The Fracture of Relational Space in Depression: Predicaments in Primary Care Help Seeking

Abstract

Primary care clinicians treat the majority of cases of depression in the United States. The primary care clinic is also a site for enactment of a disease-oriented concept of depression that locates disorder within an individual body. Drawing on theories of the self and stigma, this article highlights problematics of primary care depression treatment by examining the lived experience of depression. The data come from individuals who screened positive for depressive symptoms in primary care settings and were followed over ten years. After iterative mixed-methodological exploration of a large dataset, we analyzed interviews from a purposive sample of 46 individuals using grounded and phenomenological approaches. We describe two major results. First, we note that depression is experienced as located within and inextricable from relational space and that the self is experienced as relational, rather than autonomous, in depression. Second, we describe the ways in which the experience of depression contradicts a disease-oriented concept such that help-seeking intensifies rather than alleviates the relational problem of depression. We conclude by highlighting that an understanding of illness experience may be essential to improving primary care depression treatment and by questioning the bracketing of relational concerns in depression within the construct of stigma.

Introduction

Over just the past few decades, various complex human experiences – sadness, anxiety, loss, loneliness, mourning, guilt -- have been partitioned and reassembled into categories that lend themselves to biomedical management, such as the category of depression (Bromley 2011; Callahan and Berrios 2004). Many scholars have documented this transformation from within the psychiatric profession, tracing psychiatrists' adoption of biomedical concepts and practices (Healy 1997; Horwitz and Wakefield 2007; Kleinman 1988; Metzl 2003; Wilson 1993). Yet a focus on the intellectual, discursive, and procedural context of psychiatry overlooks the most common biomedical spaces in which depression emerges and circulates: the primary care clinic. In the U.S., primary care clinicians -- internists, family practitioners, and other generalist physicians -- identify and treat the majority of cases of depression (Callahan 2006; Kessler, et al. 2007; Pence, et al. 2012). Perhaps two-thirds of those with depression first seek care in a general medical clinic (O'Connor, et al. 2009), and they often receive antidepressant medications from generalists rather than mental health specialists (Marcus and Olfson 2010; Olfson and Marcus 2009). Moreover, primary care physicians treat depressive illnesses that are as severe as those treated by mental health specialists (Gaynes, et al. 2005).

In this paper, we chart the problematics of primary care depression treatment by examining the lived experience of depression in social context. To begin, we note that the centrality of the primary care clinic in the management of an emotional disturbance reflects the cultural predominance of a disease-oriented concept of depression. This disease-oriented concept locates depression within a specific organ (i.e., the brain) within an individual body. The disease of depression is comprised of a set of shared symptoms (e.g., low mood, sleep disruption, hopelessness, tearfulness), a recognizable course, an expected response to treatment, and (where they can be posited) shared biological markers (Ghaemi 2013). A disease-oriented concept situates depression within the individual through analogy to a physical condition like diabetes mellitus. The concept suggests treatment proceeds just as it would in physical illness: the patient recognizes symptoms as inconsistent with health, seeks help from a clinician, identifies with a diagnosis to explain the experience, and adheres to a treatment for as long as the condition persists. One appeal of a disease-oriented concept of depression is that it appears to be value-neutral. Based in abiding biological aberrations, this view suggests, the depression diagnosis should imply nothing about the sufferer's moral or social standing.

Here we use interview data from a large sample of individuals who screened positive for depressive symptoms in primary care settings to raise questions about the suitability of this analogical assertion and the value-neutrality of the disease-oriented concept. With an ontological orientation that queries the nature of depression, we illustrate that depression is experienced as located within and inextricable from relational space. We use the phrase relational space to mark the contours of the valued social contexts where one's most intimate interactions and interdependencies occur, such as the family. As we will show, the experience of depression in relational space can contradict so sharply with a disease-oriented concept of depression that the most common avenues for help become obstructed. That is, a disease-oriented concept can feel not only irrelevant but can intensify the relational problem of depression. This argument continues a line of medical social science work that compares dominant biomedical concepts of disease with everyday lived experience to identify sites of resistance, barriers to care, or novel forms of suffering generated or potentiated by biomedical interventions (Bromley 2007; Bromley 2011; Erikson 2012; Garro 1994; Lester 2009; Martin 1987; Timmermans and Buchbinder 2010). The paper also builds on theoretical understandings of stigma and ultimately raises questions about the degree to which current approaches to stigma account for the social and interpersonal challenges of the depression experience.

Background

One estimate indicates that almost 10% of primary care visits include an antidepressant prescription (Mojtabai and Olfson 2011). Yet the primary care treatment of emotional distress has had limited effectiveness. Many of those affected by depression receive little or no care (Alderson, et al. 2014; Bassuk and Beardslee 2014), and depression remains one of the most common causes of disability in the U.S. (Hall and Reynolds 2014). Substantial racial, ethnic, and socioeconomic disparities in depression treatment and outcome persist (Alegria, et al. 2008; Miranda, et al. 2008). One study estimates that primary care clinicians identify only half of those with depression, deliver adequate treatment to 9%, and achieve remission in 6% (Pence, et al. 2012). Rates of antidepressant discontinuation in primary care depression treatment are extremely high (O'Connor, et al. 2009) with over 40% of patients discontinuing an antidepressant within 30 days and another 30% discontinuing within 90 days of initiation (Olfson, et al. 2006).

Improving the quality of primary care depression treatment has long been a high priority goal. Health services researchers target factors inside the clinic such as provider knowledge and have developed models of collaborative care with mental health professionals (Katon, et al. 2010; Unutzer, et al. 2008; Unutzer, et al. 2002; Wells, et al. 2007a). Quality improvement efforts have also focused on the sociocultural context outside of the clinic that can impact treatment (Pescosolido, et al. 2013). For instance, depression researchers have identified stigma as a significant impediment to care. In depression, stigma has been associated with low intention for psychological help-seeking, delays in treatment, low rates of disclosure to clinicians and family members, poor treatment adherence, and poor treatment outcomes (Corrigan and Watson 2002; Link, et al. 1987; Perlick, et al. 2001; Reynders, et al. 2014; Sartorius and Schulze 2005; Tancredi, et al. 2013).

Stigma is often described functionally in the health services literature as a barrier to treatment, but stigma is a multilayered concept with roots in both sociological and anthropological thought. Erving Goffman (1963) first elaborated a contemporary concept of stigma to mark the extreme negative case of the performance of social identity. Goffman defined stigma as a mark that diminishes, a discrediting label that converts an individual from a usual person to one who is tainted (Corrigan and Shapiro 2010; Jurgensen, et al. 2012). The mark of stigma transforms the individual into one seen as less than fully human as a result of physical deformities, deficiencies of character, or membership in undesirable tribes (e.g., due to race, gender, or religion) (Pescosolido 2013). Goffman described mental illness as a “moral stigma” because it indicated a fundamental deficiency of character, such as a lack of willpower or self-control (Thoits 2011). Stigma as an “adverse social judgment about enduring features of identity” (Scambler 2009: 441) can lead to various forms of othering including rejection, devaluation, avoidance, and distancing in social spaces.

