predicting effects on another variable, while the moderating variables are those that can affect the direction and strength of the relationship between the IV, or antecedent, and the DV, or consequent variable. The mediator variable is one that, when introduced into the model, makes the relationship between an IV and a DV no longer significant, since the relation between them is better explained by the mediator 31 . The data were collected in tables, separating the review/meta-analysis studies from the empirical studies. The suggestions from the studies in general, regarding future research and applicability of results in organizations, were also summarized and integrated.

results
The articles analyzed are presented in four tables, in order to facilitate visualization and understanding of the data. Charts 1 and 2 present a summary of the review and meta-analysis articles (authorship/year of publication, type of review and sample, IV and DV, other variables, and conclusions). Chart 1 lists eleven studies published between 2011 and 2016, and Chart 2 presents four studies published between 2006 and 2010, totaling fifteen (15). Charts 3 and 4 present the summary of the empirical articles (authorship/ year of publication, type of study and participants, focus of the study, and conclusions). Chart 3 lists nineteen studies published between 2011 and 2016, and Chart 4 presents sixteen studies between 2006 and 2010, totaling thirty-five (35). The main positive and negative predictors of well-being at work (WBW), the environmental and individual moderators, the mediators, the positive predictors of tension, physical and psychological symptoms, and health problems are highlighted in the conclusions. The studies are arranged in descending order of publication year, with the most recent ones first, organized in alphabetical order by author name.
Considering the totality of the selected articles, it is observed that from the review and meta-analysis studies (Chart 1 and 2), eleven were published between 2011 and 2016 (73.3%), and only four pertain to the previous period (2006 to 2010), indicating a considerable increase in scientific production in the last five years. As to the empirical studies (Chart 3 and 4), nineteen appear in this more recent period (54.3%).
As for the method, Charts 3 and 4 clearly indicate the prevalence of a quantitative approach using self-reported scales in the empirical studies, corresponding to 82.8% (N = 29). Only six qualitative studies were found. A significant difference was observed regarding the inclusion of moderators and mediators, being considered in 60% of the review and meta-analysis studies and in only 31% of the quantitative empirical studies.

review and Meta-analysis Studies: predictors
The main negative predictors of well-being at work (WBW) are stressors and lack of resources at work: pressure and overload 17,22,23,25 , high demands and low control 18,20 , monotony and low decision-making power 20 , negative social interactions and lack of social support 17,20 , and negative affective events at work 23 . Individual factors such as low psychological detachment during time off 25 , negative emotions and dysfunctional emotion regulation (ER) 32 , and work-family conflict 17,23 were cited less often.
As positive predictors of WBW, personal resources such as self-esteem, active coping, opti- DV: Burnout and Quality of Work Life -QWL (satisfaction and well-being at work). -Burnout predictors: stressors such as high demands and lack of resources at work (low control, lack of support from supervisors/coworkers, and lack of training).
-Relevant factor that minimized the impact of stressors at work, improving QWL: perception of the work as rewarding.

/2015
Theoretical and empirical studies, cross-sectional and longitudinal, and meta-analyses. Workers.
Moderators: control at work, social support, other resources at work, positive affects, psychobiological recovery process. -Predictors (-) of WBW: psychosocial stressors at work, negative affective events at work, and the experience of work-family conflict.
-Moderators that minimize the effects (-) of the demands and stressors on WBW: control at work, social support and other resources at work, positive affects, and psychobiological recovery process (psychological detachment from work, experiences of relaxation, control, and self-mastery).
Workers. IV: Work factors (stressors, resources, and interpersonal relations at work); Personal resources; Work-home interface.
Moderators: Environment: positive events, autonomy, support at work, climate of security. Individuals: emotional stability, self-regulation strategies. -Predictors (-) of WBW: stressors at work, negative social interactions and lack of social support at work, work-home conflict.
-Moderators of the environment that mitigate the effects of stressors on WB: events (+), autonomy, perception of support and climate of psychological security at work.
-Individual moderators that mitigate the stressors: emotional stability and use of self-regulation strategies. -Long-term predictors (-) of WB: experience of negative emotions and habitual dysfunctional ER (suppression).
-Moderator between experience and ER that mitigates the negative impacts of affective events on the results of WB: EI.
-Moderators between experience and emotional expression, with impact (+) on relationships: EI and rules of expression. mism and self-efficacy 17 , affective commitment at work 33 , the experience of positive emotions, and functional emotion regulation (ER) 32 are highlighted. Resources at work 4 , which refer to the physical, psychological, social, or organizational aspects of the work environment that are aimed at achieving goals and reducing psychological costs, such as social support at work 17 , and positive affective events at work 23 , were pointed out. Group factors as social cohesion and collective efficacy were cited in only one study 20 .
As predictors of tension, symptoms, and physical/psychological disorders, chronic stressors, such as high demands, lack of resources and social support 20,32,34 , and low psychological detachment from work 1 stand out (Chart 1).

