Introduction
The comprehensive review in this issue by Gayer-Anderson & Morgan (Reference Gayer-Anderson and Morgan2013) establishes, beyond reasonable doubt, that the social network and support deficits characteristic of psychosis are already apparent in the very early stages of the disorder. Even at first episode, the networks of people with psychosis comprise few members, with whom, moreover, they may have little contact. Their networks tend to contain proportionately more family members. They have fewer confidants, and more people are described as acquaintances. They are less likely to be satisfied with the level of support they receive, despite receiving more than they give (i.e. despite non-reciprocity). Networks may show evidence of shrinkage before the onset of illness. From this, it appears that psychosis is characterized by a reduction of social opportunities and a parallel narrowing of focus, most consistent with it being a consequence of the illness itself. This is supported by studies showing the association of small network size with longer duration of untreated psychosis and poor premorbid functioning.
However, small networks presage the level of positive and negative symptoms in people with psychosis up to 3 years after initial assessment, independently of a range of other potential predictors. Gayer-Anderson & Morgan (Reference Gayer-Anderson and Morgan2013) also draw attention to follow-up studies of normal populations, in which small networks predict the emergence of quasi-psychotic experiences. Although relatively small social networks are typical of people with psychosis, the size and other characteristics of the networks do vary greatly from person to person (Angell & Test, Reference Angell and Test2002). It is these inter-individual differences that appear to relate to illness course and outcome.
These results suggest reciprocal relationships between reduced social networks, increasing symptoms of psychosis, and poorer outcomes. This has the potential to form an unpleasant vicious circle, and implies at least some aetiological role for social relationship deficits. In this commentary, we argue that the postulated aetiological role becomes more plausible if we can identify effective mechanisms. This in turn implies that effective treatments could be directed at any point in the circle. In practice, changing the characteristics of social networks (e.g. by befriending interventions or facilitating social support) has proved difficult (Sensky et al. Reference Sensky, Turkington, Kingdon, Scott, Scott, Siddle, O'Carroll and Barnes2000). One reason for this may be that some people with psychosis may not be motivated to maintain friendships. Thus, interventions that focus on increasing the numbers of friends available may not be targeting the right problem: changing perceptions of inadequate support may be required.
Social networks
Social relationships are clearly important for general health and well-being (Thoits, Reference Thoits2011). Those with good social relationships and social support are physically and mentally healthier and live longer (Holt-Lunstad et al. Reference Holt-Lunstad, Birmingham and Jones2008). Over the past four decades, researchers have tried to define, describe and isolate the quantitative and qualitative characteristics of social networks and social support responsible for these protective effects. Some of these effects will be generic, but others may be specific to particular conditions such as psychosis.
The various definitions of social networks in the literature represent relatively minor variations on a theme. Thus, social networks are only a subset of the totality of social contacts, a subset defined by persistence over time and a degree of significance. This is because they are held to play an important role in the maintenance of the psychological and physical integrity of the person. Primary relationships (the main source of social support) have been distinguished from less personal secondary relationships.
On the basis of these definitions, social networks can be specified further in structural and functional terms. The former denote the existence and pattern of interconnections of network members rather than the content or quality of relationships. Functional aspects refer to what is provided by, or perceived to be available from, social relationships. Social networks are also described in terms of quantitative and qualitative aspects. Quantitative aspects include network size and density, kin v. non-kin composition, marital status and whether living alone. Qualitative aspects connote the degree of satisfaction with social relationships. These include reciprocity (the extent to which the relationship is characterized by giving as well as receiving), accessibility (the ease with which network members can be contacted), multiplexity (the number of separate functions provided by relationships), social isolation (pervasive lack of social contact or communication), the presence or absence of a confidant, and loneliness.
Social support
Social networks are important because they are the vehicle for delivering social support. However, while they are necessary conditions for social support, they are not sufficient. Studies of social networks in relation to psychiatric disorder are essentially using them as a proxy for (often unspecified) elements of support. However, support does not map perfectly onto the attributes of social networks. For example, someone with a large network may actually feel lonely, and perceived social support is not necessarily greater in more extensive social networks.
Social support is a functional attribute of social relationships: frequently mentioned functions are emotional, informational and instrumental (Alloway & Bebbington, Reference Alloway and Bebbington1987; Thoits, Reference Thoits2011). A number of definitions emphasize the cognitive elements.
However, social support covers several sub-constructs. The most commonly distinguished are received and perceived support. Received social support refers to the actual provision of supportive behaviours by others; while perceived social support concerns the recipient's perception of, and satisfaction with, the general availability of support. The main difference is that perceived social support refers to anticipating help in times of need; whereas received social support involves the recall of support provided over a given period. The distinction is important because research consistently shows stronger links between health and perceived social support.