In these ways, theorists building on Goffman's work situate stigma with reference to an individual's position in a particular social context over time (Hinshaw 2007; Pescosolido and Martin 2007; Scambler 2011). Stigma may indicate a simple mark of difference but it “can be enacted only in social interaction” and its characteristic impact is “exclusion from full participation in society” (Pescosolido 2013: 3). In this frame, stigma is an inherently social event whether exclusion, judgment, and rejection are actually experienced or merely anticipated. In their modified labeling theory of stigma, Link and Phelan describe stigma as proceeding through a series of interrelated interactional processes that begin with the labeling of the individual, lead to cognitive and emotional separation (e.g., “us” versus “them”), and culminate in the loss of status and social power (Link, et al. 1989; Link, et al. 2004; Scambler 2011; Yang, et al. 2007). Stigma is understood to emerge in interactions in which one's status may be in play (Kranke, et al. 2011) and functions to define social roles by “conveying the devalued status of some identities relative to others.” (Steward, et al. 2008: 1226). Stigma can be potentiated by other types of discrimination (e.g., racism) such that stigma can replicate and reinforce relations of power that are more broadly based than categories of illness and health (Parker and Aggleton 2003).

Still other theorists identify stigma as a psychological experience, an attribute that can be incorporated into a concept of the self when the individual identifies with the stigmatizing label. Internalized stigma or self stigma (Corrigan et al, 2006) develops where self-perceptions include stigma-derived concepts. Internalized stigma can develop through mere awareness of social stereotypes, through socialization to these negative judgments (such as within the family or peer group), or through the repeated validation of stigma in social interactions. Through these and other experiences, one may learn to endorse the stigma and apply it to oneself (Corrigan, et al. 2009; Markowitz 2014; Scheff 2003; Yang, et al. 2010b). The effects of this internalization are varied (Steward, et al. 2011) but have been proposed to include concealment of the stigmatized condition, withdrawal from others, decreased self esteem, depression, and a decline in social integration (Link, et al. 1991; Link, et al. 2001; Thoits 2013). Those who fear stigma may isolate or perpetuate their own marginalization (Bromley, et al. 2013). Scambler and Hopkins' hidden distress model distinguishes discrimination from the psychological experience of stigma as shame and self-devaluation. They call the former enacted stigma and the latter felt stigma (Saunders 2014; Scambler and Hopkins 1986). In these ways, the social and psychological manifestations of stigma may be reinforcing, and stigma is understood to be a force that can undermine wellbeing not only through the creation of differential social conditions but also as a result of actual or perceived threats to one's social identity (Steele, et al. 2002).

Recently, scholars from the anthropological traditional have begun to critique these concepts and reexamine stigma experience. With an eye toward personal agency and meaning-making (Paul 1990), these scholars have encouraged a look at the intersubjective dynamics of stigma within local contexts as well as the varieties of stigma manifestations across cultures (Jenkins and Carpenter-Song 2008), a development labeled by Yang and colleagues as a “moral turn” in stigma theory. Advocating for an examination of stigma as a threat to what most matters in individuals' unique lives, these theorists argue that sociological concepts have erred in constructing stigma as predominantly dyadic, as if occurring between discrete recipients and stigmatizers, and in positing dynamics of stigma that are overly determinative, as if its emergence in specific social structures might be inevitable (Yang and Kleinman 2008; Yang, et al. 2007). Moreover, proponents of the moral turn argue that stigma theory has focused too narrowly “on institutional forces and middle-class expectations regarding normative patterns of behavior” (Jenkins and Carpenter-Song 2008: 382). Indeed, as Thoits says (Thoits 2005; Thoits 2011), the fundamental assumption of the sociological perspective on stigma assumes the presence of powerful actors capable of imposing undesirable identities on less powerful others in order to maintain social order or further elite interests.

The moral turn in stigma theory reflects broader shifts in cultural theory that attend to the importance of lived experience in context, subjectivity and intersubjectivity, social suffering, and individual agency (Biehl and Moran-Thomas 2009; Duranti 2010; Kleinman 1988; Kleinman, et al. 1997; Mattingly 2012a). Stigma gains meaning as subjectively experienced, but its content and power must be understood in relation to the dearly-held interpersonal, social, and cultural resources it threatens (Shostak, et al. 2011) as well as the avenues for endurance, resistance, and resilience that individuals construct for themselves (Biehl 2005). Stigma's “social elements” must be uncovered “through examining … a person's moral standing [as] lodged within a local social world,” a standing enacted and upheld through intrapersonal and interpersonal commitments (Yang and Kleinman 2008: 399). To these theorists, stigma matters because it “can destroy an actor's capacity to take part in the core, everyday engagements that identify one as a moral or ‘full adult’ person in a particular context” (Yang, et al. 2010a: 837). This consideration of the “moral dimensions of human experience” (Jenkins and Carpenter-Song 2008: 384) is thus more than a look at social, relational, or psychological aspects of stigma. It entails attention to the emergent and shifting concerns that matter most to individuals within the course of day-today predicaments in local worlds (Garro 2011; Garro and Yarris 2009; Linger 2010; Wikan 1990).

These theorists draw on a meaning of the moral that is quite distinct from Goffman's meaning. Everyday life, Kleinman says, is “innately moral, because it is in local worlds that the relational elements of social existence in which people have the greatest stake are played out…. [T]he fact that some things really do matter, matter desperately, is what provides local worlds with their immense power to absorb attention, orient interest, and direct action” (Kleinman 1997: 327, quoted in Garro and Yarris 2009: 561-62). With this in mind, stigma researchers are attending to commonplace sites of meaning making, such as the family, as we do in this paper. Surprisingly, the family context has to date been a relatively uncommon site for the study of stigma. Yet Moses et al (2010) found that almost half of adolescents with ADHD reported stigmatizing treatment – such as rejection, discriminatory behavior, or negative attitudes -- by members of their immediate or extended family.

