The wellbeing of health-workers is an issue with an enormous human toll and impacts on the sustainability, quality and safety of  health care systems (Dewa et al., 2014). Supporting the wellbeing of health care workers is one of four key aims for a quality health system (Bodenheimer and Sinsky, 2014). The literature has taken three major perspectives: (1) clinical, with a focus on individual psychology, mental health problems (depression, suicide, burnout); (2) organisational, occupational health and organisational psychology; (3) educational, where education has been utilised as an intervention. 

An educational perspective suggests learning as a solution (Bastounis et al., 2016; Stahl, Sharplin and Kehrwald, 2018) but the “wellbeing curriculum” has not been well defined. Educational theory is not well integrated into the occupational health and clinical discourses; and theory of workplace stress are not well integrated into educational discourse around wellbeing.

There is an urgent need for a model of wellbeing in the health workplace that mediates between the different discourses and perspectives taken. An integrated model summarises and mediates the scholarship in multiple domains and helps clinical educators and health managers implement solutions.

Develop a learning model of wellbeing that builds on existing models of health and learning that is fit for purpose to guide interventions.

Phase 1: Identifying the literature

A total of 146 studies were identified by the original search strategy in PUBMED examining reviews of psychological resilience, compassion fatigue and professional burnout. The search was then expanded to include CINAHL (n=96) and Psychinfo (n=41). An initial theme identified was the difference between the individual psychological and organisational perspective, so the search was expanded to include organisational climate and occupational health (n=55). Papers that did not deal with health workers, were opinion pieces or were not focused on the wellbeing of health workers were excluded, producing 136 papers for analysis.  

The identification of literature in Phase 1 is summarised in Figure 2 

Figure 2: Results of the literature search.  

Significant variations in the perspectives taken were identified. Although these perspectives are represented as black and white distinctions, it is acknowledged that this is a simplification and that there are continuums in perspective and significant overlaps. Perspective were classified as: Individual versus organisational; positive versus negative; and global versus focused. Burnout, as an example, is an individual negative and relatively global perspective, contrasting to the positive perspective of resilience and compared to more focal constructs of compassion fatigue or existential burnout/distress. These difference in perspective are summarised in Figure 3.  

Figure 3: Variations in perspectives taken. Global perspectives were divided into a matrix of positive versus negative and individual versus organisational.  

Papers were classified into major and minor themes. The major perspective taken was individual rather than organisational. Major themes were descriptions versus interventions:  Minor themes were: existential issues (n=10); consequences (n=6) and biological factors (2). A historical trajectory was identified in the questions being asked and the theory and terminology utilised. A critical approach identified the influence of power within these transitions as there have been strong feelings on the part of many health workers that an individual focus has been a way of attributing blame and avoiding organisational responsibility (Chamorro-Premuzic and Lusk, 2017; Aubusson, K, 2017).  

A major variation was that of focus. The Input-processes output model was identified as a useful organising structure. Using this system focus was on: (1) inputs (job demands, resources etc); (2) processes that link inputs and outputs; (3) outputs (resilience burnout, wellbeing, quality and safety of care etc); and (4) environment (consideration of the environment that these processes occurred in).  

The terminology used described a historical trajectory from workplace stress to outcomes of stress: burnout, resilience and wellbeing (Eckleberry-Hunt et al., 2009). A very broad range of overlapping terminology was identified with distinctions between concepts often problematical (Southwick et al., 2014; Bianchi, Schonfeld and Laurent, 2015; Litz and Kerig, 2019). Burnout, resilience, workplace stress, moral distress, secondary trauma, compassion fatigue, and workplace health were all identified as relevant terms within the literature. Differences in the perspective take were partly accounted for by differences between a positive (resilience) or negative (burnout) perspective and partly by variations in scope from a focused outcome (absenteeism or moral distress) to holistic concept such as wellbeing.  

There has been a historical trajectory in the theoretical perspective taken. Initial work was based on application of behavioural models of stress to the workplace with a focus on the psychological processing of stress and the identification of important concepts such as self-efficacy, autonomy and baseline personality traits. Important theoretical constructs were Lazarus and Folkman’s transactional model which highlighted that the response to stress is a process which is mediated by a process of cognitive appraisal of the work environment (Lazarus and Folkman, 1987).    

Early work focused on precise definition of terms and tools for measurement of well-defined outputs of workplace stress such as the development of the Maslach burnout inventory (Schaufeli, Maslach and Marek, 2017), asking questions around prevalence in different communities or professional groups (IsHak et al., 2013; Adriaenssens, De Gucht and Maes, 2015; Chuang et al., 2016; Dimou, Eckelbarger and Riall, 2016). Important concepts were the construction of burnout as a response to stress with dimensions of exhaustion, disengagement and lack of accomplishment. The question was often -what is the prevalence in my “tribe”: physician burnout or burnout in nurses rather than a focus on teams in the same work environment. The introduction of work hour limits in Europe and the USA led to an evaluation of the impact of changing this measure of job demand (Martini, Arfken and Balon, 2006; Landrigan et al., 2008).  