It is important to note that not all social relationships are supportive or perceived as supportive; they can be a source of stress and strain, and some are abusive. Social withdrawal may consequently be protective, by insulating individuals from stressful relationships (Wing, Reference Wing1978). However, even when someone withdraws in this way, they may still feel the effects of lack of support in the form of distress and loneliness (Duberstein et al. Reference Duberstein, Conwell, Conner, Eberly, Evinger and Caine2004).
Loneliness in psychosis
Loneliness is a distressing negative experience resulting from a discrepancy between the desired and the perceived state. It has both a cognitive element (the perception of relationship inadequacy) and an affective character (an unpleasant experience). Although loneliness is influenced by quantitative and objective aspects of social networks (e.g. size of social network and frequency of contact), it is determined more by subjective appraisals, such as satisfaction and perceived adequacy. Loneliness does not equate with objective social isolation: the number of their friends is not a good predictor of how lonely someone feels.
The wide-ranging effects of loneliness are apparent. It is associated with reduced life satisfaction, psychosocial problems (low self-esteem and reduced social competence), and mental health difficulties including anxiety and schizophrenia (Neeleman & Power, Reference Neeleman and Power1994; DeNiro, Reference DeNiro1995). It is accompanied by a variety of negative affective states such as shyness, boredom or feelings of alienation. In people with schizophrenia, DeNiro (Reference DeNiro1995) found an increase in alienation, social isolation and loneliness over the patient's lifetime, and, conversely, a decline in positive connections with others. Neeleman & Power (Reference Neeleman and Power1994) found patients with psychosis had the smallest social networks, and felt lonelier than all other patient groups, independently of network size.
Confidants for people with psychosis
A confidant is someone who provides a relationship characterized by emotional intensity, reciprocity and availability. Its absence is strongly linked to the experience of loneliness (Green et al. Reference Green, Richardson, Lago and Schatten-Jones2001). Even at first episode patients have significantly fewer confidants than controls: Morgan et al. (Reference Morgan, Kirkbride, Hutchinson, Craig, Morgan, Dazzan, Boydell, Doody, Jones, Murray, Leff and Fearon2008) recently showed that first-episode psychosis patients were seven times less likely to have a confidant.
Mechanisms of change in social networks and social support
Although the effect of poor social networks and poor social support on physical and mental health is well established, our understanding of the mechanisms by which they operate is limited. We have chosen to argue here for the (neglected) role of psychological mechanisms (perceptions and their accompanying negative cognitions and affect), on the grounds that physiological responses are likely to be influenced by the psychological consequences of suboptimal social relationships (Thoits, Reference Thoits2011), and because the effects of specific psychological processes have been progressively substantiated in recent cognitive models of psychosis (Garety et al. Reference Garety, Bebbington, Fowler, Freeman and Kuipers2007; Fowler et al. Reference Fowler, Hodgekins, Garety, Freeman, Kuipers, Dunn, Smith and Bebbington2012).
Psychosis and the family
The narrowing of social networks in the preamble to psychosis often means that social support becomes restricted to family members. The ways in which family relationships affect the course of psychotic symptoms have been extensively investigated. High levels of expressed emotion in key relatives predict patient outcomes in schizophrenia, and family intervention can reduces relapse rates (Bebbington & Kuipers, Reference Bebbington and Kuipers1994; Pharoah et al. Reference Pharoah, Mari, Rathbone and Wong2006). Patients are certainly able to perceive carer criticism accurately, and such perceptions are linked to poorer functioning (Onwumere et al. Reference Onwumere, Kuipers, Bebbington, Dunn, Freeman, Fowler and Garety2009). Criticism is the direct transmission to the patient of negative judgments, and is associated with negative affect and poor self-esteem (Barrowclough et al. Reference Barrowclough, Tarrier, Humphreys, Ward, Gregg and Andrews2003).
In contrast, carer warmth protects against relapse (Bertrando et al. Reference Bertrando, Beltz, Bressi, Clerici, Farma, Invernizzi and Cazzullo1992). Such supportive relationships are likely to operate by reducing negative affect in patients and in some trials of psychological therapy, ‘befriending’ or supportive therapy resulted in these kinds of improvements, although not sustained (Sensky et al. Reference Sensky, Turkington, Kingdon, Scott, Scott, Siddle, O'Carroll and Barnes2000). It may also explain the better response to psychological treatments of patients in contact with carers (Garety et al. Reference Garety, Fowler, Freeman, Bebbington, Dunn and Kuipers2008).