Yet we would note that, understood as imbricated within the moral content of everyday engagements, stigma would be difficult to disentangle from the illness experience itself. If stigma is experienced as a threat to what matters most, then it may be felt as intertwined with the array of self reappraisals, experiences of coming-to-terms-with, future redefinitions, speculations about culpability, and other crucibles within the experience of illness (Charmaz 1991; Conrad 1990; Frank 1995; Scambler 2009). For instance, Karp (1994) describes the recursive redefinitions of the self that occur in those experiencing depression as a “depression career.” In this career, the individual moves through phases of processing inchoate feelings of distress, recognizing that something is “really” wrong, coming to grips with the implications of the diagnosis, and perhaps defining depression as manageable. Karp's is a moral journey in which a previously whole individual can struggle with adversity and emerge as changed but whole again (Clark 2008). Other first-hand accounts of depression emphasize the ways in which depression ruptures the individual's bridge to others. Its “incomprehensible, hardly expressible,” subjective experience “robs them of everything fundamental to human existence” (Westerbeek and Mutsaers 2008: 46). Wolpert asserts that the question of culpability encircles the depression experience. He emphasizes the starkness of the deep, unbidden shame of depression, at once interpersonal and cultural: “those with depression see it as something to be ashamed of, and so kept secret” (Wolpert 2001: 222). Analyzing depression narratives, Kangas asserts that “stories of depression are stories of marginalization” (Kangas 2001: 90). They inevitably entail, according to Kangas, a loss or inability to engage in the relatedness that provides the terrain for daily life, acceptable adulthood, and engaged citizenship. The narrative data we present echo many of these themes. In these accounts, issues of labeling, separation from others, shame and loss are difficult to distinguish as a discrete experience of stigma. They are the emotional, cognitive, and behavioral manifestations of depression.

In this paper, we build on these theoretical perspectives on the moral relevance of the interpersonal contexts that comprise lived experience. We further use our data to raise fundamental questions about the ways in which the self is experienced and constructed in depression. In considering the self, we draw on Hollan's updating of Hallowell's (1959) concept to describe the self “as a locus of subjective experience” (Hollan 1992: 284). This self is a conscious product of reflection and can be described to others. That is, the self is “that part of consciousness that comes into play when a human being begins to take him- or herself as an object” (Hollan 1992: 284). Psychological anthropologists have described two models of the self, one of an egocentric, autonomous, and closed self and a second of a sociocentric, relational, interdependent, and open self. While both models are ideal types that oversimplify the variety and dynamism of experiences of the self in context (Kusserow 1999; Lindholm 1997), we use them here to frame a consideration of the experience of the self in depression, particularly in relation to help-seeking possibilities. For instance, Wikan describes the egocentric self as wholly contained and independent of others and therefore free to be ‘one's own person’, to remake oneself, and to alter or construct one's experience through one's own efforts (Wikan 1995: 274). Such a self may be free to pursue strategies to improve emotional health through alterations in thoughts and feelings -- even in the absence of contextual changes. In contrast, as Wikan says, the “relational self is seen as so inextricably woven into a fabric of culturally prescribed social roles [and] patterns of interpersonal behavior … that it is more appropriate to speak of the person as the ‘person-in-relationships,’ merged with, or co-constructed with others … rather than as a discrete, well-bounded unit” (Hollan 1992: 284). The relational self is “inextricable from context, or ‘always already’ relational” (Linger 2010: 218), and this self may always understand help-seeking as implicative in this context.

In many ways, health researchers have embraced the concept of the self as tightly bound to others, such as in the recent resurgence of interest in the family as a source of support for health care (Celano, et al. 2012; Chung, et al. 2014; Lewellen 2012; Rosenberg, et al. 2013; Sigurdardottir, et al. 2014; Timmermans and Freidin 2007) and in studies of social network impacts on health (Christakis and Fowler 2007). Indeed, a relational self has been examined as a core value among Hispanic individuals through theories of familism, and some have hypothesized that familism functions as a source of support for treatment. Yet we would also note the complexity of the question of whether strong family ties facilitate or undermine health-related behaviors. For instance, Kondo describes feeling “bound and trapped by social convention and living my life for others” in her Japanese field work (Kondo 1986: 81). Some suggest that familism can undermine help-seeking in depression if strong ties to family members overshadow the needs of individuals (Keeler, et al. 2013). Our data contribute to these debates by revealing the ways in which the self is experienced in depression as relational and by mapping the particular ways this relational self is challenged during help-seeking and treatment.

Methods

Study Site and Sample

Our informants comprise individuals with depression identified in primary care settings as part of a group-level randomized controlled trial of quality improvement interventions for primary care depression treatment, the Partners in Care (PIC) study. PIC enrolled individuals with depressive symptoms from 46 managed care clinics across the United States beginning in 1996 (Wells, et al. 2007b). At enrollment, all PIC participants reported depressive symptoms on a structured instrument and were randomized to usual care or a depression intervention. The PIC intervention included elements to improve the quality of depression treatment delivered in the clinic, such as clinician support and education regarding the detection of depression, improved access to evidence-based counseling or medication management, and resources that helped patients take control of depression management. Participants completed surveys every 6 months for 2 years then at 5- and 9-year follow-up.

At 9-year follow up, participants completed a social network interview that included questions about the proportion of network members to whom the participant had disclosed their depression diagnosis or treatment and the proportion who expressed stigmatizing attitudes (e.g., “Has [NAME] ever made you feel ashamed or embarrassed because of your mental health problems?” “If [NAME] knew about your treatment, do you think they would think less of you?”). At 10-year follow up, all African-American (n=46) and Latino (n=205) participants who completed the 9-year follow up survey were invited to participate in up to 3 semi-structured interviews about their illness experience. Thirty-seven African-Americans and 157 Latinos participated. One hundred and eight non-Latino white participants, matched on gender and clinic site to the minority sample, were also invited for interviews and 86 participated. The final qualitative sample included 280 informants. As described below, our analyses used both a random (n=40) and a purposive sample (n=46) of these informants. The RAND Human Subjects Protection Committee reviewed and approved the protocol.

Interviews at 10-year follow up asked informants about depressive symptoms, medication use, coping, and help-seeking. Other topics included how the informant shared information about their mood or treatment with others and with what response. Interviews were audiorecorded and the interviewer typed notes during and immediately after the interview. In 2012, 8 study team members reviewed notes and, where necessary, listened to audio files to generate summaries of each interview. Study team members included bachelor's- and master's-level research assistants, psychologists, psychiatrists, social scientists, and one internal medicine physician. More than one study team member generated summaries for a minority of informants (11%). The team used all three qualitative data sources (summaries, notes, audiotapes) in qualitative analyses.

Analytic Approach

Our analytic approach evolved iteratively. We were guided by theoretical perspectives on stigma, such as how labeling and distancing may impact treatment; and anticipated that data could indicate correlates of racial and ethnic disparities in depression treatment and outcome. To test these assumptions, one analyst (EB) reviewed interviews from a random sample of 40 informants stratified by race and ethnicity. Content analysis was used to divide text into categories that represented the impact of social context on disclosure, help-seeking, and treatment (Hsieh and Shannon 2005). Content analysis showed that the majority of informants confronted opposition to treatment. Of a total of 40 informants, 6 described those close to them as uniformly opposed to depression treatment (e.g., a spouse who did not want the informant going to therapy; children who opposed the use of psychiatric medications). Seven informants described both opposition and support for treatment (e.g., family members sometimes supported and sometimes derided treatment). Six informants were themselves opposed to treatment. Only two informants received specific encouragement to pursue mental health treatment. We developed a preliminary content analytic codebook to describe the phenomena of encouragement versus discouragement of treatment and its sources (e.g., one's own beliefs, others’ beliefs).

