5

Research Strategies

This chapter focuses on components of a research strategy that are essential to the design and evaluation of stigma reduction interventions and approaches: (1) formative research to assist in developing interventions and tailoring them for target audiences, (2) intervention research to assess implementation and outcomes of the specific interventions, and (3) monitoring trends over time in attitudes, beliefs, knowledge, and behaviors toward people with behavioral disorders as manifested at structural, public, and individual levels. The chapter concludes with suggested areas for future research and evaluation.

FORMATIVE RESEARCH

Launching a large-scale stigma change strategy can imbue an organization with a sense of urgency and excitement. Yet, charging forward hastily can waste resources and produce disappointing results. For a campaign to be effective and relevant to stakeholders, interventions must be well-designed and correctly targeted based on prior formative research and ongoing evaluation. Research and interventions must work in tandem and learn from one another. Formative research may also reveal new features of stigma, generate research questions, and contribute to the development of evidence-based interventions. In addition, researchers should examine naturally occurring phenomena (e.g., media events, and policy implementations) and their effect on levels of stigma as insights

from these naturally occurring phenomena could inform future research, interventions, and policy initiatives.

Formative research draws from a mix of scientific disciplines, including psychology, anthropology, and sociology, and is used to design campaigns that are geographically and culturally appropriate (Gittelsohn et al., 2006). Behavior is shaped by a range of social, psychological, and structural factors making behavior change difficult to achieve. Through formative research, strategy designers define and assess the characteristics of the target audience relevant to the behavioral health issues of interest (Gittelsohn et al., 2006). The ultimate goal of formative research in this area is to identify factors, including motivating factors that can increase the effectiveness of behavioral interventions.

Community-Based Participatory Research

Beyond conducting an assessment of the potential audience, formative research is a way to facilitate relationships between the researchers and their intended audience. Incorporating formative research in the design phase of a multicomponent national strategy can ensure that the strategy is targeted to the most appropriate populations or subgroups; and that the voices of stakeholders, particularly those with lived experience, are included in the planning and evaluation of the interventions. Formative research methods that engage communities in developing initiatives include community-based participatory research, empowerment evaluation, and participatory or community action research (Ahmed and Palermo, 2010).

Participatory approaches involve and actively engage critical stakeholders (e. g., peer experts, family members, advocates, health care practitioners, provider organizations, employers, policy makers) at every stage of the research process. Stakeholders are included as active partners in understanding the problem, describing possible approaches and interventions, describing the theory of change behind interventions, identifying methods and measures to test the approach, collecting and analyzing the data that emerge from the design, and making sense of the findings (Corrigan and Shapiro, 2010). These methods are well suited to stigma change efforts because they can provide input on the subtle and dynamic aspects of stigma and discrimination that must be translated into specific interventions and messages to increase their relevance to both those who stigmatize and those who experience stigma (Corrigan and Shapiro, 2010).

Community participatory methods are useful as a component of formative research in understanding the perspectives of specific target audiences and in determining the best mechanisms or platforms to reach them. For example, attention deficit hyperactivity disorder (ADHD) is

among the most common childhood mental illnesses and evidence-based psychosocial and pharmacological treatments are available. But white children are twice more likely to be assessed, diagnosed, and treated for the disorder than ethnic minority children. Although access to quality care is certainly determined by social and economic factors, a family’s decision to seek mental health treatments is strongly influenced by knowledge and beliefs about mental illness. In a sample of Latino families, after controlling for socioeconomic status, family cultural values and beliefs including those about gender roles and natural or spiritual harmony/disharmony predicted attitudes about ADHD as an illness as well as attitudes about child behaviors associated with inattention or hyperactivity (Lawton and Gerdes, 2014).

Principles of Local Tailoring

Communications research and strategies in public health should focus simultaneously on individuals, their social networks, larger communities, and the locales that influence behavior and health (Abroms and Maibach, 2008). For example, a study of the factors that led to an increase in the number of cases of autism diagnosed in California concluded that neighborhood parks, stores, and schools were places of key social interaction and information diffusion about the condition among parents. One way to understand what influences local communities and to tailor research and interventions to local areas is to involve local opinion leaders in the design and evaluation of interventions. In keeping with a “grassroots” approach to designing their national stigma reduction initiative, “Opening Minds,” the Mental Health Commission of Canada evaluated the outcomes of many local initiatives and invested in scaling-up and replicating those that demonstrated effectiveness (Pietrus, 2013).

