Understanding the links between social ties and health

Social relationships and health

In the most compelling evidence to date, a recent meta-analysis found that positive aspects of relationships (i.e., perceived social support) was associated with a lower risk of mortality (Holt-Lunstad, Smith, & Layton, 2010). Indeed, effect sizes from the meta-analysis appeared as large, if not larger, than standard risk factors such as smoking, exercise, and obesity. However, even relationships that are relied upon to be major sources of support are not uniformly positive and can add to a person’s distress during their time of need (e.g., ineffective information, feeling frustrated or let down by the support provider; Newsom, Nishishiba, Morgan, & Rook, 2003). This is consistent with a smaller but growing epidemiological literature documenting the health risks associated with negative social ties (De Vogli, Chandola, & Marmot, 2007).

Importantly, most of the prior studies in this literature have examined relationship positivity or negativity in isolation. This is a significant issue, because positive and negative aspects of social relationships tend to be separable factors (Newsom et al., 2003). These data suggest that relationships may differ in their positive and negative substrates, with some relationships characterized by relatively strong feelings of both positivity and negativity (Uchino, Holt-Lunstad, Uno, & Flinders, 2001).

We have developed a more general model for examining the health-related consequences of social relationships that incorporates both positive and negative aspects (see Uchino et al., 2001). A unique feature of our model is the specification of ambivalent relationships that are viewed as relatively high in both positivity and negativity (e.g., overbearing parent, volatile romance, "out of touch" friend; Fingerman, Hay, & Birditt, 2004). Despite the positivity in such relationships, we have hypothesized that the co-occurrence of negativity may be uniquely associated with worse health outcomes. This may be because ambivalent ties require heightened vigilance during social interactions, or may be frustrating and ineffective sources of support during times of need (Uchino et al., 2001). In addition, ambivalent ties are typically described as "close," and hence there is more of an overlap between self–other representations (Aron, Aron, Tudor, & Nelson, 1991), which can exacerbate any interpersonal stress with such network ties. In fact, ambivalent ties engage in more negative behaviors (e.g., criticism) and less emotionally supportive behaviors and thus appear stress-enhancing (Reblin, Uchino, & Smith, 2010).

Ambivalent ties are also not an isolated feature of most individuals' social networks. They comprise almost 50% of important network members and hence have ample opportunity to influence health-related outcomes (Campo et al., 2009). In fact, we have found consistent evidence that ambivalent ties are related to worse outcomes compared to other relationship types (e.g., primarily positive or primarily negative), such as increased cardiovascular reactivity at both conscious and less conscious levels of processing (Carlisle et al., 2012; Holt-Lunstad, Uchino, Smith, & Hicks, 2007), higher ambulatory blood pressure during daily life (Birmingham, Uchino, Smith, Light, & Carlisle, 2012; Holt-Lunstad, Uchino, Smith, Cerny, & Nealey-Moore, 2003), greater inflammation (Uchino, Bosch, et al., 2012), and even increased cellular aging (Uchino, Cawthon, et al., 2012).

On building stronger bridges between relationship science and health

Given the evidence to date linking positive and negative aspects of relationships to health, what is needed in order to move this research forward? I argue that we need to build stronger bridges between relationship science and health. A broad, simplified model to facilitate such connections is shown in Figure 1. At a basic level, relationship positivity, negativity, and ambivalence influence biological processes and physical health outcomes. However, there are two issues that are in need of greater attention in order to facilitate important “second-generation” questions. Firstly, what are the antecedent processes that are related to relationship positivity, negativity, and ambivalence from a relationship science perspective that are health-relevant? Secondly, what are the more complex (coordinated) biological pathways linking relationships to health from biomedical and neuroscience perspectives? Importantly, these questions represent critical “bridge” issues of interest to both disciplines and can facilitate more integrative and stable connections between relationship science and health.

OPEN IN VIEWER

Figure 1.

Broad model highlighting important "bridge" topics that can foster stronger links between relationship science and health.

The past several decades have seen dramatic developments in the fields of relationship science and health psychology. However, these two areas have advanced relatively independent from each other (Pietromonaco, Uchino, & Dunkel-Schetter, 2012). The net effect is that we know much less about the antecedent processes that are directly related to physical health from a relationship science perspective. In order to systematically address this question and foster theoretical modeling, one strategy is to test the health-relevance of relationship science constructs that have been linked to positivity (e.g., support), negativity (e.g., tensions), and/or ambivalence. There are several constructs that appear relevant.

Firstly, individual differences in attachment style appear important to support and conflict negotiation (Pietromonaco et al., 2012). For instance, secure individuals provide responsive support and are better at managing negativity in their close relationships (Kane et al., 2007). Although there is a literature linking attachment styles to health outcomes (e.g., Gouin et al., 2009), work would also be needed that directly models support and conflict as mediators of any such links (Pietromonaco et al., 2012). Relationship goals have also been linked to support and conflict, although none of this work to date appears to have been applied to the health domain (Canevello & Crocker, 2011; Gable, 2006). In addition, sexuality has an influence on multiple interpersonal processes in the model, but stronger direct tests of its links to health are needed (Diamond & Huebner, 2012). Finally, positivity and negativity in close relationships may be the result of complex interpersonal emotion regulation patterns (e.g., synchrony, reciprocity) that have implications for health (Butler, 2011).

