Social factors and psychotic illness

There are a variety of explanatory models for the impact of social factors on psychotic illness. Social factors may change the exposure to or action of biological risk factors or they may have a direct toxic effect on individuals and their families. Genetic and epigenetic processes interacting with social factors at the individual, group and societal levels over a life course produce a complex and interconnected web of causation. ^{Reference Kirkbride and Jones3,Reference March and Susser4}

Understanding how different social factors change the risk of psychosis has been problematic, in part because of the numerous possible mechanisms and multiple levels through which this could occur. ^{Reference Kirkbride and Jones3–Reference Susser, St Clair and He5} For instance, there is a reported association between prenatal malnutrition and an increased risk of schizophrenia in adolescent life. This could reflect the impact of nutritional stress on the hypothalamic–pituitary–adrenal (HPA) axis or direct effects of the low levels of nutrients on deoxyribonucleic acid stability or expression (to name just two theories). However, at a different level, urban or rural residence during famine may change the risk of exposure by altering access to food to supplement the diet. ^{Reference Susser, St Clair and He5}

The complex nature of this enquiry is demonstrated here by consideration of one of the explanatory models for the impact of social factors on psychotic illness: the ‘vulnerability–stress model’. ^{Reference Nuechterlein and Dawson6,Reference Myin-Germeys, Van Os, Morgan, McKenzie and Fearon7} This model posits that genetic or developmental vulnerabilities interact with social adversity to culminate in a common pathway leading to stress. ^{Reference Miller, Lawrie, Hodges, Clafferty, Cosway and Johnstone8,Reference Mason, Startup, Halpin, Schall, Conrad and Carr9}

Genetic, biochemical and neurological evidence supports the link between stress and psychosis. ^{Reference Walker and Diforio10–Reference Jones and Fernyhough12} The physiological effects of stress are demonstrated by findings of activation of the HPA axis to produce excess cortisol and adrenocorticotropic hormone. ^{Reference Ryan, Sharifi, Condren and Thakore13} In addition to genetic factors, acquired vulnerabilities such as prenatal exposures, perinatal complications, drug misuse and childhood abuse or neglect may regulate the response to stress. Links between the physiology of stress and that of psychosis have been reported. The hippocampus, known to play a key role in dampening HPA activity, is reduced in volume in psychotic disorders. ^{Reference Steen, Mull, McClure, Hamer and Lieberman14–Reference Fusar-Poli, Perez, Broome, Borgwardt, Placentino and Caverzasi17} Stress also leads to mesolimbic release of, dysregulated mesocortical response to and overall sensitisation to dopamine, a neurotransmitter that may mediate psychotic symptoms. ^{Reference Soares and Innis18–Reference Kapur20}

Experimental support for this is found, for example, in studies in which psychotic symptoms are associated with increased dopamine release after amphetamine challenge. ^{Reference Laruelle and Abi-Dargham21} Over time, in individuals who are at risk, exposure to moderate levels of stress can lead to an excessive and persistent dopamine response. Building on this theory, those with higher rates of life stressors may be at particular risk of developing a psychotic illness.

Epidemiological studies support this at multiple levels. Cumulative exposure to traumatic life events, or the number of life events experienced, is associated with an increased risk for psychosis ^{Reference Shevlin, Houston, Dorahy and Adamson22–Reference Cantor-Graae25} and the British National Psychiatric Morbidity Survey has reported that adverse life events are associated with subsequent psychotic experiences in the general population. ^{Reference Johns, Cannon, Singleton, Murray and Farrell26,Reference Wiles, Zammit, Bebbington, Singleton, Meltzer and Lewis27} In addition to traumatic major life events, the accumulation of minor events or ‘daily hassles’ have also been linked to psychotic illness. ^{Reference Myin-Germeys and van Os28}

However, it is not just the accumulation of life events that is important. As stated previously, a life event may trigger a train of other events. For instance, loss of a parent as a child (which has been associated with increased risk of later psychosis) could increase the chance of living in poverty and thus exposure to other environmental risk factors associated with psychosis.

At the level of gene–environment interaction, minor life events and daily hassles interact with polymorphisms known to be involved in dopamine neurotransmission to cause differential stress reactivity and psychotic experiences. ^{Reference van Winkel, Henquet, Rosa, Papiol, Fananás and De Hert29} Of note, the same polymorphism has been shown to correlate with differential risk for psychosis in individuals who use cannabis, ^{Reference Caspi, Moffitt, Cannon, McClay, Murray and Harrington30} further highlighting the complex interplay between multiple biological and environmental risk factors. And genetic variants and environmental exposures can interact in ways that are protective and not just harmful. Instead of focusing on ‘vulnerability genes’ that confer increased risk in the presence of certain environments, Belsky et al call for an appreciation of ‘plasticity genes’ that confer a nuanced differential susceptibility: increased risk in some environments, and decreased risk in others. ^{Reference Belsky, Jonassaint, Pluess, Stanton, Brummett and Williams31}