INTERVENTION RESEARCH

A strong intervention research component needs to be included in a multipronged national strategy to ensure that the various interventions (e.g., contact-based programs, educational programs, and mass media campaigns) are working as intended, and that they are producing the intended effects. The focus in this section is on measurement of stigma-related constructs, design considerations, and appropriate cost-benefit analyses.

Measurement of Stigma-Related Constructs

In line with the research on how stigma is defined and manifested, as discussed in Chapters 1 and 2, various domains are used to measure components of stigma. Most commonly measured domains include labeling, stereotyping, cognitive separating (i.e., us versus them), emotional reactions of the stigmatizer or of people who are stigmatized, interpersonal discrimination (i.e., expected, believed, or experienced), and structural discrimination (Link et al., 2004). In addition, three types of behavior are commonly measured: the behavior of people who have a mental illness that may serve as a stimulus to stigma, the behavior of people with mental illness in response to discrimination (e.g., avoidance or coping) (Link et al., 2004); and the behavior of people or institutions that are stigmatizing (e.g., discriminating, coercing, segregating) (Corrigan and Shapiro, 2010).

A review of measures used in mental illness stigma research that was conducted for the committee (Yang and Link, 2015) showed that prominent measures used to assess stigma among adult general community members cover most of the domains described above and include assessments of social distance, opinions about mental illness, community attitudes toward mental illness, semantic differential, attribution measures, emotional responses, and perceived devaluation-discrimination. The measures are established in their use and demonstrate good reliability and construct validity. In particular, the social distance, semantic differential, and opinion scales have a long history of use, and social distance and semantic differential scales have been used as the primary outcome in nationally representative surveys of attitudes toward people with mental illness in Australia (Reavley and Jorm, 2011) and the United States (see Pescosolido, 2015, for detail on concepts and measures; Pescosolido et al., 2010).

Program evaluators assessing outcomes of addictions stigma change efforts often rely on measures that have been adapted or developed specifically for studies about substance use (Pearson, 2015). Examples of mental health methods that have been adapted for use with substance use disorders include the Perceived Stigma of Addiction Scale (PSAS) to assess perceived public stigma in individuals in treatment for substance use problems (Luoma et al., 2010). The PSAS was found to be moderately correlated with measures of self-stigma including internalized shame and internalized stigma. In developing scales specific to stigma toward illicit drug use (Stigma of Drug Users Scale and the Drug Use Stigmatization Scale), Palamar and colleagues (2011) found that the two scales were measuring distinct forms of stigma: perceived (i.e., indirectly rating what most peoples’ attitudes are) and stigmatization (i.e., directly rating one’s own attitude). The authors also reported finding construct validity through

correlations showing that higher levels of stigmatization or greater perceived public stigma were inversely related to exposure to problem users. For assessing self-stigma, the Substance Abuse Self-Stigma Scale (Luoma et al., 2013) has four subscales, based on relational frame theory, with two subscales assessing self-stigma directly (self-devaluation, fear of enacted stigma) and two subscales assessing maladaptive reactions to self-stigma (stigma avoidance, values disengagement). The scales have shown theoretically consistent associations across a range of stigma-related constructs in residential or outpatient treatment populations (Brown et al., 2015; Luoma et al., 2013).

General versus Specific Measures

Concerns arise from scales that use “mental illness” as the referent in questions. In these cases there is no way to ascertain what respondents are thinking about (e.g., schizophrenia, depression, or attention deficit hyperactivity disorder). This becomes especially problematic as research has documented very different levels of prejudice toward individuals with different mental illness and substance use disorders (Barry et al., 2014; Martin et al., 2000).

Labeled versus Case-Based Approaches

Concerns also arise from asking about a category of mental illness or substance use disorder rather than a case description. While a category provides a clear referent, it does not match what individuals observe in the real world in terms of an individual behaving in a particular manner. Further, providing a categorical label does not allow researchers to ascertain whether individuals recognize a given condition.