To generate theory about key barriers and facilitators of treatment, we assembled a theoretical sample that maximized differences between informants. We used 9-year social network interview data to identify informants we assumed to be most and least likely to receive support for treatment. We hypothesized that informants in contexts where they disclosed depression and encountered less stigma would manage depressive symptoms and treatment differently than informants in contexts where disclosure was rare and stigma prevalent. We selected 46 informants at the extreme high or low end of scores in the entire sample (n=280) on two variables: frequency of disclosure and prevalence of stigma within the network, to yield 4 informants groups (n=13 low stigma/high disclosure; n=10 high stigma/low disclosure; n=13 low stigma/low disclosure; and n=10 high stigma/high disclosure). Six of the 46 informants were in both the random and extreme case samples. Twenty high stigma informants were sampled, compared to 26 low stigma informants, because of the relatively small number of informants reporting a high prevalence of stigma. Below we describe interview findings from this extreme case sample of 46 informants reporting high and low stigma and disclosure. We use pseudonyms to disguise informants’ identities.

We used the preliminary codebook to mark attitudes toward treatment and stigma-related phenomena. However, as in the random sample, preliminary coding demonstrated that most informants in all four sampling groups confronted significant conflict around treatment. And, the meaningfulness of sampling groups were undermined by the predominance of descriptions of stigmatizing attitudes in the social context from the 26 informants categorized as low stigma and descriptions of significant barriers to disclosure among the 23 informants from high disclosure contexts.

As a result, we used both grounded theory and phenomenological approaches to generate theory that could account for common experiences of help-seeking and treatment (Starks and Trinidad 2007). We chose a grounded theory approach because questions of how informants made meaning in their specific social contexts appeared integral to their depression experience and their choices about treatment (Hamilton, et al. 2011). We used open coding to explore how the informant and others perceived and managed the depression diagnosis and its treatment. We also adopted a phenomenological approach (Smith, et al. 2009), summarizing each informant's predominant lived experience of depression-in-context (e.g., depression as exile; depression as threat to family cohesion). Both open coding and phenomenological findings were reviewed iteratively with 5 members of the research team and revised based on consensus. Then, using a constant comparative approach, we compared open coded segments and phenomenological descriptions to draw connections between concepts. Through these steps, we schematized the role of a relational space in depression experience and then further delineated themes that addressed the types of fractures that interfered with help-seeking. A second analyst (MM) then independently coded a quarter (11 cases) of the sample and agreed on the application of thematic categories in all but one case. Where we present counts, they reflect the informant's predominant experience of depression-in-context. Nonetheless many informants described experiences consistent with more than one theme.

Results

Despite a theoretical case sampling method designed to include informants in supportive and unsupportive social contexts, the majority of informants described significant conflict about depression and its treatment. Only two of 46 informants, a 63-year-old man and a 54-year-old woman, described engaging in and receiving effective depression treatment that they experienced as helpful and appropriate. For instance, the male informant attributed his depression to stress at work, worries about drug use in a daughter, and financial strain. He said that several years before his interviews, he had visited a counselor at work and was encouraged to start Prozac. He took it consistently, describing that it evened his mood and helped his sleep. Residual depressive symptoms led him to ask his doctor for an increase in his dose. He reported that his wife liked that he took it. The female informant described strong family ties and mental health care that she drew on in times of stress. She described crying with family and friends and seeking advice and support for coping during challenges.

Both informants said that supportive family members or friends noticed and validated their symptoms, endorsed their depression diagnosis, and encouraged them to access help. Neither informant described depression as a threat to relationships, to family cohesion, or to their own sense of value to others. But the remainder of the sample, 44 of 46 informants, experienced depression as a less tractable pathology of the relational space. They described depression as an experience in which their closest relationships or their value in their social context were threatened. They moreover described depression's ability to fracture relational space as intensified by help-seeking and treatment. Even when treatment was effective, they described detaching their treatment from their relational life (e.g., through concealment). Some informants described treatment as ineffective because of its irrelevance to the relational issues they understood to be central to their depression. We detail each theme in turn, beginning with the experience of exile and concluding with the experience of the ineffectiveness of effective treatment.

Depression as Exile

For many informants, depression undermined or obliterated a capacity for interpersonal connection. About a quarter of our informants (13 of 46) seemed to find depression inarticulable to the receptive interviewer. These informants were tentative, derailed, and often enigmatic in their interviews. When the interviewer asked about their own feelings, they often talked about others' problems: the emotional needs of grandchildren, health issues in a daughter, or a husband's financial and alcohol problems. Erin, a 42-year-old woman, provides one of the most evocative narratives of the exile of depression. She began her interview with a cryptic set of ruminations. When asked to describe her feelings, she said,

I was kind of feeling sad about some things going on in the family and thinking perhaps if I had made some different choices, wondering if I was responsible, or, whether things would've been different if I'd done things differently, if things would've been different if I hadn't been depressed at the time -- and maybe more in control of my emotions -- and what I said and what I did and how I behaved. So that's kind of made me feel sad, made me cry, and sort of stress out a lot about it.

Despite prompts from the interviewer, Erin did not explain the conflict. She did clarify that, “I haven't talked with anybody” about it. Erin said that her usual strategy for coping with stress was to avoid considering it. Prior to a work evaluation, “I didn't sleep. I was having nightmares, racing thoughts and just -- it was pretty awful,” but she mentioned this to no one because “I saw it as something that would resolve itself.” Similarly, when her feelings were negative, “Generally I try to talk myself out of them if possible.” When she felt she was about to cry, “Usually I just try to stay away from being around other people when that happens.” The interviewer asked if anybody noticed her distress. Erin responded that, “I think sometimes just by looking at me my friends can pick up on it. I'll be at church and they'll say, ‘How you doing?’ and I'll smile and they'll say, ‘Oh, you're not doing well.’ And then there's the obvious, that there's tears in your eyes.”

If Erin seemed to live in a kind of exile circumscribed by her own suffering, other informants described depression as a state of the threat of exile. In a struggle to preserve their position within or the cohesion of a valued relational world, almost a quarter (11 of 46) of informants worked to hide depression. For many, this became a self-perpetuating bind. One man who described himself as deeply lonely talked to his daughter about his sadness, but he minimized its severity because he felt it would burden her to know of it. Naomi, a 54-year-old woman, suggested that depression left her alone in a cycle in which concealment was preferable to the magnification of conflict. She said that she hid her medications from her husband because she understood his anger to be the cause of her depression, and disclosing her treatment to him would anger him. Patricia, a 46-year-old woman similarly said, “I usually feel like crying,” because of “the man I live with, I guess,” who she described as “too jealous.” They had a “big fight” in a store because her husband accused her of walking away from him; Patricia became so upset that she cried uncontrollably. After another fight, “I just started crying and I just stayed like that for 2 or 3 days.” To avoid angering him, and because she did not know whether she wanted to leave him, “I keep [my sadness] to myself and I hate it because it builds up.” With a tone of desperation she added that, “I just shut up. I don't say nothing; I just keep my mouth shut. I just don't say nothing. I shut up.”

