Understanding the impact of exposure to adverse socioeconomic conditions on chronic stress from a complexity science perspective

Chronic stress is defined as the repeated occurrence of the stress response, which involves cognitive, affective and biological reactions induced by a stressor [1], over an extended period of time, i.e. several hours per day during weeks to months [2]. Chronic stress increases the risk of chronic diseases such as type 2 diabetes [3] and cardiovascular disease [4]. Exposure to adverse socioeconomic conditions has been shown to be associated with chronic stress [5, 6]. Chronic stress is therefore likely to be one of the explanatory factors for the higher occurrence of adverse health outcomes among people with low socioeconomic status, who are more frequently exposed to such conditions, independent of health-related behaviour [7]. In epidemiological research, socioeconomic status is usually defined based on education, income and/or occupation [8].

One way to explain why exposure to adverse socioeconomic conditions is associated with a higher prevalence of chronic stress is to look at the conditions that are necessary for a stress response to arise: high demands and limited access to resources to cope with these demands. The stress response can be conceptualised as being determined by the balance between the demands of a stressor and the resources available to cope with these demands [1, 9]. Specifically, a stress response is more likely to occur when a stressor is assessed as being uncontrollable, i.e. when it poses demands that exceed available resources [10‐13]. People exposed to adverse socioeconomic conditions are worse off on both sides: they are likely to endure high demands and to have limited access to resources. First, exposure to stressors that pose high demands, such as financial hardship, family conflict, death of a loved one, poor living conditions, crime, violence and (status) discrimination, is correlated with adverse socioeconomic conditions [6, 14]. Second, exposure to adverse socioeconomic conditions implies having limited access to resources, such as money, knowledge, power, prestige and beneficial social connections, to accommodate these demands [15]. In addition to the severity of these stressors as expressed in demands versus resources, exposure to adverse socioeconomic conditions is associated with recurrent stressor exposure [16], increasing the probability of suffering from chronic stress.

Epidemiological studies concerning the effect of stressors on health originally focused on finding a linear relationship between stressors—initially expressed as the number of “life events” ranked in severity according to “life change units”—and various outcome measures including distress (i.e. emotional suffering as a function of stressor exposure [17]) [18]. However, it was soon realised that other psychological and social factors are important in defining the impact of these stressors. In fact, across studies it was found that life events only explain anywhere between 1 and 12% of the variance in distress [18]. While availability of resources to accommodate a stressor’s demands are increasingly being taken into account to explain whether or not a stressor evokes a stress response, as described above, it can be argued that looking at the balance between demands and resources is still too simplistic to explain why those exposed to adverse socioeconomic conditions more often suffer from chronic stress. Various reasons can be distinguished that make the relationship between exposure to adverse socioeconomic conditions and chronic stress a complex issue.

First, there are feedback loops between stressor exposure and psychological processes, which results in complexity. This can be illustrated by the reciprocal interaction between stressor exposure and the appraisal of additional stimuli. Recurrent exposure to stressors can lead to an accumulation of stressors, finally resulting in a constant background of chronic stress [2]. This persisting pressure may lead to additional stimuli being assessed as stressors because they are more threatening given the circumstances of chronic stress [1, 11], which results in an amplifying feedback loop.

Second, such interactions are transpiring over different temporal scales. For instance, whether or not a stressor is appraised as controllable is indirectly influenced by past stressor exposure. Here, past experiences of low control can diminish perceived controllability of subsequent stressors [19]. In turn, the effect that past stressor exposure has on perceived controllability is affected by concurrent access to socioeconomic resources [20, 21].

Third, there is an interplay between biological, psychological and social processes, i.e. different spatial scales. For example, because cognitive bandwidth (which relates to the working memory) is a finite resource and each stressor takes up cognitive bandwidth, an accumulation of stressors [6, 14]) can prevent people from addressing some of the stressors [22‐25]. The impossibility of dealing with them all at once and the trade-off that therefore needs to be made can leave some of the stressors unresolved [26, 27].