Construct Validity

Construct validity is the extent to which a scale or test is actually measuring what it intends to measure. This is important in stigma research at all stages from planning to evaluation because of the many perspectives and constructs that underlie the phenomenon. For example, an instrument or scale may be intended to measure public, self, or structural stigma; perceived attitudes of other people or one’s own attitudes; or discrimination that is experienced or anticipated. Findings related to construct validity in stigma research are mixed. While evidence of construct validity has been found in some measures that assess public attitudes and emotions (Link et al., 2004), a recent meta-analysis (Stevelink et al., 2012) found that only one out of six mental illness-related stigma scales reviewed had accept-

able construct validity. This suggests that, when selecting measures of stigma and reviewing their psychometric properties, particular attention should be paid to tests of their construct validity and the extent to which the constructs being measured correlate with other similar measures.

Social Validity

When designing programs to change behavior, it is important to ensure that the intervention is considered socially important, ethical, and acceptable to both the target of the intervention and outside observers. Social validity (sometimes referred to as ecological validity or cultural validity) is a multidimensional concept that traditionally serves as an assessment of both the importance of a behavioral intervention and the acceptability of the intervention. It is often assessed across three factors: the goals of the intervention, the intervention procedures, and the intervention outcomes (Foster and Mash, 1999). Although the discussion of how best to assess social validity is evolving, historically researchers rely on normative comparisons to assess the importance of an intervention and subjective evaluations to assess the acceptability of the intervention (Foster and Mash, 1999). Because social validity is unique to each intervention, it is critical to involve stakeholders (e.g., people with serious mental illnesses, employers, family members, providers) in formative research during the evaluation planning stages to appropriately assess the social validity of the planned intervention (Corrigan and Shapiro, 2010).

Research Methods and Design Considerations

In developing methods to evaluate specific interventions in a national anti-stigma initiative, the type of intervention and subject matter require special attention. This section briefly covers designs for evaluating the effectiveness of interventions and issues that could affect the research designs, including social desirability, external validity, and fidelity or internal validity.

Research Designs

Research designs used for evaluating the effectiveness of stigma reduction interventions are the same as those used in other social and behavioral sciences. Randomized controlled trials are considered to be the gold standard because random assignment of participants to the intervention being implemented versus a control group or comparison group allows for a causal inference to be drawn between the intervention and the outcome (Shadish et al., 2002). Randomized controlled trials are useful

for everything from laboratory tests of the effectiveness of specific content and methods of persuasion to field tests of entire programs. However, randomized controlled trials are rarely used in evaluating large media campaigns because of the difficulties in random assignment and controlling conditions at community or larger scale levels (Hornik, 2002; Noar, 2006). It is also difficult to identify a comparison group that was not exposed to the campaign’s message (Hornik, 2002). Time-series designs that assess processes and outcomes of campaigns at multiple points can be useful in comparing expected and actual trends in stigma and discrimination reduction (Collins et al., 2012). Alternatively, a dose-response test can be used to measure the effect of a campaign’s message across different communities with different levels of exposure (Collins et al., 2012).

Social Desirability

In research studies assessing stigma, social desirability can interfere if respondents want to offer a positive image of themselves rather than respond with their truest attitude (Tourangeau and Yan, 2007). This can introduce bias in measures, and it can also pose problems in research designs that include pre- and posttest assessment or repeated measure designs when study participants may anticipate the desired response (Corrigan and Shapiro, 2010). A recent study found that “differences” (i.e., between a survey respondent and a person with mental illness) are more likely to be endorsed than “stereotypes” because differences may be considered neutral, while stereotypes are generally perceived to be negative (Corrigan et al., 2015b). Therefore, using measures of difference and incorporating repeated measures in the design may yield more sensitive indicators of stigma change.

Fidelity of an Intervention and Internal Validity

Internal validity is the degree to which researchers can conclude that findings are due to deliberate experimental manipulation rather than unaccounted for or confounding variables. Closely related, intervention fidelity is the degree to which the intervention was delivered as planned, rather than deviating from original design. Just as in other forms of intervention research, attention has to be paid to identifying the hypothesized effective ingredients of interventions and assessing the fidelity with which those intervention components are being delivered. For example, in a contact-based intervention the effective ingredients may be the scripted introduction, the description of the purpose, or the dialogue (Collins et al., 2012). The extent to which the quality of the interaction varies from the ideal model could pose one type of threat to internal validity and

affect results (Shadish et al., 2002). Evans-Lacko and colleagues (2013b) reported on the active ingredients of a contact program delivered as part of the anti-stigma marketing campaign in England from 2009 to 2011. They found that personal contact predicted positive changes in knowledge and attitudes for the school students. Consumers’ stories about their mental health problems and of their contact with a range of services had the greatest impact on the target audiences in terms of reducing mental illness stigma.