For these informants, the depression experience threated their inclusion within the family. For instance, Linda, a 45-year-old woman, described her depression as equivalent to her inability to contribute: “I always get kind of depressed because I can't go to work and I can't drive anymore.” She had recently been turned down for disability, and she was afraid to reveal this to her family. “I don't get no disability or social security; I don't get no kind of benefits, and it's really really tight. It's really bad.” Describing her situation as “stressful, depressed,” she began to cry, admitting that, “sometimes I feel like it's my fault because I can't bring any money in.” She hoped to increase the resources available to her family, and her depression deepened because she could not: “I can't help. I can't help pay the bills, and it just makes me feel bad.” Linda said that she rarely discussed her distress with her family but that they responded in a manner that seemed appropriate to her. “They say, well, Mom, just go get another job.”

In this, Linda shared her understanding of an effective treatment for depression with us: it would have restored her relational value. Like her family, she viewed productive re-engagement in the family's shared struggle as the appropriate solution to depression. As she said of her daughter's comment about returning to work, “What else could she say? I mean, I know she cares, but she can't do anything about it.” She spent her time pursuing a cure as she envisioned it: “thinking of ways that I can get money, earn money. I'll go look for a job, wherever would hire me. Or, I'm even thinking of selling my car that's sitting in the garage.” But short of being able to affect this solution, concealment remained a useful strategy for minimizing depression's threat of exile. Like Linda, Mabel, a 39-year old woman, said that she hid her bouts of crying from her husband because she believed her distress about their shared financial situation would trouble him and lead him to encourage her to find a job, which she had been unable to do.

Help-Seeking as Relational Threat

As these informants indicated, the exile of depression could be intensified in asking for, receiving, or accepting help for it. Another quarter of the sample (11 of 46) foregrounded this conflict. These informants described help-seeking as likely to make them less acceptable to others, to exacerbate conflict, or to cement their isolation within their immediate context. They weighed this cost against the actual and potential benefits of receiving help. For some, family members' ambivalence about treatment validated the negative relational consequences of help-seeking. Molly, for example, a 49-year-old woman, contended with the confusing attitudes of her family while understanding herself as benefiting from treatment. Her husband did not like that she took medications but simultaneously acknowledged that they helped her. Because her son did not want her to take it, she asked her doctor to lower her Prozac dose, despite continued symptoms. She hid her medications from her grandchildren and her friends. Molly's acquiescence to her family members' views suggest not only that she shared their ambivalence but also that she viewed her depression and its treatment as burdens borne by all of them.

Informants described several aspects of help-seeking as relational threat including acknowledging sad feelings to one's self and to others, accepting a diagnosis of depression, disclosing the diagnosis, and taking psychiatric medications. Diana, who screened positive for depression at almost all 6 follow up waves, told a poignant story of the relational costs of help-seeking. Diana felt “judged” by her doctors for not wanting to take an antidepressant. Yet it seemed to be Diana who judged herself, as she agonized over whether it was preferable to suffer with depression or with her own derogatory feelings that accompanied the acceptance of treatment. She described her family, her closest friend, and herself as opposed to medications. She felt, at the age of 38, that she was “too young” to be on psychiatric medications. Yet her depression was severe and debilitating. She “just couldn't” get up to do household chores and felt “like I have to force myself to go have coffee with a friend but not talk about any thing important.” Social outings required “physically forcing myself to get out of the house….just get myself showered and try to put some makeup on.” She felt like staying in bed but worked hard “to avoid sitting there because it just gets worse if I give into it.” In fact, “It feels like it would just swallow me whole.”

Diana acknowledged that this effortful trial of going “through the motions until I feel better” was a consequence of her inability to engage in treatment. As she said, “part of [the difficulty] goes back to, because I just can't make myself take medications.” Even though “I know there is medications that can help because I've used it before,” she felt the medication side effects were worse than the depression itself; or, admitting that this idea may not have been true in reality, she clarified that when she was taking medications, “that's just how I saw it.” In response, “I just have tried to do these other things [instead of taking medications] because even if it is a cycle, this just seems more natural, to just keep doing other things, what I believe I should be doing, as much as I can until this passes.” Her incongruous phrase, “even if it is a cycle,” suggests she may have recognized in her experience a version of the medical model of depression as a recurrent condition with characteristic symptoms. In spite or because of this recognition, she suspected there might be something unnatural about interfering with its course. Thus, her decision that it was “more natural” to live with debilitating depression without asking for help was one she admitted to the interviewer may be slightly suspect. But she chose to prioritize the solution for which she had sole responsibility, one she could pursue without risk of judgment: the fulfillment of her obligations to herself and others -- doing “what I believe I should be doing” – rather than medical treatment.

The interviewer asked whether she had sought help from clinicians. Laughing gently and with a tone of self-deprecation, she replied that,

I tried, really hard, one day during that week [in which she could not sleep]. I tried really hard to go to the doctor knowing that I probably really should. [Interviewer: but you didn't…] I left the house. I parked at the clinic, but I just couldn't make myself go in. That was just….[trailing off] I guess I just felt like there was too much, and I wouldn't even have known what to talk about. And they are health care providers, and I do have health issues, and I guess I could have started there, but I just became too overwhelmed and I just came back home.

Describing a few nights she took an over-the-counter sleeping pill, she said, “I guess I want to say that it did help,” but “I really have to work myself up just to even take one thing.” She reiterated that “Making myself do it [take the pill] was very difficult.” As the interviewer gently probed to understand her reluctance, Diana tried to end the speculation with a non-specific statement that “I just am really paranoid about these drugs.” But she continued, suggesting she avoided treatment to fend off a tainted identity that both she and those around her perceived. She said, “I kind of equate taking something….as being weak. Can't have anybody thinking I'm a weakling, I don't know [laughs].” In concluding that “I just don't think that pills are the answer to everything; in fact I think they're the answer to very little,” she suggested that working to prove to herself and to others that she was, in fact, a strong person felt like a form of effective treatment.

In advocating for the eschewal of treatment as a form of strength, Diana echoed Linda's perception that returning to work would treat her depression and Mabel's pride that she could hide her worries and thereby not exacerbate her husband's. Both would have endorsed Diana's efforts to do “what I believe I should be doing, as much as I can” while depressed. These women could show themselves to be especially strong in their ability to hide their distress or in their avoidance of treatment, and this form of suffering while demonstrating strength preserved their position with their relational space. In contrast, options for help-seeking seemed only to offer further debasement of their value to themselves and to others.