The dynamics described above, stemming from feedback loops and the interconnectedness between temporal and spatial scales, could explain why the relationship between stressors and distress is not linear and why it is necessary to consider biological, psychological and social factors interacting with each other to put this relationship into context [18]. In this paper, we therefore examined the dynamics underlying the relationship between exposure to adverse socioeconomic conditions and chronic stress from the perspective of complexity science, which focuses on dynamics within a system, i.e. on how all parts interact within the whole as opposed to on how each part relates to the whole in isolation [28].

Correspondingly, the research question of this paper was: which dynamics drive chronic stress in a context of adverse socioeconomic conditions? We applied the complexity science perspective to our research question by developing a causal loop diagram (CLD) of the involved dynamics. From this perspective, we examined the impact of a context of adverse socioeconomic conditions on individual-level biological and psychological processes, leading to a biopsychosocial model of stress in line with Lazarus’ model of stress as a relational outcome between people and their environment [11]. Specifically, given that every biopsychosocial model describes an outcome as arising from a combination of the biological, psychological and socio-environmental scales [29], the added value of a complexity science perspective is that it provides methods to explicitly study the interactions between these spatial scales over the life course.

In a CLD, empirical research and theoretical knowledge (i.e. well-reasoned hypotheses) can be combined to compose a network of variables and causal links [30]. A CLD is thus a visual representation of “initial hypotheses, or sets of linked hypotheses, about the processes involved” [31]. It is a conceptual framework that can help us interpret the current evidence base.

A CLD is an appropriate methodology to apply the complexity science perspective to our research question for several reasons. First, dynamics can be portrayed in a way that captures important individual differences while retaining a structure that provides an understanding of variation at the population level [32]. Specifically, each causal link refers to a causal mechanism that may act as a driver of chronic stress in a context of adverse socioeconomic conditions. Depending on the exact combination, the dominant mechanisms leading to chronic stress may differ between people exposed to such conditions. Second, a CLD can reveal non-linearity between stressor exposure and chronic stress: it could for instance show where differences in initial conditions may induce large and unforeseeable differences in outcomes through amplifying feedback loops involving biological, psychological and social processes [28]. Third, a CLD allows for interactions between multiple spatial and temporal scales [32].

We conducted a literature review on the mechanisms underlying the relationship between exposure to adverse socioeconomic conditions and chronic stress from a complexity science perspective to develop the CLD, where we focused on the characteristic indicative of a complex system referred to as emergence [31]. Emergent properties “arise from interactions between parts of a system” and “are not seen in any one part of a complex system nor are they summations of individual parts” [31]. This was dictated by our research question addressing the dynamics that drive chronic stress, which may be classified as an emergent property, in a context of adverse socioeconomic conditions. We characterised the stress response—the repeated occurrence of which constitutes chronic stress [2]—as emerging from interactions within a biopsychosocial system [11].

In this paper, we aimed to identify the dynamics that drive chronic stress in a context of adverse socioeconomic conditions. We started the introduction with the conceptualisation of the stress response as being determined by the balance between the demands of a stressor and the resources available to accommodate these demands. By taking a complexity science perspective, we have identified multiple amplifying feedback loops which may be elicited following exposure to a context of adverse socioeconomic conditions, showing how these demands and resources are interlinked in a complex way. For example, stressor exposure affects psychological processes that are important in accommodating subsequent stressors, which can result in a non-linear relationship between stressors and chronic stress. In a context of adverse socioeconomic conditions, it seemed that the impact of stressors on chronic stress increases over time, dictated by the continuous accumulation of stressors and the elimination of options for problem approaching. The fact that a context of adverse socioeconomic conditions comes with a constant exposure to stressors over the life course appeared to be essential in the dynamics identified. The combination of recurrent exposure to stressors and undermined access to coping resources due to past exposure to uncontrollable stressors can inform future assessments of stressors as uncontrollable. Coupled with the experience that stressors could not be overcome in the past, this can lead to feelings of helplessness in relation to current stressors, thereby increasing chronic stress. As a consequence of this chronic stress, cognitive bandwidth is taxed, which can leave stressors unresolved, amplifying chronic stress. In addition, a background of chronic stress can increase the susceptibility to appraising additional stimuli as threats. Because these feedback loops are interlinked, these dynamics could be conceptualised as one vicious cycle. Given the higher probability of recurrent exposure to stressors, this vicious cycle is more likely to occur and to be unceasingly sustained over the life course for people exposed to adverse socioeconomic conditions.