External Validity

Studies designed to understand social phenomena and to assess social interventions are often conducted with restricted populations (e.g., college students), which limits generalizability of the findings. Targeted and local stigma change has to be implemented and evaluated in real world settings to translate theory into practice and allow for more realistic tests of how, for example, employers in a large city or small town landlords respond to anti-stigma programs (Collins et al., 2012; Corrigan and Shapiro, 2010).

Cost-Benefit Analyses

Cost-benefit analyses are critically important in large national initiatives in which the return on investment is under scrutiny. As an example, data on the evaluation of the national Time to Change initiative in England (see Chapter 4) were combined with their social marketing campaign expenditure data to estimate the economic impact on employment for people with depression (Evans-Lacko et al., 2013a). Based on average national social marketing campaign costs, analysts found that the economic benefits outweighed costs and concluded that the campaign’s anti-stigma social marketing component is a potentially cost-effective and low-cost intervention for reducing the impact of stigma on people with mental health problems. Canada is also incorporating economic analyses of workplace-based anti-stigma and mental illness awareness programs in their evaluations of the Opening Minds initiative (Pietrus, 2013). They are developing an economic model to estimate a breakeven point with the idea that stigma reduction may lead to reduction of a major source of disability costs (i.e., those from mental illness).

MONITORING NATIONAL TRENDS

In addition to evaluating the multiple interventions that constitute a national strategy for reducing stigma, resources also need to be invested in monitoring national trends in attitudes, beliefs, knowledge, and behaviors

toward people with behavioral disorders. These trends need to be monitored to detect changes in structural, public, and self-stigma. Such trends would provide feedback loops on successful interventions or identify the possible need for course correction or new interventions. At present, for example, the United States lacks basic surveillance data and standards for reporting suicide and suicide attempts. This is one of the main challenges to the implementation of the National Strategy for Suicide Prevention (Caine, 2013). The rest of this section briefly suggests possible methods for monitoring changes in institutional structures (structural stigma), changes in social norms at the population level (public stigma), and changes in the lives of people who have been stigmatized including levels of self-stigma.

Monitoring Changes in Institutional Structures

Much research still needs to be carried out before measures of structural stigma can be operationalized and measured in order to monitor national trends. But there is promise in the methods being developed through studies conducted of:

• legislation and legal restrictions across many states, which document discrimination against people with mental illness (see, e.g., Burton, 1990; Corrigan et al., 2005b; Hemmens et al., 2002);
• discrimination suits filed under the Americans with Disabilities Act (ADA) for mental disorders (Colker, 2001; Scheid, 1999, 2005; Stuart, 2006);
• employment discrimination suits (Swanson et al., 2006); and
• disparity of funding for mental health services and research compared to general physical health (Kelly, 2006; Mark et al., 2014).

In addition, there are models from other countries. One example is how Canada is evaluating some of its initiatives to reduce sructural stigma by monitoring change in the media portrayals of peole with mental illness (the Media Monitoring Project) (Mental Health Commission of Canada, 2013).

Monitoring Changes in Social Norms at the Population Level

In earlier chapters, we presented findings from studies that examined social norms (public attitudes, beliefs, and behaviors) about people with mental and substance use disorders that prevail in the United States. Most survey items relating to stigma were added to surveys examining the prevalence of mental illness, substance use, or other conditions but were not necessarily carried out on a periodic basis. However, to monitor

trends of a multipronged national initiative to reduce stigma, these types of items on population-based surveys would need to be re-administered on a periodic basis. Items would need to be reviewed and updated to capture the most recent research findings on indicators of stigma (e.g., attitudes, beliefs, behaviors, discrimination, access to care). The results would be important indicators of public health risk, in the same way as other contextual factors that are included on surveys (e.g., social support, housing, income, employment).