Detached in Treatment

Diana's statement that “I just don't think that pills are the answer to everything” alluded to another way in which help-seeking potentiated the relational fracture of depression. Our informants described medications as a particularly potent reminder of the debasement that accompanied help-seeking. To manage, many informants hid them. For instance, Ana, a 33-year-old woman, acknowledged that her antidepressant had been helpful, but she hid her medication from her 17-year-old daughter. Asked to explain, she said, “I just don't want her to feel that medicine is the easy way out,” and, more precisely, that, “I just don't want her to think that I'm looking for an easy way out of the situation.” She hid her medications to minimize her own shame, concluding that, “I guess I would say I'm a little bit embarrassed that I have to take the medication.” The problem with Ana's effective treatment was that it had not repaired depression's injury to her social self. For many informants, the shame, exile, or distance brought to the fore by the depression experience remained when symptoms were gone.

About a fifth of our informants (9 of 46) described themselves as closeted beneficiaries of treatment. Sadie, a 68-year-old woman who described the very positive impact of medications, individual therapy, and group therapy on her depression during her interview, did not talk to those closest to her about her diagnosis or her treatment. Janice, a 54-year-old woman said, with a defensive tone, that she did not know what her family thought about her medications because, “I don't go around to the world announcing that I'm taking it.” She preferred to think that her family members were unaware and uninvolved: “It's just not something that comes up. If it comes up it's not a problem.” Detached from her interpersonal experiences, depression – even in its remission -- still disconnected her from others. Treatment was a necessary chore pursued alone and acknowledged, if possible, not at all.

John, a 43-year-old man, echoed these attitudes, suspecting that others would see him negatively if they knew of his depression. John, who screened positive for depression 5 of 6 times during the 9 years of follow up, described his mood as “down in the dumps.” “I just stay in the house and I stay isolated.” Despite his positive screens, he felt that his medications, psychotherapy, and the collaboration between his mental health and primary care providers had had a positive impact on his life. Yet when the interviewer asked what he told friends, he said that, “My friends don't really know about it. One friend knows that I take the medications. He doesn't really comment about it.” His girlfriend, he said, knew about his psychiatric medications and expressed concerns that they may have long-term side effects. As the interviewer asked why his other friends did not know, John interrupted: “Oh, I don't tell people I take all these medications.” Referencing a recent surgery, he continued, “They know what I've been through…[and]… I'm sure anybody would guess that somebody who has been through surgery is going to take medications.” When the interviewer asked if he felt the need to hide the fact that he took antidepressants from anyone in particular, he replied, “Ummm…Not right now. I think when I, when I have to, probably, head back into work, get back into the working [world]… if I'm asked, I might have to say that I don't [take them].” He elaborated that, “I don't want to tell them that I've got ‘Brian Wilson Syndrome’ or something.” In other words, he believed that if others knew, he would be labeled, distanced, and devalued: an experience of Goffman's discreditable identity (Chaudoir, et al. 2013) and Steele's stereotype threat (Steele, et al. 2002).

A disease-oriented concept of depression seemed to facilitate John's treatment adherence. Like John, Teresa, a 35-year-old woman, believed others would think less of her if she were to take medications, and she accepted them and became engaged in her treatment only after she concluded she “wasn't crazy.” Teresa accomplished this by describing her emotional distress as akin to a physical condition. Nonetheless, she told people about her medications only if they asked, and she otherwise hid them from view. A few others also suggested like Teresa and John that a disease-oriented concept of depression, in which their difficulties were material and located within the brain, helped them to accept treatment. Yet their degree of activation in treatment seemed minimal. Susan, for instance, a 54-year old woman, told the interviewer that she has been recently feeling “nervous” but quickly clarified that this was “not depression,” or, “not in the way that you guys mean it.” As she worked to fend off the feeling of debasement she perceived in the label of depression, she drew on a disease-oriented concept to depersonalize and mystify her treatment: “I really don't like taking all the pills but if it keeps my brain at an even keel, then that's great, or my body, whatever it does.” As she pulled selectively on neurobiological metaphors to generate new notions of the self and its embodiment (Wilson 2004), Susan was able to deflect shame and open a space for self-care.

For John, Teresa, and others, the ability to literally and figuratively detach treatment and its implications for the self from the relational context made it possible to passively accept it. Other informants, though, for whom this detachment was also necessary and helpful, identified treatment's detachment from the relational space as precisely its lack. For them, treatment simultaneously worked and felt ineffective because, they said, a treatment that did not repair the relational fracture at the core of depression could not be effective: it was “mere covering over,” as one said. Ray, a 53-year-old man, for instance, attributed his depression to his powerlessness to change the fact that his children were abusing drugs. He felt better with treatment, he said, but “at the end of the day” he continued to suffer because medications did not secure his children from peril. Similarly, Mary, a 39-year-old woman, admitted that antidepressants kept her from wanting to stay in bed all day and having angry outbursts. But she described feeling no better, because what she really wanted was for her children's behavior to change. And, Carmen, a 28-year-old woman, described her antidepressant as a “bandaid.” For these informants, treatment addressed their symptoms, but failed to address the strained relationships and lost opportunities in which they located their malady. Like Diana and Ana, some hinted at a lingering suspicion that accepting this “bandaid” could keep them from finding a more fitting solution, one through which their relational context or their perception of their own relational value could be restored.

Conclusion

The Relational Self in Depression

In both their narrative richness and specificity, as well as in their thematic consistencies, these findings demonstrate the value of studying complex illness experiences in their lived contexts. This large interview sample provides a view of the implacably relational nature of the depression experience, including the impact of that experience on critical clinical and public health concerns like access to and engagement in depression treatment. In elaborating these arguments, we intend to make an ontological claim that depression is best conceptualized as existing in relational space, not as a pathology within the individual sufferer. Our informants described depression as a state of anxiety about the safety of one's intimate engagements and relational value. At stake in depression -- to a certain extent though with variable manifestations -- was a link to the immediate social context. The salience of their interdependencies and the exigencies of the relational space led informants to perceive depression as burdensome for not only the sufferer. Our informants' most common response to the experience of fracture of relational space was concealment: keeping all markers of depression (feelings, overt outbursts of emotion, diagnoses, clinician visits, medications, treatments) sequestered from relational life. Only two informants of 46 found ways to acknowledge depression and treatment, enroll others to support them in pursuing it, and share positive experiences of treatment.

These narratives suggest that the self in depression is experienced and constructed as interconnected with others. Our informants described exile, loss, and threat as vulnerabilities of their interdependent self. They perceived and described the self in depression by noting its inability to contribute to family needs or its debasement or devaluation in relation to valued others. This self and its depressive experience were inextricable from the local social space. Indeed, this relational space was often described as the site of pathology, such as when Ray and Mary discussed depression as a condition situated in their relationships with their children. Elizabeth Wilson proposes that, rather than positing a derangement in the executive brain, an apt biological metaphor for depression would characterize it as an impairment in “an organic capacity to connect” (Wilson 2006: 130). Our claim that an interdependent self suffers from depression extends Wilson's metaphor to take seriously that health and illness are always brought into existence in the company of others (Garro 2011; Mattingly 1998; Mol 2002).