The first reason why we took a complexity science approach was that it enabled us to do justice to the interactions between the involved biological, psychological and social factors that are needed to put the relationship between stressors and distress into context. This perspective can therefore lead to a new outlook on how these factors should be addressed in interventions. The apparently decisive role of these interactions implies that there may be room for improvement in the way we address the role of stress in interventions. The CLD shows that a context of exposure to adverse socioeconomic conditions can mean that problem approaching coping strategies have repeatedly been proven to be unsuccessful because stressors are insolvable, which can result in “the belief that stress factors cannot be overcome” [52] and affect the appraisal of subsequent stressors. This indicates that improving access to socioeconomic resources as a strategy to reduce chronic stress might need to be accompanied by attention to past experiences that may have substantiated feelings of helplessness.

Consequently, we should take care not to conceptualise chronic stress as an individual-level issue but as the emergent outcome of a set of conditions. However, individual-level interventions focusing on such psychological processes have increasingly become established practice in this context over the past years. The pitfall of this type of interventions is that they can draw attention away from the contextual circumstances that cause recurrent exposure to stressors and shift responsibility to the individual. An example of an intervention concentrating on psychological processes is Mobility Mentoring™, which aims to “teach clients how to ‘mentor’ themselves, so that they can better analyse their own problems, regulate their own behaviours, independently set long-term and short-term goals, build mastery and self-control, and transfer those skills to others by mentoring their own children, communities, and networks” [70]. Translated to the CLD, the goal of this intervention is to invoke behaviour change in people regarding coping resources and coping strategies. The CLD illustrates that if the social context remains the same, individual-level psychological processes might not be the most productive place to start to mitigate chronic stress. This is because, as demonstrated by the CLD, psychological processes are affected by exposure to stressors that pose high demands. Despite its focus on psychological processes, Mobility Mentoring™ is advocated as a means to ultimately move out of poverty [71]. In other words, the assumption is that devoting effort to psychological processes should eventually go hand in hand with mobilising socioeconomic resources. The CLD serves to emphasise that even if interventions such as the one described above can realise shifts in psychological processes, these improvements can only be maintained if the accumulation of stressors and their repeated insolvability are also alleviated. The amplifying feedback loops relating to learned helplessness and to deterioration of access to coping resources because of repeated insolvability of stressors both illustrate the necessity of simultaneously targeting contextual circumstances and psychological processes.

Analogously, in the context of the obesity epidemic, public health researchers now agree that obesity is more than an individual-level energy imbalance [72]. It is the final embodied outcome of a constellation of factors ranging from epigenetics and social norms to the built environment and market forces [73]. It seems appropriate to create a similar narrative for chronic stress. Specifically, in the context of obesity it is now realised that “a siloed focus on individual responsibility leads to a failure to address these wider factors for which government policy can and should take a leading role” [73]. Looking at chronic stress as purely a mental health concern could mean erroneously individualising a social issue. Paying undue attention to individual-level psychological processes independent of access to socioeconomic resources could steer in the direction of blaming the victim.