Several surveys could be considered as possible candidates if they could be administered on a regular periodic basis. For example, a mental health module of the General Social Survey (GSS) was administered in 1996 and repeated in 2006. As detailed in Chapter 2, substantial changes were seen over time in relation to respondents’ understanding about mental illness and treatment (Pescosolido et al., 2010). Another example is the Behavioral Risk Factor Surveillance System (BRFSS) that incorporated a mental illness and stigma module in 2007 in which 35 states participated. Smaller numbers of states have participated in subsequent years because it is a voluntary add-on module to the BRFSS (Centers for Disease Control and Prevention et al., 2012). If these modules of the GSS or BRFSS were more broadly and continuously supported, they might be candidates for monitoring trends in public attitudes.

Another possible candidate is the National Survey of Drug Use and Health sponsored by the Substance Use and Mental Health Services Administration, which is administered annually and includes questions related to accessing services and unmet service needs. Respondents have a wide array of possible responses pertaining to reasons for not accessing services or treatments. Some of these reasons include stigma-related variables, such as concerns about confidentiality, potential negative opinions of neighbors and employers, and fear of being committed to a hospital. Differences from year to year in the numbers of people who access service, have unmet needs, and have reasons for not accessing services make this survey a candidate for monitoring trends in other attitudes, beliefs, and behaviors related to stigma.

Social media also have emerged as valuable resources for researchers collecting data on and monitoring trends in public attitudes and behaviors toward health. Twitter, in particular, has been used to track disease outbreaks, such as influenza, and also to successfully monitor suicide risk (Jashinsky et al., 2014). Due to the low cost and relative ease of use, researchers need to continue to explore social media platforms for monitoring behavioral health conditions, assessing public attitudes toward people with the conditions, and delivering anti-stigma messages (Jashinsky et al., 2014; Korda and Itani, 2013).

Monitoring Changes in the Lived Experience of People Who Are Stigmatized

Monitoring changes in the lived experiences of people with mental and substance use disorders provides a ground-level view that would supplement results of monitoring changes in structural stigma and public stigma, as well as an evaluation of outcomes of anti-stigma campaigns or social inclusion interventions (Brohan et al., 2010).

Constructs of the lived experience of stigma include perceived public stigma, experienced and anticipated discrimination, self-devaluation, and stigma avoidance (Brohan et al., 2010; Luoma, et al., 2013), as well as more distal measures of the influence of stigma on treatment-seeking, relationships with family and friends, employment, and housing. As change in stigma would be measured over time, an ideal research design would be longitudinal. Repeated measure designs have been used to study stigma constructs, such as:

• the enduring effects of stigma on the well-being of men with dual diagnoses of mental disorder and substance abuse (Link et al., 1997);
• differences in the effects of court-ordered treatment versus no court order on stigmatization, symptoms, treatment engagement, self-esteem, quality of life, and functioning among people with serious mental illness (Link et al, 2008);
• whether psychological consequences of stigma associated with HIV are lasting or transitory (Kang et al., 2006);
• discrimination suits filed under ADA for mental disorders (Colker, 2001; Scheid, 1999, 2005; Stuart, 2006);
• employment discrimination suits (Swanson et al., 2006);
• the nature and variety of discrimination experiences in the community and their impact on self-stigma, psychosocial resources, and the course of illness (Wright et al., 2000); and
• disparity of funding between research for mental health services and research for general physical health (Kelly, 2006; Mark et al., 2014).

SUGGESTED AREAS FOR FUTURE RESEARCH

Although there is evidence in many areas of stigma research as summarized in earlier chapters of this report, critical gaps remain. For example, as described above, it should be a priority for the field to move beyond periodic surveys and develop a strategic plan for monitoring change on an ongoing basis in attitudes, beliefs, and behaviors at structural, public,

and individual levels. Based on the information drawn from the public workshops, commissioned papers, and deliberations, the committee suggests areas for future stigma research in no order of priority, along with specific research questions related to these topical areas.