review and Meta-analysis Studies: moderators
Environmental and individual moderators that ease the negative impact of stressors on workers' well-being deserve attention. The principal moderators of the environment are the resources at work, such as control and social support at work 11,18,23 , positive events at work, worker autonomy, perception of organizational support, and the climate of security 17 .
Whereas personal resources, in addition to reducing the negative repercussion of stressors, increase the positive effect of organizational resources on well-being at work (WBW) 4 . These are: emotional stability and self-regulation strat-

Type of review and Sample iv and Dv Other variables Conclusions
Häusser et al. 11

Conclusions
Bell et al. 21 /2012 Correlational and crosssectional. Academics from Australian Universities. Investigate the relationship between perceived stress at work (threat and pressure), work-life interface, and well-being/malaise. -Predictor (-) of WB and (+) malaise: threat stress.
-Predictor of work-life conflicts: threat and pressure stress at work.
Bell et al. 36 /2012 Correlational and crosssectional. Academics from Australian Universities. Investigate the influence of spirituality at work (SAW) on WB and stress, and the moderating effect of SAW on the relationship between stress and WB/ malaise. -Predictors of resilience: emotional (EI and empathy) and social competencies.
-Resilience was a predictor (-) of psychological distress and mediator between EI and PWB.
Lee et al. 56 /2011 Correlational and crosssectional. Hospital nurses in Taiwan. Analyze the relationships between stress at work, coping strategies, and promotion of a healthy lifestyle. -> Stress was associated with > use of coping and < with healthy lifestyle promotion.
-< stress was associated with > coping for control and support and > healthy lifestyle (health responsibility and keeping updated). Pharmacists, South Africa. Investigate whether stress at work and coping strategies predict WBW (burnout/engagement). -Predictor (+) of WBW: > problem-focused coping and < avoidance coping.
-Predictor (-) of WBW: high demands, lack of resources, and avoidance coping. workers in Japan. Examine the effects of problem-focused coping and of distraction on well-being (responses to stress; performance at work). -> Use of problem-focused coping is associated with < stress in individuals with high distraction, and > stress in those who had low distraction.
-There was a short-term effect (+) of problem-focused coping on WB and performance it continues

Authorship/ Publication
Year

Conclusions
Uncu et al. 57 /2007 Correlational and crosssectional. Attending physicians in Turkey. Analyze the relationships between affective wellbeing at work and stress, anxiety, and depression. -Affective WB: correlation (-) with stress, depression, and anxiety.
-Affections (-) at work correlate with stress, anxiety, and depression. workers. Test the relationship between resilience and social support and the effects on stress and satisfaction at work, and the influence of negative affects on the relationship between stress and satisfaction at work. -Resiliency resources, support from supervisor and coworkers has a protective function that minimizes stress and improves satisfaction at work.
-The > the negative affects, the < the job satisfaction.
-When the other variables were included as predictors in the model, there was no correlation between negative affects and job satisfaction.
Source: The authors. egies 17 , emotional intelligence (IE) 32 , greater psychological detachment 25 , and age. Elderly adults are more effective in avoiding and regulating negative affects resulting from work events, reducing tension 9 (Chart 2).

empirical Studies: predictors
As negative predictors that affect work well-being (WBW), here we also see mainly stressors and lack of resources at work: psychosocial occupational risks 35 , perceived stress at work 2 , occupational stress 1 , stress from psychological threat 21,36 , high demands and lack of resources 6 , illegitimate tasks 37 (seen as unnecessary), and lack of social support 38 . Reduced managerial support was associated with a greater risk to mental health in the study by Chen et al. 1 . Factors related to affection and its regulation were pointed out in fewer number, such as destructive emotions 39 , maladaptive stress coping strategies 40 , and avoidance coping 6 .
Notable positive predictors of WBW include personal resources such as high psychological capital (self-efficacy, optimism, and resilience) and constructive emotions 39 , emotional competencies 41 , increased awareness and management of emotions 42 , emotional intelligence and willingness to flow in work 24 , in addition to proactive problem-focused coping strategies 6,26,40,[43][44][45] . Work resources such as organizational support, growth, and career opportunities proved to be WBW promoters in the research by Jackson et al. 46 .
The main predictors of malaise, tension, psychological stress, and other physical and psychological disorders were: psychosocial risk factors at work, that is, characteristics and conditions at work that act as stressors 35,47,48 , and notable among these are illegitimate tasks 37 , threat stress 21 , time and deadline pressure, work overload, and high demands and low resources 7,49,50 . The consequences of constant exposure to stressors can be detrimental to well-being, damaging workers' health 7,35,49,51 (Chart 3).