One potential antecedent process from relationship science that we have recently begun testing is related to partner knowledge (Fletcher & Kerr, 2010). We have been examining a specific form of partner knowledge that we coined “attitude familiarity.” Decades of work in social psychology has found that personal attitudes are functional in that they facilitate information processes, guide decisions, and behaviors (Fazio, 1990). We reasoned that knowledge of a partner's attitudes is similarly functional in that it might enable one to avoid conflict or provide more responsive support (Neff & Karney, 2005). Consistent with this possibility, our initial study showed that greater attitude familiarity with a spouse was associated with greater daily life perceived partner responsiveness and interaction positivity, as well as lower perceived interaction negativity (Sanbonmatsu, Uchino, & Birmingham, 2011). Importantly, greater attitude familiarity also predicted lower ambulatory blood pressure during daily life (Sanbonmatsu et al., 2011). This preliminary research highlights the utility of conceptualizing how constructs in relationship science might be linked to health-relevant interpersonal and biological outcomes. Such work can extend existing theoretical models in terms of their domains of applicability and facilitate more integrative connections to other constructs and processes that need strong consideration in a health context (e.g., dyadic coping with chronic health conditions).

Most of the prior work in relationships and health has examined a limited number of biological outcomes, often within the same system (e.g., cardiovascular). However, these biological systems represented coordinated responses aimed at fostering adaptation, but which may also have long-term costs (Miller, Chen, & Cole, 2009). There is elegant work from our biological colleagues examining how these systems are coordinated, typically during exposure to stress. This work should now be expanded to examine similar coordinated processes as a function of our relationships.

There are many opportunities for more complex biological modeling both within and across systems. Health and relationships researchers have typically examined indices of cardiovascular reactivity, especially heart rate and blood pressure. Heart rate is a function of both sympathetic and parasympathetic influences and there are good non-invasive measures of these inputs in terms of pre-ejection period and respiratory sinus arrhythmia (RSA), respectively (Cacioppo et al., 1994). In one study, we found that the links between ambivalent ties and heart rate were mediated by RSA changes which might reflect a disruption in self-regulatory processes (Carlisle et al., 2012). Blood pressure is similarly a multiply-determined endpoint reflecting both cardiac output and peripheral resistance. Both of these can be estimated non-invasively using impedance cardiography. It is important to model these underlying determinants, because peripheral resistance changes are more closely linked to threat appraisals and hence might be more health-relevant (Seery, 2011).

Modeling relevant biological responses across systems is also important. Health researchers have focused much attention recently to inflammation as it appears to influence multiple disease endpoints (Hawkley, Bosch, Engeland, Marucha, & Cacioppo, 2007). An important cytokine receiving attention is IL-6, which has both pro and anti-inflammatory properties (Hawkley et al., 2007). One important “trigger” of IL-6 is activation of the sympathetic nervous system (Soszynski, Lozak, Conn, Rudolph, & Kluger, 1996). Once released, IL-6 is a potent stimulator of the hypothalamic-pituitary-adrenal (HPA) axis resulting in the release of cortisol, which is often an attempt to control inflammation (Webster, Oakley, Jewell, & Cidlowski, 2001). However, under conditions of chronic stress, cytokine-producing cells can become less sensitive to the inhibitory effects of glucocorticoids, hence bypassing one important control mechanism (Miller, Cohen, & Ritchey, 2002). Given the potency of social processes (Miller et al., 2002; Padgett et al., 1998), modeling coordinated biological responses need strong consideration in the context of our social relationships.

A second opportunity to model responses across systems is related to recent neuroimaging work. Exposure to stress reliably increases activity in the dorsal anterior cingulate cortex (dACC), anterior insula, periaqueductal gray (PAG), and deceases activity in the ventromedial prefrontal cortex (vmPFC; Gianaros & O’Connor, 2011). This is important, because these brain regions have direct and indirect links to the peripheral biological responses often studied by relationship researchers and thus can help us understand how the brain coordinates these responses. There are several predictions one might make based on prior relationship work. Firstly, supportive ties are associated with increased activity in the vmPFC (Eisenberger et al., 2011), which might decrease peripheral physiological responses such as heart rate and blood pressure. In comparison, ambivalent ties appear to exacerbate interpersonal stress and hence might be associated with greater activity in the dACC, anterior insula, and PAG, which in turn would potentiate peripheral physiological responses (Wager et al., 2009).

Conclusions

Relationship scientists have made tremendous progress in understanding the processes underlying relationship formation, maintenance, and dissolution. At the same time, health psychologists have documented the health consequences of basic relationship processes, such as support, negativity, and ambivalence. In order to move work on relationships and health forward, greater integration of these two areas is required. I highlighted two potential “bridge“ areas of interest to both disciplines, but there are certainly others that might be gleaned from the broad model. Given the dynamic nature of relationships (Butler, 2011), the antecedent processes depicted in Figure 1 may also serve as direct mediators of links between basic relationship processes and health, as very little work exists on this question as well (Pietromonaco et al., 2012). For instance, relationships are strongly linked to the experience of emotions (e.g., Berscheid & Ammazzalorso, 2001) and, hence, might be important mediators of links to health. The pursuit of such integrative questions is likely to be fostered by (a) broadening the training of researchers in both areas (e.g., postdocs, sabbaticals) and/or (b) teams of interdisciplinary experts. I believe such integrative work will play a central role in helping us understand why relationships are health-relevant and what might be done about it.