Stigma of Substance Use Disorders

Significantly more is known about the stigma of mental illness and related processes than about the stigma associated with substance use, abuse, and addiction. Surveys of public attitudes about multiple stigmatizing labels indicate that drug and alcohol misuse are viewed significantly more negatively than depression or schizophrenia (Schomerus et al., 2011). More research is needed on the nature, extent, and dynamics of stigma toward people with substance use disorders and the associated social- and psychological-related processes to better inform intervention and behavior change-related research. While some comparative stigma research suggests that there are common elements across stigmatized conditions (Fife and Wright, 2000; Pescosolido et al., 2010), more work is needed to identify the unique dimensions of stigma related to substance use disorders and what it has in common with mental illness and other stigmatizing conditions. Areas of future research might include

• What psychological, interpersonal, community, or social factors influence stigma toward people with substance use disorders are most amenable to change?
• Why is the public more stigmatizing of people with substance use disorders than people with mental illness?
• What is the role of cultural belief systems about health and illness related to stigma on substance disorders?
• Do people with co-occurring substance use and mental disorders experience stigma from multiple sources?
• Has criminalization of certain substance use worsened the impact of addiction on individuals and communities?
• Do 12-step programs diminish stigma? Do these programs promote shame, which worsens the impact of stigma?
• How do other public health approaches, such as harm reduction and “Housing First,” interact with stigma prevention and reduction programs?
• What are the long-term outcomes of substance use disorder initiatives?

Reducing Structural Stigma

A large proportion of anti-stigma interventions have focused on public rather than structural stigma. Although there is also a dearth of research related to structural stigma, there is a relevant body of work examining the structural barriers that people with mental and substance use disorders face. Examples include discriminatory legislation that places restrictions on jury service, voting, holding political office, and parental custody rights; and health datasets that document negative outcomes of structural stigma on individuals, including treatment utilization, chronicity, or persistence of symptoms (e.g., Hatzenbuehler and Link, 2014; Hatzenbuehler et al., 2014). Studies are also needed to examine how structural, public, and self-stigma, interact with each other.

To fully address stigma at this level, research is needed to identify and target manifestations of structural stigma that are less easily recognized, such as the unintended consequences of anti-stigma efforts, and to understand how structural stigma persists in the presence of national legislative efforts to protect stigmatized groups such as parity laws (Corrigan et al. 2004b). Research is needed to clarify the impacts of new policies, along with their implementation and enforcement, on mental health and health-related disparities among people with behavioral health disorders.

The disproportionate representation of people with mental and substance use disorders in the criminal justice system is well established (e.g., Abramsky, 2015; Epperson and Pettus-Davis, 2015; Giliberti, 2015; Gingrich and Jones, 2015; Wisniewski, 2015). However, more research is needed to examine the relationships and pathways between criminality and the indicators of structural stigma and both criminal justice and health outcomes.

Additional Manifestations of Stigma

This section briefly identifies and defines manifestations of stigma that are not characterized as discrete in the same way that structural, public, and self-stigma are characterized in the research and intervention literature. These forms of stigma (label avoidance, double and intersecting stigmas, and courtesy or family stigma) require further study to understand their nature and role in the larger stigma complex, which have implications for new interventions.

Label Avoidance

The term “label avoidance” is related to both public and self-stigma. It occurs when people with behavioral disorders perceive stigma and

opt to avoid the labeling it engenders by not seeking treatment or help that would identify them as a member of the stigmatized group (Abramsky, 2015; Corrigan et al., 2009a, 2014; Epperson and Pettus-Davis, 2015; Giliberti, 2015; Gingrich and Jones, 2015; Wisniewski, 2015). A systematic review by Clement and colleagues (2015) found that fear of the impact of disclosing one’s mental or substance use disorder was the most commonly reported barrier to help-seeking. Label avoidance also undermines the outcomes of antidiscrimination legislation; for example, if an individual does not disclose to an employer, employment protections in the ADA, and its amendments, are not relevant (Cummings et al., 2013).

Results of a national survey examining the prevalence of mental disorders and treatment utilization found that approximately 10 percent of people who reported not seeking treatment avoided treatment for their disorders because they were afraid it would have negative impacts on neighbors and their jobs (Substance Abuse and Mental Health Services Administration [SAMHSA], 2013). Label avoidance and concerns about disclosure can also keep people from obtaining peer services and community support.

Double and Intersecting Stigmas

People with mental and substance use disorders such as HIV/AIDS may experience stigma related to other physical disorders, racial or ethnic heritage, sexual orientation, or gender identity. Gary (2005) used the term “double stigma” to describe the experience of mental illness across four ethnic minority groups in the United States. The author observed that the double stigma not only impeded treatment-seeking but may have contributed to the development of co-morbidities such as substance use disorders that might also go untreated.

