Stigmatization of people with alcohol use disorders: An updated systematic review of population studies

Background: We summarize research on the public stigmatization of persons with alcohol use disorder (AUD) We conducted a systematic for studies on the We identified 20,561 records, of which 24 met the inclusion criteria, reporting results from 16 unique studies conducted in 9 different countries. Compared to substance- unrelated mental disorders, persons with AUD were generally less likely to mentally ill, while they were perceived as being more dangerous and responsible for their condition. Further, the public desire for social distance was consistently higher for people with AUD. We found no consistent differences in the public stigma toward persons with AUD in comparison with other substance use disorders.


| INTRODUC TI ON
Stigmatization of people with mental disorders is a key contributor to healthcare inequality, as highlighted in the report of the World Health Organization's Special Initiative for Mental Health -2023World Health Organization, 2019a). Consequently, understanding the public stigmatization of people with mental disorders or substance use disorder (SUD) is essential to reduce or eliminate negative consequences for those affected and warrants particular attention in research. Yet, most research on the public's mental health stigma has employed descriptive rather than systematic approaches or conceptual models which are needed for an in-depth understanding (Corrigan et al., 2017). In this systematic review, we summarize population-based research on the public stigmatization of people with AUD compared to the stigma of other conditions (i.e., mental disorders and other SUD).
Several established models of stigma exist (Corrigan et al., 2017;Link & Phelan, 2013;Stangl et al., 2019). Based on this work, we define stigma for the purpose of this review as a process in which people are firstly labeled and thereby assigned to an out-group, secondly subjected to stereotypes and prejudices, and thirdly exposed to discrimination and social distance. This dynamic process implies a nonlinear and interactive stigma process, in which each component can mutually reinforce the others. It is noteworthy that although the stigma process commences with labeling and stereotyping, this may not necessarily be perceptible to the stigmatized person. Latter components such as discrimination, however, are usually tangible for those affected and are more likely to be perceived as harmful.
The components labeling, stereotyping and prejudice, and discrimination will constitute the analytical framework of the current review (see Figure 1) and are outlined in more detail hereafter.
Labeling is the categorization of a person using a label such as a diagnosis, which on the one hand is mandatory to provide treatment, but on the other hand assigns the person to a group, which in turn facilitates stereotyping (Link & Phelan, 2013). This transition is fostered when a label already implies a negative connotation, such as "alcoholic," which in turn is highly culturally shaped. The second component includes stereotypes and prejudices, closely related concepts that are largely linked to negative, harmful, and disrespectful beliefs about a person or group (Corrigan et al., 2017;Link & Phelan, 2013;Stangl et al., 2019). While stereotypes are imminent and culturally conditioned thoughts, such as believing someone to be unpredictable or dangerous, prejudices are the approval of stereotypes, which may lead to negative emotions and evaluations such as fear or anger and finally to discrimination (Corrigan et al., 2017). Blame also belongs to this component and is covered in this review by its 2 facets: onset (e.g., someone is perceived to be responsible for developing an alcohol use problem) and offset responsibility (e.g., someone is perceived to be responsible for not recovering). The final component discrimination can manifest in the desire for social distance and structural discrimination through, for example, social exclusion and loss of status through workplace discrimination or unemployment (Corrigan et al., 2017;Link & Phelan, 2013;Stangl et al., 2019).
This work is an update of a previous systematic review , in which attitudes toward people with F I G U R E 1 Analytical framework of the stigma process developed for this systematic review [Color figure can be viewed at wileyonlinelibrary.com] AUD were compared to attitudes toward people with other mental disorders or other SUD in 17 different studies. The authors of this review concluded that "people suffering from alcohol dependence (and from other addictions) are particularly severely stigmatized" (p. 109). High levels of blame projected onto people with AUD may further reflect a specific moral component of addiction stigma. They argued that the stigma of SUD can be understood as a reaction to unacceptable behavior-that is, as a function of norm enforcement-to "keep people in" (Phelan et al. 2008). In this updated systematic review, we summarize studies published since October 2010 investigating the stigma of AUD in comparison with the stigma of other mental disorders. We present and discuss our findings in the context of the stigma process and highlight research studies that have investigated: (i) changes in the stigma process since 1990 or (ii) the interrelation of the different aspects of the stigma process in AUD. In order to provide an update of the 2011 systematic review by   Schomerus and colleagues, we conducted a literature search of all   representative population-based studies on public attitudes toward   persons with AUD and toward persons with any other mental health   condition, published between October 1 , 2010, andDecember 20, 2020. The protocol of this systematic review was published on PROSPERO (registration number: CRD42020173054) based on the PRISMA guidelines (Appendix S1). Potential risk of bias was determined based on an adapted version of the ROBINS-I tool (Sterne et al., 2016). The tool was adapted to evaluate the bias of cross-sectional observational studies, rather than randomized controlled trials (available upon request).