empirical Studies: moderators and mediators
Some authors investigated the protective factors against stress (moderators), that is, those personal or environmental characteristics, or the perception held about them, that can reduce the negative effects that the stress process can have on health and well-being, including: well-defined roles, leadership committed to the job and to the employees, possibility of breaking the routine, quality of resources, teamwork, pleasant environment with good interpersonal relationships 52 ; resiliency and support of supervisor/coworkers 53 , and the use of cognitive re-evaluation strategies in the interaction with coworkers 54 . Emotional competencies had both a protective role against stress as well as being a predictor of resilience and psychological well-being 41 . On the other hand, there were moderators that exacerbated the effect of stressors on well-being, such as the low perception of organizational support 12 , the strategies of emotion regulation (ER) when used in the interaction with people from outside the organization, aimed at meeting occupational or organizational rules of expression (emotional labor) 54 , and palliative or avoidant coping 26 .
The investigation of mediators was found in three studies. The first is that of Jackson et al. 46 who tested and confirmed the mediation of burnout in the relationship between work demands and health problems, and the mediation of work engagement in the relationship between resources and commitment at work. The study by Karlsen et al. 27 concluded that the effect of stress on well-being was positively mediated by a problem-focused coping process, associated with a growth component, leading to improved subjective well-being (SWB), and on the other hand, stress was negatively mediated by an avoidance coping process and a distress component, leading to a worsening of SWB. In the study by Kinman and Grant 41 , resilience mediated the relationship between emotional intelligence (EI) and psychological well-being (PWB). Without imputing causality, they suggest that EI has an indirect effect on PWB through resilience (Chart 4).

Discussion
Analysis of the data reveals a distinction in the way WBW is approached. Some studies cover WBW in a combination of positive and negative aspects 6,23,55 . Others differentiate positive WBW and malaise/tension (negative facet of WBW) by adopting separate measures 4,36 . The measurement instruments addressed positive aspects such as psychological (PWB), affective (AWB), and subjective (SWB) well-being; mental and physical health; engagement at work; satisfaction with life and work; and Quality of Work Life (QWL). Negative aspects were addressed by measuring psychological stress, tension, physical and psychological symptoms/disorders, and burnout. This variety indicates that there is amplitude in the proposed concept/models of well-being related to the complexity of the phenomenon.
It is observed that due to the great diversity of studies, some variables had different roles in the models tested, appearing either as predictors or moderators, positive or negative, of well-being, broadening the understanding of the dynamics of this construct. The lack of social support from peers and leaders was a negative predictor of WBW 17,20 , and its presence was a positive predictor of WBW 17 , also playing a moderating role in the study by McCalister et al. 53 by mitigating the negative effect of stressors. In summary, the lack of social support had a greater negative impact on well-being than its presence contributed to well-being on the positive side 44 . In the study by Lee et al. 56 , the search for social support used as a coping strategy was associated with lower stress and a healthier lifestyle, demonstrating the importance of this factor for employees. The perception of organizational support, in turn, minimized the effect of stressors on WBW 17 , while the low perception of this support had an inverse effect, with negative impact on well-being 12 .
Functional emotion regulation (ER) strategies and emotional competencies were positive predictors of well-being 32,41 as well as protectors against stress in the studies by Sonnentag 17 and Kinman and Grant 41 . In addition, affective well-being was associated with lower levels of stress, depression, and anxiety, in the study by Uncu et al. 57 . In some studies, although moderators were not tested, certain factors may have acted to reduce the effects of stress. This is the case in the study by Goto et al. 58 in which the workers had a positive perception of well-being (satisfaction with life and work), despite the presence of various stressors. It can be assumed that the strategies used to handle them were effective. In the research by Bragard et al. 59 there was high job satisfaction associated with the positive perception of the work itself, although the employees reported many stressors and burnout.
One can see that the work factors (e.g., stressors, lack of resources, difficulties in relationships) were more consolidated as negative predictors that undermine WBW 1,3,6,17,18,[21][22][23]25,[35][36][37][38] , and personal resources (e.g., self-efficacy, emotion regulation) as positive predictors of WBW 6,17,24,32,33,[39][40][41][42][43][44] . According to the Sonnentag 17 review, the results about personal resources are inconclusive for negative indicators of well-being. It is relevant to highlight controversial and ambivalent aspects found about coping strategies and ER. While, on the one hand, the use of problem-focused coping proves to be strongly consolidated in the literature as a positive predictor of WBW 6,40,[43][44][45] , as well as related to lower stress and better health status 58 , on the other, the results related to avoidance coping are contradictory. Although this strategy is mainly associated with negative indicators of well-being and health 6,44 , in some situations it appears to have beneficial effects 40 , which reinforces the idea of the importance of regulatory flexibility depending on the context 60 . That is, the use of different ER strategies, depending on the situation, facilitates adaptation to the stressors. In this sense, Lawrence et al. 32 emphasize that it is not the mere use of certain regulatory strategies considered dysfunctional (e.g., suppression of emotions) that causes damage to health, but their habitual use.