Relational regulation theory
Building on attachment research, Relational Regulation theory (RRT; Lakey & Orehek, Reference Lakey and Orehek2011) contains explicitly cognitive and affective elements. Specifically RRT asserts that ‘main effects of social support on mental health occur when people regulate their affect, thought and action through ordinary yet affectively consequential conversations and shared activities, rather than through conversations about how to cope with stress’ (p. 482). According to RRT ‘perceived support typically does not directly cause affect but emerges from the types of social interaction that successfully regulate affect’ (p. 490).
Social support and cognitive models of psychosis: common ground
Impairments in social networks and support are associated with increased rates of dysphoric mood, specifically anxiety and depression. They are also associated with effects on a range of cognitions. Cognitive models of psychosis (e.g. Garety et al. Reference Garety, Bebbington, Fowler, Freeman and Kuipers2007) propose that the emergence and maintenance of psychotic disorder involves both cognitive and emotional processes. The models postulate that environmental factors operate through psychological mechanisms including an increased vulnerability to anxiety, depression and low self-esteem, and enduring cognitive biases affecting the processing of negative events and experiences. The latter emerge from the consequences of adversity on expectations, and includes schematic beliefs about self and others, and distortions in attributional style, such as a tendency to attribute power and control to others (Birchwood et al. Reference Birchwood, Gilbert, Gilbert, Trower, Meaden, Hay, Murray and Miles2004; Garety et al. Reference Garety, Bebbington, Fowler, Freeman and Kuipers2007).
Sündermann et al. (submitted for publication) recently suggested that the relationship between loneliness and paranoia may be mediated by affective processes such as anxiety or depression. This is because loneliness may distort thinking processes by exaggerating threat appraisals, or may make it harder for lonely patients to think of alternatives to their unusual ideas because they have no one to discuss them with. This may raise anxiety levels, leading in turn to increased paranoia.
You have
Access
Introduction
The comprehensive review in this issue by Gayer-Anderson & Morgan (Reference Gayer-Anderson and Morgan2013) establishes, beyond reasonable doubt, that the social network and support deficits characteristic of psychosis are already apparent in the very early stages of the disorder. Even at first episode, the networks of people with psychosis comprise few members, with whom, moreover, they may have little contact. Their networks tend to contain proportionately more family members. They have fewer confidants, and more people are described as acquaintances. They are less likely to be satisfied with the level of support they receive, despite receiving more than they give (i.e. despite non-reciprocity). Networks may show evidence of shrinkage before the onset of illness. From this, it appears that psychosis is characterized by a reduction of social opportunities and a parallel narrowing of focus, most consistent with it being a consequence of the illness itself. This is supported by studies showing the association of small network size with longer duration of untreated psychosis and poor premorbid functioning.
However, small networks presage the level of positive and negative symptoms in people with psychosis up to 3 years after initial assessment, independently of a range of other potential predictors. Gayer-Anderson & Morgan (Reference Gayer-Anderson and Morgan2013) also draw attention to follow-up studies of normal populations, in which small networks predict the emergence of quasi-psychotic experiences. Although relatively small social networks are typical of people with psychosis, the size and other characteristics of the networks do vary greatly from person to person (Angell & Test, Reference Angell and Test2002). It is these inter-individual differences that appear to relate to illness course and outcome.
These results suggest reciprocal relationships between reduced social networks, increasing symptoms of psychosis, and poorer outcomes. This has the potential to form an unpleasant vicious circle, and implies at least some aetiological role for social relationship deficits. In this commentary, we argue that the postulated aetiological role becomes more plausible if we can identify effective mechanisms. This in turn implies that effective treatments could be directed at any point in the circle. In practice, changing the characteristics of social networks (e.g. by befriending interventions or facilitating social support) has proved difficult (Sensky et al. Reference Sensky, Turkington, Kingdon, Scott, Scott, Siddle, O'Carroll and Barnes2000). One reason for this may be that some people with psychosis may not be motivated to maintain friendships. Thus, interventions that focus on increasing the numbers of friends available may not be targeting the right problem: changing perceptions of inadequate support may be required.
Social networks
Social relationships are clearly important for general health and well-being (Thoits, Reference Thoits2011). Those with good social relationships and social support are physically and mentally healthier and live longer (Holt-Lunstad et al. Reference Holt-Lunstad, Birmingham and Jones2008). Over the past four decades, researchers have tried to define, describe and isolate the quantitative and qualitative characteristics of social networks and social support responsible for these protective effects. Some of these effects will be generic, but others may be specific to particular conditions such as psychosis.