In a qualitative study exploring the experiences of lesbian, gay, bisexual, and transgender (LGBT) individuals, Kidd and colleagues (2011) observed complex intersections of stigma experienced by the study participants in relation to mental illness and their LGBT identities. The resulting first-person narratives suggested that building supportive networks can smooth the path to resilience and recovery.

Courtesy or Family Stigma

Families and friends of individuals with mental and substance use disorders may experience a “courtesy” burden of stigma. Stigmatizing attitudes toward families of people with substance use disorders have been found to be greater than for other health conditions, and family

members are more likely to be blamed for onsets and relapses of a relative’s mental than physical disorder. These families were also more likely to say that they had been avoided socially by others (Corrigan et al., 2006a).

The experience of courtesy or family stigma can lead to self-stigma and self-blame, and it can undermine the support of friends and families for care-seeking (Corrigan et al., 2014). Public stigma also influences a family’s previous experiences with treatment providers and with others in wider family and social circles. Hopeful beliefs about the possibility of recovery and positive attitudes toward people with mental illness and substance use disorders would encourage family members, neighbors, and friends to offer their support and encouragement.

Targeting and Evaluating Contact and Education Programs

There is robust evidence that contact is related to lower levels of stigma making contact-based interventions a prime strategy for implementing change. Optimal target groups for contact and education programs are those with influential roles in the lives of people with mental and substance use disorders (Corrigan, 2004, 2011). While targeted interventions have rightly addressed stigma in health care providers, law enforcement, journalists, and students, other important sources of stigma have seldom been sought out for research or interventions. These include landlords, religious leaders, policy makers, and human resources gatekeepers.

One potential, but largely untested, target group is government officials who often have influence over the funding of mental health and substance use treatment, the siting of treatment facilities, and the enactment and enforcement of laws and policies protecting (or harming) people with mental and substance use disorders.

Another target for research and intervention would be members of the criminal justice system in addition to law enforcement officers, including court officials and probation officers, all of whom have frequent contact with people with mental and substance use disorders, partly due to the criminalization of some substance use activity.

The media represent another potential target. Evaluations of interventions targeted at the media are sparse in the literature, and those that are published vary in quality (Borschmann et al., 2014; Clement et al., 2013; Corrigan, 2012). Social media, such as Facebook and Twitter, have been used in national efforts in Australia, New Zealand, Spain, and Sweden (Betton et al., 2015), but a comprehensive review was unable to identify

interventions using multiple types of media or any that evaluated television, radio, cellphone, or movies (Clement et al., 2013).

Teachers and other lay advisors play important roles in promoting the mental health of children and adolescents and are another potential target for research and intervention. The public support for lay advisors such as teachers suggests that faith-based and community leaders are also potential audiences for contact-based intervention efforts, in particular, due to the supportive and affirmative role they have within their communities (Pescosolido et al., 2008a).

Emergency responders who are tasked with being the first on the scene in a wide range of crises could also be targets for anti-stigma interventions. Many people with substance use disorders receive their only treatment in the emergency room (Cohen et al., 2007; Pescosolido et al., 2008b).

Family and friends of people with mental and substance use disorders are among the most frequently cited perpetrators of discrimination in several national studies (Corker et al., 2013; Thornicroft et al., 2014), which suggests that research on and interventions tailored to these groups are essential elements of stigma change campaigns.

Children and adolescents are rarely targets of stigma change initiatives (Livingston, 2012). Peer-reviewed research on classroom anti-stigma programs for youth are particularly sparse in the literature (Mellor, 2014), although the Opening Minds initiative is currently evaluating several youth programs (Mental Health Commission of Canada, 2013). No media-based anti-stigma interventions for children have yet been fully evaluated (Clement et al., 2013).

In addition to research that identifies effective targets for interventions, more research is needed to understand the outcomes of contact and education programs, including internally valid hypothesis testing. Questions for future research might include

• How do outcomes of contact programs compare to education or combination (contact and education) programs?
• How do outcomes of contact and education programs compare to protest, advocacy, and legislative change strategies? Do they work in combination?
• Do early childhood, school-based stigma change initiatives influence the development of negative behavioral health social norms later in life?
• Given that many contact-based, anti-stigma initiatives come from within peer and activist communities, what are the factors that contribute to success of contact-based programs?

• What are the characteristics of peers who effectively deliver the desired message(s)?
• What are the most effective peer service training programs?