The accumulating literature on prevalence across disciplines has led to the concept of a crisis and increasing calls for action highlighted by high profile and personal examples. The COVID-19 pandemic has highlighted the implications of the physical and emotional health of health workers in a dramatic way.  

A major focus, particularly of early models was identification on the relative importance and interactions between various inputs into wellbeing. This was reflected in the demonstration of the importance of imbalance models such as job demands and resources (Bakker and Demerouti, 2007), control (Demerouti et al., 2001) or rewards (Basińska and Wilczek-Rużyczka, 2013; Morgan, Dill and Kalleberg, 2013). The question was often: what are the important determinants of (stress, burnout, resilience) in this context? Study of the outputs of workplace stress have moved from a negative (burnout) to a positive perspective (resilience) to a more holistic perspective (wellbeing). These shifts have coincided with a parallel discourse around how we define health (Huber et al., 2011). This is reflected in a move from an absence of illness-based definition, to a more holistic assessment of human needs. Influential perspectives include Maslow’s self-actualisation theory (Maslow, 1970), and Sens capability approach (Sen, 1993).  

Subsequently the focus has moved to interventions, particularly psychological interventions conceived as “stress management “(Reynolds, Taylor and Shapiro, 1993; Bragard et al., 2010); from cognitive behavioural therapy (Gardner et al., 2005) to massage (Brennan and DeBate, 2006) and aroma therapy (Hansen, Hansen and Ringdal, 2006). More recently mindfulness-based stress reduction has been a dominant method (Goodman and Schorling, 2012; Luken and Sammons, 2016; Daya and Hearn, 2018), and there has been a more dynamic approach focused on recovery, and a more holistic consideration of environment to include home and (work-life balance).  

Consideration of workplace stress as a process highlighted the mediating role of specific behaviours between perception and outcomes. These behaviours included practices of clinical care, selfcare and learning. Learning was identified as the key to adoption of adaptive or mal-adaptive behaviours leading to positive or negative outcomes.  

The organisational perspective reflects a fundamental recognition about the inadequacy of an individual perspective and the importance of the environment. This also has undergone a historical trajectory that has moved from a focus definition and measurement of summative concepts such as organisational culture and climate towards a more differentiated approach to construct climate in multiple ways e.g. a safety climate or learning climate. The organisational perspective has also moved towards a more dynamic consideration of how adaptation occurs, with a greater focus on recovery, as well as a more holistic approach to the way that work interacts with home (work-life balance or integration).  

The key concepts and theoretical frameworks identified, definitions and associated references are outlined in Supplementary File 1.  

Phase 2: Identification of initial model assumptions  

These considerations lead to the adoption of the input process and outcomes model (IPO) as a basic framework to provide a trusted organising conceptual framework. The foundational model for stress-response has been the transactional stress model, so this was fitted to an IPO model. “Objective” inputs into the workplace were identified as stressors/threats, outputs of the system divided into wellbeing and performance (individual or system). The mediating processes are: (1) cognitive appraisal in identifying the perceived stressors; and (2) the identification of specific behaviours (clinical practices, self-care behaviours, etc). The adaptation to workplace stress occurs within a specific work climate or organisational environment. The IPO model, Lazarus workplace transactional stress model and organisational climate literatures, were therefore identified as key elements for the initial model. Characteristics of the initial model are given in Figure 4 and published online at Mindmeister. 

Figure 4: The initial mapping of concepts from phase 1 onto a mind map. Concepts identified in the literature review are mapped onto this framework.  

Phase 3 Iterative improvement  

Improving comprehensiveness  

The model was revised by targeted exploration of theory around the concepts outlined, expanding searches, where necessary beyond systematic reviews, beyond health and by recruiting new theoretical perspectives. McGrath generalised the role of perception in stress to include perceived demand, perceived ability to cope, perceived impact (McGrath et al., 1970). Inputs of stress have been summarised by Maslach’s six areas of  worklife (Leiter and Maslach, 1999) and the concept of wellbeing elaborated by Warr’s vitamin theory (Warr, 2008). Concepts of recovery were strengthened by the stressor-detachment model (Sonnentag and Fritz, 2015). The issue of environment was broadened to differentiate between different aspects of environment such as the climate of psychological safety (Dollard and Bakker, 2010) and learning (Lases et al., 2019) which impact in different ways. The coherence of these concepts with the initial model increased its trustworthiness.  