| RE SULTS
A total of 20,570 records were identified of which 24 research articles reporting data collected in 16 unique studies have been included in the final review (see Figure 2 and and 2018 (study #11), microdata were publicly available, and for a German survey conducted in 2011 (study #3), results were provided upon request by the authors. These additional findings provide insights on the extent of public stigma toward persons with AUD and other mental health conditions, which were not published elsewhere but are included in this review.
In the following, key findings relevant to this systematic review are descriptively summarized (for details on the study's methodology, see Appendix S5). Unless otherwise stated, comparisons of stigma aspects between mental health conditions were usually not subject to inferential statistical analyses in the original studies.

| Labeling
Six studies investigated whether AUD was considered a disease or a mental illness, with only 5 studies presenting their findings (no information was available for study #12). In the Finnish study by Tikkinen et al., (2012Tikkinen et al., ( , 2019, more than 1500 adults were asked whether they considered a wide range of mental disorders or other conditions to be a disease (label stimulus "alcoholism"). While more than three-quarters indicated that schizophrenia and autism were a disease and more than half of the respondents considered depression, eating disorders, panic disorder, and attention deficit hyperactivity disorder (ADHD) to be diseases, the results were less clear for AUD, drug addiction, social anxiety disorder, generalized anxiety disorder, and insomnia. Almost 50% agreed, at least to some extent,

| Stereotypes and prejudice
Nine studies examined stereotypes concerning dangerousness or unpredictability of AUD and other mental health conditions. Among vignette studies examining perceived dangerousness (7 studies), persons with AUD were consistently considered to be more dan-  were the highest, followed by schizophrenia (M = 2.0, 95% CI: 2.0-2.0) and depression (M = 1.9, 95% CI: 1.9-2.0).
We identified 8 studies asking for the onset responsibility on a disease. Compared to both depression (9.6%) and schizophrenia mean over all substances: M = 3.2, SD = 0.6). We also considered causal beliefs as 1 aspect of onset responsibility when "bad character" was included as 1 option to explain the origin for a disease (compared to options like "chemistry in the brain" or "stress" or (2018, Singapore, study #15) conducted a factor analysis with different causal beliefs and identified a "personality" factor determined by high factor loads on the items "weak character" and "being a nervous person." More than 80% of respondents endorsed causal attribution to this "personality" factor in alcohol abuse.
Four studies examined aspects related to the offset responsibility of various mental health conditions. With regard to the individuals' likelihood to "get better if they wanted to" (Subramaniam et al., 2017, Singapore, study #15), the highest approval was observed in vignettes describing a person with alcohol abuse (94.5%), followed by depression (92.1%), OCD (90.4%), schizophrenia (88.0%), and dementia (81.3%). Similarly, offset responsibility was rated higher for persons with AUD (35.3%) than for those with depression (24.8%) or schizophrenia (20.7%) in a German sample (see Appendix S7, study #3). In the 2 other studies, both using disease labels, the responsibility to recover in people with AUD compared to other addictions was compared. In the first case (Blomqvist, 2012, Sweden, study #4), respondents were asked about the person's likelihood of self-change with and without treatment. On average, respondents indicated a similar probability for alcohol misusers to recover without treatment compared to those with cannabis addiction and a lower probability compared to tobacco addiction. The likelihood that people with a dependence on hard drugs would recover was considered to be the lowest. Finally, the individual responsibility to recover was investi- We found only 2 studies examining negative emotions toward persons with AUD and other mental health conditions that were published since the 2011 review. The first study included feelings of fear (e.g., "the person makes me feel insecure" or "the person scares me") and anger (e.g., "I feel annoyed by the person" or "I am amused by something like that") toward people with AUD, schizophrenia, and depression (Angermeyer et al., 2013, Germany, study #3

| Discrimination
Twelve studies included questions on the desire for social distance themselves from persons with AUD and other mental health con- An overview of key findings presented above with regard to AUD, depression, schizophrenia, drug, and tobacco addiction is provided in Table 2.

