Alcohol stigma and persistence of alcohol and other psychiatric disorders: A modified labeling theory approach

Abstract

Background

We sought to apply modified labeling theory in a cross-sectional study of alcohol use disorder (AUD) to investigate the mechanisms through which perceived alcohol stigma (PAS) may lead to the persistence of AUD and risk of psychiatric disorder.

Methods

We conducted structural equation modeling (SEM) including moderated mediation analyses of two waves (W1 and W2) of data from the National Epidemiologic Survey on Alcohol and Related Conditions. We analyzed validated measures of PAS, perceived social support, social network involvement, and psychiatric disorders among (n = 3608) adults with two or more DSM-5 AUD symptoms in the first two of the three years between the W1 and W2 survey. Cross-sectional analyses were conducted owing to the assessment of PAS only at W2.

Results

Per mediation analyses, lower levels of perceived social support explained the association of PAS with past-year AUD and past-year internalizing psychiatric disorder at W2. The size of the mediated relationship was significantly larger for those classified as labeled (i.e., alcoholic) per their prior alcohol treatment or perceived need (n = 938) as compared to unlabeled (n = 2634), confirming a hypothesis of moderated mediation. Unexpectedly, mediation was also present for unlabeled individuals.

Conclusions

Lower levels of social support may be an important intermediate outcome of alcohol stigma. Longitudinal data are needed to establish the temporal precedence of PAS and its hypothesized intermediate and distal outcomes. Research is needed to evaluate direct measures of labeling that could replace proxy measures (e.g., prior treatment status) commonly employed in studies of the stigma of psychiatric disorders.

Introduction

Alcohol use disorders (AUDs) are perhaps one of the most stigmatized medical or psychiatric conditions (Schomerus et al., 2010). Over half of the general public attributes the cause of AUDs to one's “own bad character” (Link et al., 1999) or believes that individuals are to blame for their illness (Crisp et al., 2000). Perceived stigma, which develops during socialization, is defined as individuals’ awareness of the discrimination and devaluation directed toward those with conditions that are viewed unfavorably (Link, 1987). For persons who acquire stigmatized conditions, perceived stigma becomes personally relevant, which may provoke fear of being rejected by others (Link, 1987). Among people affected by substance use disorders, perceived stigma is associated with a number of adverse outcomes that complicate recovery, including poorer mental health functioning (Smith et al., 2010), higher depression scores (Luoma et al., 2010), lower rates of treatment utilization (Keyes et al., 2010), lower quality of life (Luoma et al., 2007) and poorer physical health (Ahern et al., 2007). Scholars maintain that stigma is detrimental to achieving and sustaining recovery from addiction (Laudet, 2008, White, 2007, White, 2009).

Modified labeling theory (MLT) elucidates the mechanisms via which stigma leads to adverse consequences for those affected by psychiatric conditions (Link, 1987, Link et al., 1989). A major proposition of MLT is that the consequences of perceived stigma for people with psychiatric conditions are dependent on labeling, such that perceived stigma has personal relevance to those who carry a stigmatized label (e.g., alcoholic). Individuals may be labeled during various social exchanges, and in particular, MLT describes that labeling may occur if and when one receives treatment through the assignment of a formal diagnosis (Link, 1987).

According to MLT, to avoid further stigmatization, labeled individuals may employ specific coping orientations, such as maintaining secrecy about a psychiatric condition or avoiding potentially uncomfortable or threatening social interactions (Link et al., 1991). Secrecy and avoidance are thought to damage social ties and result in other negative outcomes despite their beneficial appearance (Link et al., 1991). Consistent with this notion, meta-analyses have identified a reliable association between perceived stigma and weakened social support or integration (Livingston and Boyd, 2010). Employing ineffective coping orientations has been linked to diminished self-esteem, self-efficacy, general well being, and job market participation and earnings (Link et al., 1987, Link et al., 1989, Link et al., 1997, Wahl, 1999, Wright et al., 2000), and is thought to ultimately increase individuals’ likelihood of relapse and exacerbation of psychiatric conditions (Link et al., 1989). While MLT focuses on social and individual level exchanges, we note that stigma can also result in discrimination at the institutional or structural level (Deacon, 2006, Link and Phelan, 2001, Williams et al., 2012).

While intervention research has begun to focus on preventing or ameliorating the outcomes of addiction stigma (Livingston et al., 2012), we are aware of few studies in the alcohol literature that have investigated mechanisms through which perceived alcohol stigma (PAS) may lead to adverse consequences. A recent study found a negative association between PAS and perceived social support among individuals with AUDs in the United States general population, which was stronger among individuals classified as labeled as compared to unlabeled (Glass et al., 2013). In contrast, Luoma et al. (2010) did not find an association between perceived substance use stigma (drug or alcohol stigma) and social support in an addictions treatment sample. However, their study did find a positive association between perceived stigma and self-concealment. Rather than focusing on coping strategies, others have analyzed the internalization of PAS. Schomerus et al.’s et al. (2011) analysis of an alcohol detoxification sample found positive correlations between PAS (which they deem “stereotype awareness”), self-stereotyping based on one's AUD, and self-esteem decrement. In line with these findings, an association between perceived substance use stigma (i.e., alcohol or drug) and internalized stigma has been found in other addiction treatment samples (Luoma et al., 2007, Luoma et al., 2010).

While these studies offer insight into intermediate outcomes, we are aware of no alcohol research formally investigating mediators of hypothesized stigma outcomes. While higher PAS has been linked to lower levels of social support (Glass et al., 2013), only alcohol research outside of the stigma literature has linked social network measures such as social support and related constructs to negative outcomes. For example, a smaller social network and lack of social relationships or social support is known to be a risk factor for increased alcohol consumption (Pressman et al., 2005) and depressive symptoms (Booth et al., 1992). Conversely, alcohol-abstinent social networks and treatment-supportive relationships predict alcohol dependence recovery (Hunter-Reel et al., 2010). To summarize, while stigma, social relationships, and outcomes have been linked in separate lines of alcohol research, formal tests of mediation are warranted to better understand the mechanisms of alcohol stigma that may worsen psychiatric outcomes.

An ideal data source to test an application of MLT would provide population-based data on labeled and unlabeled persons with AUDs, and the requisite measures of PAS, social network variables, and psychiatric outcomes (see Fig. 1). Such data are available within the National Epidemiologic Survey of Alcohol and Related Conditions (NESARC), a representative survey of the United States general population. While NESARC provides longitudinal assessments, several key variables (e.g., PAS) were only assessed at the follow-up interview, which precluded our ability to establish temporal precedence. Hence, we conducted cross-sectional analyses of NESARC using structural equation modeling (SEM) to test two basic hypotheses of MLT: (1) the relationship of PAS with past-year AUD and past-year internalizing psychiatric disorders would be positive and mediated by individuals’ social network involvement and perceived social support, and (2) the mediated relationship would exist for labeled, but not unlabeled persons. While MLT assumes perceived stigma is inconsequential for unlabeled individuals, we deemed it worthwhile to evaluate this empirically for alcohol stigma.