Measuring Behavior Change

Recent reviews of stigma change initiatives have found few studies that targeted and measured actual behavior toward people with mental illness (Dalky, 2012). As behavioral change is the ultimate goal of stigma reduction, more research efforts are required in this area. More and varied methods are needed for measuring behavior change in target populations because research suggests that greater awareness of mental illness or mental health literacy alone will not decrease stigma or discrimination against people with mental illness (Evans-Lacko et al., 2013b; Pescosolido, 2013; Schomerus et al., 2012).

In the broader prejudice and discrimination literature, the past 2 decades have seen a surge in the amount of work examining implicit measures of prejudice toward a wide range of social groups (Greenwald et al., 2009). People may hold automatic biases of which they are unaware that can in turn result in discriminatory behaviors that place people with substance use disorder at a disadvantage. Although these measures have received minimal attention in the stigma toward mental illness literature (O’Driscoll et al., 2012), they hold much promise to examine aspects of both self-stigma and public stigma, which are not necessarily accessible to one’s conscious mental processes.

Finally, psychometric testing and validation of measures are needed within target populations and cannot be presumed to be generalizable based on strong reliability and validity in other populations (Brown, 2011).

The Stigma Complex

Much of what the literature suggests about influences on public stigma is theory based or speculative with little empirical backing. Even less is known about the development and predictors of self-stigma. This is due in part to the dearth of longitudinal research in this area (Kulesza et al., 2013). Robust longitudinal data would contribute to knowledge of stigma’s antecedents and consequences, as well as the development of negative social norms over time in individuals and among the general public. Also important, longitudinal data would support efforts to monitor stigma change over time in the United States, including change across

regions, and how change varies by demographic variables, such as sex, age, race, and ethnicity.

It would be important to follow the progress of successful anti-stigma initiatives in other countries, including the evolution of national norms and opinions over time, and specific knowledge gained about what changed public opinion, how long it took to effect change, and how the changes endure or continue to evolve in the desired direction(s).

Key issues to monitor in the United States include variability across states in implementation and enforcement of key legislation, including the ADA, Affordable Care Act, and Mental Health Parity Act and related outcomes for people with mental and substance use disorders.

Finally, research on stigma change efforts tend to focus on decreasing stigma and discrimination, but it is also important to monitor change in affirming attitudes and behaviors, for example, the belief that with proper support people with mental illnesses can perform and maintain jobs like people without mental illness (Corrigan and Shapiro, 2010).

Strategic Prioritization of Research

The topics above are not listed in order of priority since priorities change over time. SAMHSA will identify other research questions concerning its previous initiatives to improve behavioral health social norms.

SAMHSA has articulated its strategic program initiatives in areas of prevention of substance abuse and mental illness, health care and health systems integration, trauma and justice, recovery support, health information technology, and workforce development (Substance Abuse and Mental Health Services Administration, 2014). An important next step will be for SAMHSA to strategically prioritize research initiatives that will be informed by the science of changing behavioral health social norms.

A model for such prioritization of research was developed by the Transportation Research Board (2013). Although the methods were focused on transportation safety research, they can be applied across other fields. The overarching principles emphasize developing a data-driven approach of prioritization, encouraging collaboration and communication among stakeholders, encouraging bottom-up initiation of projects, joint funding to leverage resources for common purposes, focusing on “large, multi-modal research efforts of national importance” (Transportation Research Board, 2013, p. 8), and ensuring transparency of the prioritization process and results to all stakeholders.

The prioritization process includes several steps. Most important, the selection criteria need to be developed with a structure for weighting possible research projects. Considerations under this step include a project’s contribution to the strategic vision; whether it has appropriate

perspectives at the national, regional, or local level; how it fills a gap in the knowledge base; how it addresses an urgent or longer term needs; feasibility given existing resources, including the level of financial investment and risk; and potential value added in terms of economics, benefits to the end users, or support of other mission goals and objectives.

Subsequent steps involve publishing the selection criteria; developing a process for submitting proposals for future research projects; soliciting input from all stakeholders and all levels of organizations; developing a prioritized list of research projects that are based on the selection criteria and fit with mission, vision, and goals; and finally the identification of potential tradeoffs and alternate scenarios based on potential or varying levels of investments.

This page intentionally left blank.