The various definitions of social networks in the literature represent relatively minor variations on a theme. Thus, social networks are only a subset of the totality of social contacts, a subset defined by persistence over time and a degree of significance. This is because they are held to play an important role in the maintenance of the psychological and physical integrity of the person. Primary relationships (the main source of social support) have been distinguished from less personal secondary relationships.
On the basis of these definitions, social networks can be specified further in structural and functional terms. The former denote the existence and pattern of interconnections of network members rather than the content or quality of relationships. Functional aspects refer to what is provided by, or perceived to be available from, social relationships. Social networks are also described in terms of quantitative and qualitative aspects. Quantitative aspects include network size and density, kin v. non-kin composition, marital status and whether living alone. Qualitative aspects connote the degree of satisfaction with social relationships. These include reciprocity (the extent to which the relationship is characterized by giving as well as receiving), accessibility (the ease with which network members can be contacted), multiplexity (the number of separate functions provided by relationships), social isolation (pervasive lack of social contact or communication), the presence or absence of a confidant, and loneliness.
Social support
Social networks are important because they are the vehicle for delivering social support. However, while they are necessary conditions for social support, they are not sufficient. Studies of social networks in relation to psychiatric disorder are essentially using them as a proxy for (often unspecified) elements of support. However, support does not map perfectly onto the attributes of social networks. For example, someone with a large network may actually feel lonely, and perceived social support is not necessarily greater in more extensive social networks.
Social support is a functional attribute of social relationships: frequently mentioned functions are emotional, informational and instrumental (Alloway & Bebbington, Reference Alloway and Bebbington1987; Thoits, Reference Thoits2011). A number of definitions emphasize the cognitive elements.
However, social support covers several sub-constructs. The most commonly distinguished are received and perceived support. Received social support refers to the actual provision of supportive behaviours by others; while perceived social support concerns the recipient's perception of, and satisfaction with, the general availability of support. The main difference is that perceived social support refers to anticipating help in times of need; whereas received social support involves the recall of support provided over a given period. The distinction is important because research consistently shows stronger links between health and perceived social support.
It is important to note that not all social relationships are supportive or perceived as supportive; they can be a source of stress and strain, and some are abusive. Social withdrawal may consequently be protective, by insulating individuals from stressful relationships (Wing, Reference Wing1978). However, even when someone withdraws in this way, they may still feel the effects of lack of support in the form of distress and loneliness (Duberstein et al. Reference Duberstein, Conwell, Conner, Eberly, Evinger and Caine2004).
Loneliness in psychosis
Loneliness is a distressing negative experience resulting from a discrepancy between the desired and the perceived state. It has both a cognitive element (the perception of relationship inadequacy) and an affective character (an unpleasant experience). Although loneliness is influenced by quantitative and objective aspects of social networks (e.g. size of social network and frequency of contact), it is determined more by subjective appraisals, such as satisfaction and perceived adequacy. Loneliness does not equate with objective social isolation: the number of their friends is not a good predictor of how lonely someone feels.
The wide-ranging effects of loneliness are apparent. It is associated with reduced life satisfaction, psychosocial problems (low self-esteem and reduced social competence), and mental health difficulties including anxiety and schizophrenia (Neeleman & Power, Reference Neeleman and Power1994; DeNiro, Reference DeNiro1995). It is accompanied by a variety of negative affective states such as shyness, boredom or feelings of alienation. In people with schizophrenia, DeNiro (Reference DeNiro1995) found an increase in alienation, social isolation and loneliness over the patient's lifetime, and, conversely, a decline in positive connections with others. Neeleman & Power (Reference Neeleman and Power1994) found patients with psychosis had the smallest social networks, and felt lonelier than all other patient groups, independently of network size.
Confidants for people with psychosis
A confidant is someone who provides a relationship characterized by emotional intensity, reciprocity and availability. Its absence is strongly linked to the experience of loneliness (Green et al. Reference Green, Richardson, Lago and Schatten-Jones2001). Even at first episode patients have significantly fewer confidants than controls: Morgan et al. (Reference Morgan, Kirkbride, Hutchinson, Craig, Morgan, Dazzan, Boydell, Doody, Jones, Murray, Leff and Fearon2008) recently showed that first-episode psychosis patients were seven times less likely to have a confidant.