We contrast this concept of depression within relational space with a disease-oriented concept of depression that locates pathology within an individual body. In a disease-oriented model of depression, symptoms are pathological attributes of the individual sufferer. Even symptoms that allude to the relational aspects of depression – such as guilt and low self-esteem – are attributed to the sufferer's perception of reality rather than to actual vulnerabilities of the self within that reality. It is the case that some informants used a disease-oriented concept of depression to put relational concerns to the side and engage in treatment (Schreiber and Hartrick 2002). But most informants voiced fundamental disagreements about the therapeutic practices forwarded by a disease-oriented concept, particularly their site of action, the mechanism, and their impact. Susan's idea that the medications keep her brain on an even keel and Teresa's insistence that she could take medications and not be crazy reflect ambivalent acquiescence to a disease model for instrumental reasons. The assumption of a disease-oriented model that symptoms are altered through action on the individual depends on a notion that the self is autonomous, self-interested, and capable of remaking experience. On the whole, our informants did not experience or construct themselves in quite that way.

Our data indicate that, for many individuals, a disease-oriented concept can exacerbate depression's relational burden, increasing concerns about one's safety in relation to cherished others. For Diana, Linda, Mabel and others, refusing treatment as a demonstration of strength was a viable and appealing strategy for preserving relational possibilities, even though they continued to be profoundly burdened by symptoms. For others, hiding or eliding depression and its treatment were partial solutions to the fact that asking for help, acknowledging feelings, or disclosing a diagnosis all posed some relational threat. As Westerbeek and Mutsaers (2008) characterize it, those who “dare to be open” about depression describe, “yielding to circumstances far beyond their control.” As crystallized in the image of Diana sitting paralyzed in the clinic parking lot, many informants grappled with the quandry that they could find help only by accepting a debased status. A model of depression in which the individual identifies symptoms, recognizes their pathology, locates them within a diseased organ of the body, and brings them to the clinic for attention sounds value-neutral, but was experienced by our informants as a way to confirm their inadequacy to those they depended upon and valued most. For some, even effective treatments that work on the self as if it were autonomous could feel ineffective.

In an interview-based study of older adults with depressive symptoms, Wilby and Barusch found that the most common attitude toward treatment was that it was something to be avoided. Like our informants, their participants voiced a heroic and unrealistic expectation “that they should be able to handle their problems alone.” (Wilby and Barusch 2012: 5) The case of Clara, a 46-year-old woman, illustrates the predicament that a disease-oriented model of depression can present to the sufferer. Clara screened positive for depression on 4 of 6 follow up surveys over 9 years of the study. During her 10-year interviews, she suffered from headaches that she attributed to a series of especially tense fights with her husband. One day, careful to avoid his discovery, she snuck out of the house to see her primary care clinician. She told the clinician about the fights and asked for help with her headaches. She reported that her clinician encouraged her to reframe her distress as manageable, and he recommended relaxation exercises and cognitive-behaviors techniques that might allow her to revise her feelings about her circumstances. From her point of view, this positioned her as responsible for her own distress and its relief. She said, “All he did was, he told me I had to learn to deal with stress. That if I decided I was going to stay with this man, then I had to learn to deal with this man.” His advice felt irrelevant because it focused on her own capacity to control herself: “He just told me to just go into a room and do a relax[ation] technique thing that I already do.” She repeated with incredulity, mocking her clinician, “‘Go into a room and relax.’” Clara understood her anxiety to be inextricable from her situation with her husband, thus she rejected a solution that located anxiety exclusively within herself, addressable (literally and figuratively) in isolation with a technique applied to the body. In the next interview, she said she had given up seeking help for her anxiety and had chosen not to talk to “anybody at all” about her distress. Disease-oriented therapeutic practices had exacerbated her distress and isolated her from support.

In reviewing the data, we were alarmed by informants' persistent efforts to hide their medications and by the pointed embarrassment they harbored about taking them, reflected in the statements that they are “an easy way out” or are only for “weaklings.” It may be that, for some informants, psychiatric medications constitute a specific nidus for the shame of depression. Carpenter-Song, in her study of children taking psychiatric medications for ADHD, argues that medication “take on significance as the site and source of difference” in her informants' construction of self. They interpreted their diagnoses as placing them at the extremes of normal behavior, making them, as one says, “just a little bit different.” But they were unable to deny their difference from others in the face of medication-taking (Carpenter-Song 2009: 265). The physicality of medications -- that one unambiguously takes them in or not, and repeatedly -- sets them starkly apart from both the abstract emotional states of depression and treatments that aim to make meaning of those emotions (Ecks 2013; van der Geest and Whyte 1989). However, on the whole our findings indicate that medications may not always present qualitatively different threats to one's sense of self than other forms of treatment. We would suggest that medications constitute the most concrete, recurring, and unavoidable reminder of the debasement of help-seeking via a disease-oriented concept of depression. This reworks Goffman's assertion (1963) that the “visual conspicuousness of the stigmatized attribute” (Chaudoir, et al. 2013: 75) can be a critical influence on the intrapersonal experience of illness. In this frame, medications may materialize and symbolize the unresolved relational fractures exposed in the disease-oriented treatment of depression.

The Moral and the Structural Picture of Stigma

In what ways might a notion of depression as an experience of the relational self inform concepts of stigma? First, our findings challenge the meaning of closed-ended assessments, like those we used, that construct stigma as a process of labeling, distancing, and status loss. In our sample, structured assessments of stigma and disclosure at year 9 did not correspond to interview content at year 10. Though we sampled a group that reported high or low rates of disclosure in year 9, in interviews in year 10 the majority described employing some degree of concealment and non-disclosure, from discussing treatment only obliquely to actively hiding medications and mental health visits. In fact, the two informants who described supportive contexts in their interviews did not come from the low stigma/high disclosure subsample. In interviews, neither mentioned stigmatizing attitudes from family members or friends despite that one came from the high stigma subsample. Moreover, both described a high degree of openness about depression with their family even though both were drawn from the low disclosure subsample. These findings upended our original assumption that we could use a structured assessment of stigma and disclosure to mark supportive versus unsupportive social contexts.

In addition, phenomenological and grounded approaches to the interview data showed that processes of labeling, distancing and shaming were almost ubiquitous, and those informants that persisted in treatment did so not because of a lack of exposure to them but because of strategies they developed to ignore them. Agentic resistance of the stigma of mental illness, such as that described by Thoits (2011), was rare. A small number of informants (e.g., John, Teresa) worked to deflect stigmatizing labels by claiming they were “not me,” but we did not find clear examples of active challenge of stigmatizing views. It should be noted, however, that interviews did not include questions designed to elicit resistance to stigma.