| Changes in the stigma in AUD since 1990
Three articles of those identified additionally assessed changes in various aspects of the stigma process over time. Pescosolido et al., (2019) analyzed and compared data from 3 waves of the General Social Survey covering a time period over more than twenty years

| Cross-sectional testing of stigma framework assumptions
In addition to stigmatizing attitudes toward people with AUD and other mental disorders, we were also interested in reviewing evidence concerning the relationship between the different components of the stigma process ( Figure 1). We report findings of analyses that are grounded on the assumption that certain lowerlevel components, such as labeling or stereotypes, constitute the foundation for developing higher-order stigma components, such as discriminatory attitudes or social distance. We identified 3 articles, based on the same German general population sample (study #3), which examined cross-sectional associations of different stigma components. All analyses used vignettes as stimulus and adjusted for sociodemographics.
In the first 2 studies, the researchers examined whether labeling and stereotypes (i.e., beliefs regarding onset responsibility) predicted discrimination in terms of social distance. They found that respondents who identified AUD as a mental illness  and who attributed AUD onset to chemical imbalances in the brain or to a brain disease (Speerforck et al., 2014) reported a significant lower desire for social distancing. In a third study, Schomerus et al., (2014a) performed path analyses to study how stereotypes and prejudice (including beliefs regarding on-and offset responsibility and perceived dangerousness) relate to the desire for social distance. According to their findings, beliefs in biogenetic causes for AUD were directly associated with lower desire for social distance. This relationship was in part mediated by opposing associations with blame and perceived dangerousness: Respondents who endorsed biogenetic causal beliefs were less likely to blame people with AUD for their condition, which was related to lower desire for social distance. However, endorsement of biogenetic causal beliefs was also associated with increased perceived dangerousness of people with AUD, which in turn was related to higher desire for social distance. Indirect effects of causal beliefs on discrimination were found for the associations between beliefs in childhood adversities or current stress and social distance. Beliefs in childhood adversities as causes for AUD were linked to increased perceptions of dangerousness of people with AUD and, in turn, higher desire for social distance. On the other hand, beliefs in current stress as cause for AUD was associated with less blame and less perceived dangerousness, which was linked to lower desire for social distance.