Mechanisms of change in social networks and social support
Although the effect of poor social networks and poor social support on physical and mental health is well established, our understanding of the mechanisms by which they operate is limited. We have chosen to argue here for the (neglected) role of psychological mechanisms (perceptions and their accompanying negative cognitions and affect), on the grounds that physiological responses are likely to be influenced by the psychological consequences of suboptimal social relationships (Thoits, Reference Thoits2011), and because the effects of specific psychological processes have been progressively substantiated in recent cognitive models of psychosis (Garety et al. Reference Garety, Bebbington, Fowler, Freeman and Kuipers2007; Fowler et al. Reference Fowler, Hodgekins, Garety, Freeman, Kuipers, Dunn, Smith and Bebbington2012).
Psychosis and the family
The narrowing of social networks in the preamble to psychosis often means that social support becomes restricted to family members. The ways in which family relationships affect the course of psychotic symptoms have been extensively investigated. High levels of expressed emotion in key relatives predict patient outcomes in schizophrenia, and family intervention can reduces relapse rates (Bebbington & Kuipers, Reference Bebbington and Kuipers1994; Pharoah et al. Reference Pharoah, Mari, Rathbone and Wong2006). Patients are certainly able to perceive carer criticism accurately, and such perceptions are linked to poorer functioning (Onwumere et al. Reference Onwumere, Kuipers, Bebbington, Dunn, Freeman, Fowler and Garety2009). Criticism is the direct transmission to the patient of negative judgments, and is associated with negative affect and poor self-esteem (Barrowclough et al. Reference Barrowclough, Tarrier, Humphreys, Ward, Gregg and Andrews2003).
In contrast, carer warmth protects against relapse (Bertrando et al. Reference Bertrando, Beltz, Bressi, Clerici, Farma, Invernizzi and Cazzullo1992). Such supportive relationships are likely to operate by reducing negative affect in patients and in some trials of psychological therapy, ‘befriending’ or supportive therapy resulted in these kinds of improvements, although not sustained (Sensky et al. Reference Sensky, Turkington, Kingdon, Scott, Scott, Siddle, O'Carroll and Barnes2000). It may also explain the better response to psychological treatments of patients in contact with carers (Garety et al. Reference Garety, Fowler, Freeman, Bebbington, Dunn and Kuipers2008).
Relational regulation theory
Building on attachment research, Relational Regulation theory (RRT; Lakey & Orehek, Reference Lakey and Orehek2011) contains explicitly cognitive and affective elements. Specifically RRT asserts that ‘main effects of social support on mental health occur when people regulate their affect, thought and action through ordinary yet affectively consequential conversations and shared activities, rather than through conversations about how to cope with stress’ (p. 482). According to RRT ‘perceived support typically does not directly cause affect but emerges from the types of social interaction that successfully regulate affect’ (p. 490).
Social support and cognitive models of psychosis: common ground
Impairments in social networks and support are associated with increased rates of dysphoric mood, specifically anxiety and depression. They are also associated with effects on a range of cognitions. Cognitive models of psychosis (e.g. Garety et al. Reference Garety, Bebbington, Fowler, Freeman and Kuipers2007) propose that the emergence and maintenance of psychotic disorder involves both cognitive and emotional processes. The models postulate that environmental factors operate through psychological mechanisms including an increased vulnerability to anxiety, depression and low self-esteem, and enduring cognitive biases affecting the processing of negative events and experiences. The latter emerge from the consequences of adversity on expectations, and includes schematic beliefs about self and others, and distortions in attributional style, such as a tendency to attribute power and control to others (Birchwood et al. Reference Birchwood, Gilbert, Gilbert, Trower, Meaden, Hay, Murray and Miles2004; Garety et al. Reference Garety, Bebbington, Fowler, Freeman and Kuipers2007).
Sündermann et al. (submitted for publication) recently suggested that the relationship between loneliness and paranoia may be mediated by affective processes such as anxiety or depression. This is because loneliness may distort thinking processes by exaggerating threat appraisals, or may make it harder for lonely patients to think of alternatives to their unusual ideas because they have no one to discuss them with. This may raise anxiety levels, leading in turn to increased paranoia.
Conclusion
Social relationships are important in facilitating recovery in psychosis. Social support, or the lack of it, shapes many of the affective changes and cognitions seen as significant in the development and maintenance of psychotic symptoms. This complementarity implies potential mechanisms linking deficits in social support and psychosis. The field is open for further research. It would be particularly helpful to design and test cognitive behavioural interventions aimed specifically at the perceived effects of poor social support in psychosis, in a productive way. Focusing on the mitigation of cognitions that relate to loneliness is likely to be a helpful starting point.
Conflict of Interest
None