Physiologically and cognitively, exposure to early developmental stressors (such as childhood trauma) may act by sensitising people to later adverse events, major or minor. They may increase the likelihood of adverse events for instance by shaping the capacity of individuals to form relationships. ^{Reference Myin-Germeys, Van Os, Morgan, McKenzie and Fearon7,Reference Wright, Turkington, Kingdon and Ramirez Basco32} Alternately, they may interface with a person’s attributional style – potentially making them more prone to psychotic thinking. ^{Reference Bentall, Kinderman and Kaney33}

Moving from the individual to the ecological, a major branch of investigation into the social aetiology of psychotic illness began in 1939 with Faris & Dunham’s work on the urban environment. ^{Reference Faris and Dunham34} Since then, research has focused on better defining its role as a risk factor for psychosis. Studies have demonstrated that the higher risk of psychotic illness for those living in cities is not the result of social drift but is associated with being born and growing up in an urban environment in childhood. Moreover, the bigger the city the bigger the risk. ^{Reference Spauwen and Van Os35–Reference Boydell, McKenzie, Morgan, McKenzie and Fearon38}

Living in an urban environment must, however, be a marker for a number of exposures. Allardyce et al found that after adjusting for social fragmentation and deprivation, there was only a non-significant trend towards an association between urbanicity and schizophrenia. ^{Reference Allardyce, Gilmore, Atkinson, Rapson, Bishop and McCreadie36}

Other factors may be important for the association between place and psychosis. For instance, Silver et al documented that the proportion of people moving in and out of an area was associated with higher rates of schizophrenia. ^{Reference Silver, Mulvey and Swanson39} Another pilot study has reported that increased social cohesion and social efficacy in areas in London, UK were associated with a reduced incidence of psychosis, ^{Reference Boydell, van Os, McKenzie, Allardyce, Goel and McCreadie40} and after controlling for individual deprivation, Boydell et al have shown an increased rate of schizophrenia with increasing neighbourhood inequality, but only in more deprived areas. ^{Reference Boydell, van Os, McKenzie and Murray41}

Kirkbride et al recently used multilevel Poisson regression simultaneously to model individual- and neighbourhood-level factors. ^{Reference Kirkbride, Morgan, Fearon, Dazzan, Murray and Jones42} They reported that 23% of the incidence of schizophrenia could be attributed to neighbourhood-level social risk factors including socioeconomic deprivation, voter turnout (proxy for social capital), ethnic fragmentation (proxy for segregation) and ethnic density (95% CI 9.9–42.2). A 1% increase in voter turnout (incidence rate ratio (IRR) = 0.95, 95% CI 0.92–0.99) and ethnic segregation (IRR = 0.95, 95% CI 0.92–0.99) were both independently associated with a reduced incidence of 5%, independent of age, gender, ethnicity, deprivation and population density. ^{Reference Kirkbride, Morgan, Fearon, Dazzan, Murray and Jones42}

However, there are a number of different mechanisms through which ecological factors could have their impact on psychosis. Stress may be just one of a number of mechanisms. ^{Reference Susser, St Clair and He5}

Taken as a whole, the literature demonstrates that social factors are associated with psychosis risk at an individual and an ecological level. There is evidence of an increased risk with greater exposure to some individual risk factors and with a greater number of risk factors. There are multiple mechanisms and an interplay between genetic, epigenetic and environmental factors.

Conceptual framework

A framework for categorising or organising different associations between psychotic illnesses and risk factors may be helpful to those who are considering developing prevention/promotion initiatives. It may also further research in the field. Although we have highlighted the stress hypothesis, this umbrella conceptual framework we present would allow the inclusion of a variety of perspectives on causation and different mechanisms. A simplified conceptual model would follow what we know from the literature and posit that there are four dimensions of interest:

1. 1 exposure to individual-level social factors linked to psychosis;

2. 2 exposure to ecological-level social factors linked to psychosis;

3. 3 the interaction between individual and ecological social factors; and

4. 4 time.

Developing interventions

This simplified model may be useful in considering possible prevention and promotion initiatives, since each dimension can be linked to types of strategies. Prevention at an individual level will aim to decrease the amount of exposure or limit the number of different social factors to which a person is exposed. Promotion efforts would enhance behaviours and expose individuals to social factors known to strengthen mental health. At an ecological level, the aim would be to develop healthy environments and limit exposure to psychologically noxious social environments. Where interactions between individual-ecological factors lead to an amplification or spiralling of risk, targeting such interactions may be important. If neither individual risk factors nor environmental risk factors can be reduced, then remedial strategies aimed at uncoupling the links between factors could be effective. And, of course, time is all-important. Understanding the stages of development at which particular risk factors act is important for targeting prevention strategies for maximal effect. Understanding the longer-term impacts of particular risk factors may give an indication of how long preventive strategies should last. Finally, understanding the timescales over which social factors may have their effects helps evaluators to pick the right time frame for assessment of the effectiveness of interventions.

Conclusion

We have outlined here a framework for considering how a range of social factors, acting over time, could individually or together make a contribution to the development of psychotic illness. The importance of such a model for conceptualising the social aetiology of psychosis lies in demonstrating the complex and multifaceted interplay between individual and ecological dimensions, their interactions, and time, while taking into account gaps in knowledge or the current evidence base. We hope that this model will help in research; in the design and planning of interventions aimed at particular risk and protective factors; in the training of health professionals; and in the development of health and social systems that more effectively support populations at risk of developing psychosis.