| DISCUSS ION
Our Note: Greater than and less than signs indicate considerable difference, based on either statistical significance or a descriptive criterion of at least 10 percentage point difference.
Abbreviations: AUD, alcohol use disorder; D, depression; S, schizophrenia; SUD, substance use disorder; T, tobacco addiction. a Negative stereotypes including unpredictability and dangerousness.; b Being unpredictable.; c Getting into trouble with law.; d Only studies that asked for blame or the causes "bad character" or "moral weakness." other mental disorders. The synthesis of findings revealed that stigmatizing beliefs and behaviors toward people with AUD were pervasive in the general population and usually more pronounced than toward persons with depression or schizophrenia. More specifically, people with AUD tend to be perceived as more dangerous and more responsible for their condition, as well as being faced with a greater desire for social distance and a higher degree of acceptance of structural discrimination than people with substance-unrelated disorders.
However, we found no clear evidence for substantial differences in these stigma aspects between AUD and other SUD or indications for changes in stigmatizing beliefs or behaviors toward persons with AUD since 1990 in the United States or West Germany. Thus, AUD remains among the most stigmatized mental disorders, corroborating findings of the 2011 review by . Further, the sociocognitive processes underpinning the stigmatization of people with AUD appears to follow distinctly different patterns as compared to depression and schizophrenia Schomerus et al., 2014;Speerforck et al., 2014).
Before we discuss our findings, we would like to outline the limitations of this systematic review. First, the risk of bias assessment identified a serious risk in eleven articles (see Appendix S4), of which the majority did not provide information on missing values, while 3 articles failed to report response rates. Due to these shortcomings, a final evaluation of potential bias due to missing data or nonresponses could not be undertaken for all publications. In the remaining articles, 6 were found to have a low and 6 to have a medium risk of bias.
Second, the interrater reliability in abstract and full-text screenings was low, which was counteracted by discussing ambiguous decisions with the team. Additionally, reference lists of other systematic reviews in the field were screened and experts were consulted for relevant additional literature resulting in 9 additional records included (see Figure 2). We observed differences in stigma aspects between studies that used labels or vignettes to describe persons with AUD as stimuli. In vignette studies in which a person fulfilling the criteria of a mental disease was described, AUD seemed to be more likely to be considered a mental illness by viewers (60% to 74% in studies #10, #11, #15; Chong et al., 2016;DePierre et al., 2014;Perry et al., 2020), while in the study of Tikkinen, which used the label "alcoholism", only half of the sample recognized AUD as a disease (Tikkinen et al., 2012(Tikkinen et al., , 2019. This underlines the potential impact of applying either a labeling or a vignette-based approach to capture public stigma. While certain labels of AUD could reinforce stigmatizing attitudes, labeling AUD as a disease could have a beneficial impact, potentially reducing blame and increasing prosocial reactions. However, in the German study by Schomerus et al., (2013)  States. While we found no indication for changes in perceived dangerousness and approval for structural discrimination across the US studies, findings from Germany suggested that public stigma toward persons with AUD has changed somewhat for the worse. We identified 2 other articles using the same data in which the results for AUD were not compared with other mental disorders, so they were not part of this systematic review. As compared to 1990 survey, no sufficient evidence was found pointing to a change in the stigma of AUD since 2011 (Schomerus et al., 2014b;Schomerus et al., 2014).
Taken together, the results do not point to substantial changes in the stigma of AUD since 1990. However, limitations such as the crosssectional nature of the data and the regional restrictions need to be taken into account and demand further studies to detect possible changes over time.
Very few studies reported findings that could be used to (in) validate the proposed analytical framework. We identified 3 studies from the same cross-sectional sample of German respondents reporting associations of stigma components. While findings should be interpreted cautiously and neither causality nor the direction of a possible causal link of the stigma components is known at this stage, an important observation can be made: The association between stigma components is not as clear and unambiguous as specified in the stigma process framework. The findings suggest that the development of stigma toward people with AUD is complex. For example, labeling AUD as mental illness was associated with a lower likelihood of expressing anger, but with an increased likelihood of expressing fear toward those with an AUD . This could indicate that 2 contrasting conceptualizations shape our responses to AUD: If regarded as a stable and trait-like condition, related to assumptions on "bad character," blame and feelings of anger might be less pronounced but fear and social exclusion nevertheless high.
Conversely, if regarded as a "bad behavior"-that is, a state that needs to be overcome-moral judgments and blame of people with AUD could be harsher, possibly leading to more discrimination and social exclusion. Such seemingly contradictory patterns highlight the particular nature of SUD, being regarded as both a behavior and an illness. So far, both conceptualizations seem to work to the disadvantage of persons with AUD. It is unclear how interventions need to frame SUD in order to minimize discrimination, and this tension between behavior and disease possibly cannot be resolved.
However, continuum models of substance use problems might offer a third way to conceptualize SUD in order to make SUDs of differing severity more relatable (Morris et al., 2020).

| CON CLUS ION
People with AUD were shown to experience a high level of stigmatization, unchanged over the past decades (e.g., no changes since the last systematic review of . The stigmatization seems to be somewhat unique and higher than for other mental disorders, with the exception of other SUDs. However, while the situation appears to be unique for AUD, the attention to this fact is minimal, in terms of both scientific and clinical implications.
As for scientific implications, most studies are based on simple cross-sectional surveys, which are limited in tracking changes over time, as sampling frames differ, with overall increases in nonresponse (Rehm et al., 2021) and survey methodology not being standardized or tested with respect to comparability. Given the methodological limitations identified in the reviewed studies, a homogenization of the methods in studies on the public stigma seems warranted, for example, via standardized protocols. As demonstrated in other fields (e.g., Guidelines for Accurate and Transparent Health Estimates Reporting GATHER; Stevens et al 2016), adherence to research standards is key for a comparative assessment across locations and time, which is required for theoretical advances and evaluation of stigma campaigns. Yet, even if the methodology was comparable, there is need for improved research design-which would allow for stronger conclusions to be drawn with respect to causality, such as experimental designs or truly longitudinal follow-ups of larger cohorts. Additionally, only a couple of studies exist so far that explore links between different stigma aspects.
With respect to clinical interventions, the last systematic review found interventions for self-stigmatization, stigmatization in the general population, and structural stigmatization in health (Livingston et al., 2012). However, effect sizes were small, and overall, the authors indicated that several interventions demonstrated promise for achieving meaningful improvements in stigma related to SUDs. Obviously, for changing the long-lasting stigmatization of people with SUD, this may not be enough.