The second reason why we took a complexity science perspective was for its potential to explain non-linearity as arising from amplifying feedback loops. Amplifying feedback loops in a system can cause small differences in initial conditions to be amplified over time, resulting in large differences in the emergent properties of the system. The CLD shows how exposure to uncontrollable stressors can be understood as an initial condition that may steer the system towards a certain outcome, highlighting this condition as a promising leverage point for changing system-level outcomes. In complexity science, leverage points refer to “places in the system where a small change could lead to a large shift in [system-level] behavior” [54]. As differences in initial conditions may start out small but gradually push people towards entirely different trajectories, the timing of interventions within the life course appears crucial. Even though we defined “uncontrollable childhood and life course stressors” as a non-feedback variable in the CLD, in reality this initial condition is not independent of context and will be driven by social structures. Accordingly, the system that we described is evolving within a larger system, which should be considered in order to affect this initial condition, meaning that exposure to adverse socioeconomic conditions and limited access to socioeconomic resources are important leverage points. This “larger system” might also include temporal scales other than those included in the causal loop diagram. For example, changes in psychological processes at the individual level might, if accompanied by improved access to socioeconomic resources, eventually affect the initial conditions for the next generation.

By describing variables as part of feedback loops, we also take a step towards identifying whether intervening on a particular variable can have a balancing or reinforcing effect within the entire system, driving it towards a different trajectory. For example, if the necessary precondition that access to socioeconomic resources is improved is met, the amplifying feedback loop that underlies either the deterioration or the development of coping resources can serve as a leverage point within the system as it can act either as a vicious or a virtuous cycle. While we should not expect that access to coping resources is enough to break conditions of chronic stress, an intervention that supports problem approaching coping strategies to meet their objective may contribute to the restoration of access to coping resources. When coping resources such as mastery and self-esteem are built up in a specific state, this may translate to a global trait belief, which can then enable problem approaching coping strategies. Importantly, two of the five feedback loops that were distinguished are related to access to coping resources. If a particular variable plays a key role in several feedback loops, unveiling overlapping burdens, this may suggest that this variable is a potential leverage point within the system. Psychological interventions that aim to enhance resilience, which has shown to be affected by factors such as self-esteem, active coping, self-efficacy and optimism [74], may act upon this leverage point. In combination with improving access to socioeconomic resources, such an individual-level approach might then contribute to diminishing the dominance of certain mechanisms within the system.

In this paper, we described the mechanisms underlying the relationship between exposure to adverse socioeconomic conditions and chronic stress from a complexity science perspective, focusing on amplifying feedback loops across different scales. By taking this perspective, we aimed to contextualise the existing portrayal of these mechanisms in the literature by showing that they are interlinked and embedded in social structures. For example, coping resources are defined in the literature as “relatively stable individual differences” [34], where especially neuroticism is often referred to as a character trait resulting from inter-individual variation in “negative emotional response to threat, frustration, or loss” and typified by “irritability, anger, sadness, anxiety, worry, hostility, self-consciousness, and vulnerability” [47]. By demonstrating that those factors that we commonly refer to as “coping resources” are affected via mechanisms such as the repeated insolvability of stressors, we intended to contribute to seeing these differences in a new light. Similarly, as previously addressed by others [69], the negative connotations that accompany problem avoidance coping strategies in the literature, suggesting that they are maladaptive in all circumstances, should be reassessed. By describing the mechanisms that could evoke problem avoidance in a context of adverse socioeconomic conditions, we aimed to shift the focus away from the individual to the context. This is why, despite the connotations of victim blaming that the terms “coping resources” and “problem avoidance” may convey due to how they are currently used in the literature, we have used this terminology in the CLD.

Recognising that chronic stress is one of the interlocking pieces that can explain the effect of exposure to adverse socioeconomic conditions on mental health may be important for clinical practice, for example when treating patients with depressive symptoms in relation to chronic stress. The value of the CLD in this setting seems to lie in particular in that it contextualises individual-level psychological processes. It shows that these psychological processes are part of several interacting mechanisms that, in combination, result in chronic stress and are driven by a context of adverse socioeconomic conditions.