Our findings concur with those of Pescosolido and colleagues that the “backbone” of stigma appears in intimate settings (Pescosolido 2013; Pescosolido, et al. 2013), what we have called the relational space. The shame and devaluation of depression emerges where it calls intimate relationality into question. Scambler (2009) points out that attention to the relationships in which stigma can circulate is consistent with Goffman's original theoretical perspective, which explored stigma as an extreme demonstration of the rules that structure everyday interactions. Goffman used stigma to illustrate how an individual must “maintain face” (i.e., avoid being discredited) in everyday interactions through a consistent and affirming presentation of self (Yang and Kleinman 2008). Goffman's notion that an attribute is not discrediting in itself but can become so in specific social interactions mirrors Diana and John's experience that forceful censoring of depression, not a direct reckoning with it, allowed for the maintenance of face with those who meant most to them. The moral turn in stigma theory correctly asserts that these specific relational circumstances matter to the people with a condition and therefore should matter to clinicians, researchers, and others with a stake in addressing distress.

Simultaneously, though, our informants suggest that the moral turn in stigma theory may give short shrift to a core aspect of stigma: the social, economic, cultural, political, or interpersonal power it requires to operate (Castro and Farmer 2005; Link and Phelan 2001). Even within the family, maneuvers of differentiation and devaluation constitute and reinforce the existing order and shape the possibilities for those with depression, such as for Linda, who perceives financial participation to be a necessary avenue for full inclusion within her family. But more thoroughly, these narratives of illness experience make transparent the exercise of biomedical power through a disease-oriented concept of depression. In their reluctance to adopt an idea of depression as located within the individual body and in the ambivalence with which they adopt treatment as a project of the self, informants indicate that the institutions of biomedicine are neither value neutral nor passive (Bromley 2012). Indeed, many of our informants posited biomedicine as comprised of powerful actors with the means of discrediting vulnerable others (Thoits 2005; Thoits 2011). Thus, Susan had to bat away the interviewer with the insistence that she was not depressed “in the way that you guys mean it.” Clara scoffed at her doctor's order that she “go into a room and relax,” but clearly the advice delivered a potent sting. These findings are consistent with studies that distinguish the uniquely stigmatizing impact of specific everyday interactions with biomedicine, such as the particular shame associated with clinic visits, medication-taking, and hospitalization identified by Lee and colleagues among individuals with schizophrenia in Hong Kong (Lee, et al. 2006). Biomedical actors are one type of social actor with the authority to adjudicate and impose differential status. Informants, “[p]recisely because they are subjected to an overwhelmingly powerful symbolic apparatus,” had few avenues for resistance (Parker and Aggleton 2003: 18).

A comprehensive concept of stigma would need to account for biomedicine's particular potency in elaborating and enforcing notions of normality, in determining blame, and in communicating an individual's worth. Instead, as Scambler notes, many stigma reduction campaigns still assume biomedical concepts to be value-neutral. They adopt a trope of empowerment that emphasizes personal responsibility and therapeutic engagement (Scambler 2009). Such efforts may merely repackage biomedical power and thereby exacerbate the same relations of power that stigmatize (Hansen, et al. 2014). Our findings suggest that what is needed is a model of stigma that accounts for the importance of individual meaning-making while recognizing the critical dynamics of power, both its operation in intimate contexts and through biomedical practices. An attention to stigma as a reflection of “the complexities of motive, of moral deliberation and moral creativity, especially as elements of ordinary life” (Mattingly 2012b: 177) should also include a focus on everyday encounters with structural power.

Predicaments in Primary Care Treatment

These data are limited by the absence of interviews from family members and the lack of an ethnographic portrait of informants' everyday contexts. Both types of data might have allowed us to more fully demonstrate the nature of the relational conflict precipitated by depression and the possibilities of the relational space in which depression is experienced. But a unique strength of the dataset is that informants were identified at the point of the most common site of entry into depression treatment, the primary care clinic, and, after 9 years of survey follow up, were interviewed multiple times about their depression experience. These features suggest the data provide a highly unusual look at the commonplace predicaments facing those who live with depression. Despite the decade that had elapsed since identification of the sample, many were still in an uneasy relationship with the concept of depression and its treatment. Like Diana sitting in her parked car, many informants seemed still to struggle literally or figuratively to enter the clinic, with the primary care site still functioning as the gatekeeper. These were challenges we assumed would be surmounted at screening and in early study phases. In addition to underscoring the pervading centrality of the primary care clinic in the medicalization and pharmaceuticalization of distress (Bell and Figert 2012), the findings highlight the importance of longitudinal research on the lived experience of depression.

Because we noticed consistent themes that cross-cut ethnic and other subgroups, we did not explore relationships between informants' study arm (i.e., usual care versus quality improvement intervention) and depression experience at 10-year follow-up. Our informants received a range of treatments, and some received no treatment despite a positive screen. However, in a separate analysis, participants from the PIC study treated by primary care clinicians receiving enhanced resources for the delivery of therapy showed reduced stigma concerns compared to other participants; while those participants treated by clinicians receiving enhanced resources to support medication management showed increased stigma concerns compared to others (Klap, et al. 2009). As do informants' narratives, these analyses suggest that primary care pharmaceutical treatment of depression may increase stigma. Effectively incorporating antidepressant treatment into primary care may require a clinically sophisticated approach that explores and lessens self stigma and that fosters possibilities for resisting social stigma. Moreover, as does a high-quality psychotherapy, this approach need not aim to change life circumstances but to moderate interpersonal relationships by helping the individual not blame him or herself for their limitations.

Strategies that can support patients in articulating their depression experience as it intercalates their relational world may lessen the burden of help-seeking, as they approximate the treatment approach to the nature of the depression experience (Bussing, et al. 2006). Our findings endorse a set of validated approaches to primary care depression treatment that provide avenues for the patient to consider herself to be a competent participant in a interpersonal treatment process. These approaches include developing a shared notion of the problem, activating the patient in his or her care, collaborating care between specialists and primary care clinicians, instituting a personalized treatment plan, incorporating psychotherapy, and pursuing proactive follow up (Callahan 2006; Murphy, et al. 2014; Rubenstein 2006; Sinclair, et al. 2014; Unutzer, et al. 2002). Enrolling family members or loved ones to support treatment can be seen as a core goal in depression treatment, yet the risks of this enrollment must be taken seriously. Finally, with an awareness of their own structural position, primary care clinicians can recognize their clinical interactions, their diagnostic practices, and their therapeutic interventions as forms of structural power, capable of shaping patients' views of their moral and social value. Our data indicate that these strategies may both reduce stigma and improve the effectiveness of primary care depression treatment.