Conclusion
The present review fulfills its objective in providing an overall view of national and international studies that relate stress and well-being at work, highlighting predictors, moderators, and mediators of this relationship. At the national level, it offers an additional contribution, since the two Brazilian review studies that were found analyzed only the predictors of WBW, without considering moderators and mediators. The review also reveals the emergence of new, more complex and sophisticated models, offering new insights into the processes and dynamics of well-being. In addition to the work factors, personal resources (resilience, self-efficacy, emotional competencies, ER, psychological detachment), work-life interface, and group-level factors are included in these models, clarifying the role they play in the relationship between stress and well-being.
Regarding organizational factors, the results indicate that lack of support, of resources, of good relations, and of other aspects related to work have a more negative impact on WBW indicators than their presence contributes to WBW on the positive side. The perception of organizational support and resources at work, especially autonomy and the social support of peers and managers, act as moderators, reducing the negative impact of stress on WBW. Regarding personal resources, they prove capable of letting a person preserve their well-being at work even in adverse conditions. Emotional competencies and functional emotion regulation strategies, besides being predictors of WBW, act as effective moder-ators, minimizing the negative effects of stressors. It is worth mentioning that although some strategies used to deal with stressful situations are considered adaptive or maladaptive, for being associated more frequently with positive or negative indicators of well-being and health, the consequences may vary depending on the context. The problem-focused coping strategy, for example, carries more benefits than that of avoidance, although the latter may be provisionally useful in situations where control over an adverse environment cannot be managed.
Considering the complexity of this issue, more studies should explore intrapersonal ER strategies, testing their different roles (moderators and mediators), in different situations and occupations, different age groups, including as well the emotion regulation strategies of the other people with whom they interact (interpersonal). In addition, it is noted that few mediators have been tested, representing a gap to be filled. Although it has not been the focus of this review, one can note that there are few studies that explore the effectiveness of interventions in personal resources (regulation skills and emotional competencies) and organizational changes, assessing their effects on stress and well-being.
Another suggestion is to test broader and more integrated models, which consider individual variability regarding demands and stressors, and the long-term positive and negative results on health and well-being. For this, more longitudinal studies are necessary to infer causality and also to understand the process and the dynamics of well-being at work. Another gap regarding the method refers to broadening the measurement beyond the self-report, with the use of measures of body responses, observation of work, daily monitoring, multi-level measures (e.g., interpersonal, group), integrating quantitative and qualitative analyses.
Regarding the limitations of this study, one can point out the selection of journals, which occurred in various databases and journals relevant to the area, although there may be others not consulted. And there may also be other levels of analysis and approaches about stress and well-being at work, beyond those used in this study.
As a practical contribution, the results found are relevant and can help support policies and programs to improve well-being and reduce stress of workers, from a more careful look at stressors, work resources, organizational and social support, interpersonal relationships, personal resources, and work-life balance. These factors have proven to be highly important and, if not properly treated, may constitute psychosocial risks with harmful consequences for workers' health.

Collaborations
ALT Hirschle worked on the design, introduction, bibliographic research, methodology, data analysis and interpretation, conclusion and final writing. SMG Gondim worked on analyzing and interpreting the data, guiding and correcting the conclusion, critical review and final writing.