The CLD can also be expanded to include countervailing mechanisms, i.e. mechanisms that prevent people from developing chronic stress despite being exposed to adverse socioeconomic conditions. These countervailing mechanisms may in some cases serve to compensate for the feedback loop mechanisms leading to chronic stress, where through the principles of non-linearity a certain feedback loop may become dominant or remain inactive through small changes in the system. For example, if social support would be added to a person’s constellation of factors, this may result in the effect of the other causal mechanisms being counterbalanced for that specific person [75]. Extending the CLD with such mechanisms, as well as the interactions between socioeconomic resources (such as between money and knowledge [15]), would be a next step in obtaining a full understanding of why people exposed to adverse socioeconomic conditions more often suffer from chronic stress.

The CLD is subject to some methodological considerations. A CLD that was established on the basis of GMB should always be considered as a reflection of the collective understanding of a specific group of domain experts, in relation to a specific research question. In other words, a single, “true” CLD modelling “the dynamics driving chronic stress in a context of adverse socioeconomic conditions” cannot exist. More generally, it cannot be reiterated often enough that “all models are wrong” [76], as they represent an abstraction of reality meant as a useful medium that aids in understanding that reality. In line with this, the CLD should be seen as a summary of the current evidence base that is limited to the specific research question addressed in this paper. Nevertheless, to increase the structural validity, we substantiated the CLD by literature references and evaluated its structure by presenting it to eight independent domain experts who were not involved in the GMB process. Additional variables and causal links could be integrated into the CLD to further explicate the relationship between exposure to adverse socioeconomic conditions and chronic stress in more detail, using updated knowledge, under different contexts, for different research questions or reflecting the viewpoint of a different set of researchers. This highlights that a developed CLD may well be in continuous flux. Notwithstanding, we argue that at a qualitative level the present CLD contains the most prominent mechanisms that drive chronic stress in a context of adverse socioeconomic conditions which can at least serve to anticipate important non-linear behaviours.

We defined the boundaries and which variables were regarded as exogenous and as endogenous in relation to our research question, where different boundaries and a different role for particular variables may be called for when applying this CLD to another context. For example, we have taken access to socioeconomic resources as an initial condition that cannot be influenced from within the modelled system, whereas in fact this access is determined by social structures. For the purposes of the CLD, the distribution of socioeconomic resources is thus taken as a given, while this could actually be addressed in policy, as we have discussed above. On the other hand, while we recognise that the modelled system is evolving within a larger system, we do expect the mechanisms between the endogenous variables included in the CLD to be generalisable. For instance, even if a mechanism influencing socioeconomic resources would be included, we presume that the role of socioeconomic resources within the modelled causal links remains the same.

We have analysed and discussed the CLD’s feedback loops and hypothesised dynamics in a qualitative manner. In order get a more precise indication of the importance of the various mechanisms within the system ideally a computational model, such as a quantitative system dynamics model, would have to be developed based on the conceptual model (as we have previously realised in [77]). Related to this, even though we have made a first step by identifying the temporal scales that the variables and causal links act upon, the exact implications that the different durations of the modelled mechanisms eventually have on the system’s outcomes remain elusive until quantification is achievable. Nonetheless, we are convinced that a conceptual model does justice to the objectives of this paper.

To conclude, we developed a CLD to apply a complexity science perspective to understanding the impact of exposure to adverse socioeconomic conditions on chronic stress, which might serve as a basis for developing novel hypotheses for future research on this topic. Taking a complexity science perspective clarifies that the social structures causing some people to be exposed to adverse socioeconomic conditions may sustain amplifying feedback loops with individual-level biological and psychological consequences. As such, we should take care not to conceptualise chronic stress as an individual-level issue but as the emergent outcome of a set of conditions. The CLD describes that exposure to adverse socioeconomic conditions implies recurrent stressor exposure which impacts chronic stress via amplifying feedback loops that together could be conceptualised as one vicious cycle. This means that in order for individual-level psychological interventions to be effective, the context of exposure to adverse socioeconomic conditions also needs to be addressed.