Reducing complexity  

Subsequent improvement of the model focused on reducing complexity and synthesis of overlapping concepts within the existing model. Much of the complexity to the model comes from integrating organisational and individual perspectives so the model was simplified from a comprehensive model (simultaneously dealing with individual and organisation levels) to a multi-level model (able to be applied to each level). Identification of learning as a key process led to consideration of learning at either as individual learning (Sherwood and Zomorodi, 2014). Inputs were consolidated using the six areas of worklife (Leiter and Maslach, 1999) and the literature on mindsets; (Sherwood and Zomorodi, 2014, Yeager and Dweck, 2012; Ben-Avi, Toker and Heller, 2018) was introduced as both an input to the system and a potential learning focus. The concept of a mindset or schema allowed collation of multiple beliefs about stress, clinical practice and improvement in ways that impact on the response to stress. Thus, a growth mindset, where the ability to cope with stress is not considered a fixed trait, promotes resilience (Yeager and Dweck, 2012). Promoting and teaching such a mindset becomes a critical strategic priority.  

Outputs were consolidated into those related to wellbeing and performance with a continuum of outcomes between negative (burnout, clinical error) and positive (resilience wellbeing and high-quality care).  

Identification of additional perspectives  

Two areas were identified as under-theorised in the initial model. The first was the relationship of the proposed model to the Job demand-resources model which was identified as a highly trusted model in the existing literature. The second area was that of learning which had a role in many of the processes identified and potential interventions. These areas were selected for further iterations of inquiry.  

Re-appraisal against existing models (JD-R) and focus on learning  

The major difference from the simple presentation of the JD-R model was the highlighting of the difference between appraised job demands/resources and the environmental job demands. Consideration of learning in practice to be a critical mediator of wellbeing raises the question as to what drives this cycle? We drew on the JD-R models conceptualisation of the key role of energy and motivation to highlight these as drivers. The major difference between this model and other models, was the focus on learning in the workplace. This raised the question as to what is learnt. We therefore consolidated relevant learning domains as increasing circles of focus from the inner self to the iteration with others. Thus, drawing on professional identity, the practitioner learns to be a doctor or a nurse, reflecting the critical existential domains of burnout and wellbeing. Drawing on the literature on cognitive appraisal and mindsets there are important beliefs which are both inputs into the system and learnt outcomes. Beliefs such as self-efficacy, a growth mindset – where obstacles are opportunities for learning and beliefs that they system is supportive of you – or not. A further internal aspect of wellbeing related to the affective domain of learning and the learning of skills in emotional self-regulation and self-care.  

A critical domain of learning is what practitioners do, of practice. This includes learning to communicate better, learning to insert an airway efficiently while minimising risk of self-contamination, learning to set aside time for reflection and study. The team-based nature of contemporary health practice does require moving beyond the individual perspective and here the practitioner needs to learn how to interact. Skills in teamwork, dealing with interpersonal conflict, asking for help etc, are important skills for avoiding burnout, thriving and delivering high quality care. The final level of learning moves from a static perspective to a dynamic one and that is learning to adapt. Learning to learn, learning how to respond to adverse events, planning self-care when strategies are not working are necessary in a dynamic model. The domains of what is learnt are represented in Figure 5.  

Figure 5: Domains of learning relevant to stress and wellbeing. The domains of learning are represented as moving in increasing circles from the personal to the social. 

The final model therefore makes explicit that the wellbeing can be considered at different levels as the wellbeing of the individual, the organisation or the society; and that each level goes through a process of appraisal of demands and resources, which mediates stress. The main process is an iterative process of learning and adaptation of behaviours. The drivers of this process are motivation and energy. These produce outputs of wellbeing and performance. Each locus of concern interacts with its own environment which moves from the micro to macro approach. The environment however can be deconstructed into three main domains- the “objective” environment of physical resources and demands, the appraisal environment which influences how an individual or organisation views the situation and a learning environment which influence show behaviours can adapt.  

The issue of what learns and what adaptation occurs can be viewed similarly as circles moving from critical identity questions of “being”, beliefs, emotional response and practice to questions of interaction and adaptation. The final model is represented in Figure 6.  

Figure 6: A model of learning to be well in the workplace in response to stress. 

Paracelsus is credited with the statement that all models are wrong, but some models are useful. The strength of this model is that it integrates multiple discourses about health, wellbeing, burnout, occupational health and education, towards a model that is potentially useful in guiding the scope of interventions. Organisations need a conceptual model and a methodology for improving the wellbeing at both individual and organisational levels. The wellbeing of the health workforce, as well as the quality, safety and sustainability of the health system, demand that we get this right.