Background

Stigma is the devaluation of a group or individual based on a characteristic that is discredited by society (Goffman, 2009), such as having a mental illness, physical health condition, minority sexual orientation or gender identity, female sex, higher body weight, disability, and minority ethnic/racial heritage. The intersectionality literature highlights how different characteristics interact and overlap to affect an individual’s experiences of discrimination and privilege (Crenshaw, 1989). Stigma processes include negative attitudes, opinions, and feelings (prejudice), unfair treatment when these attitudes are acted upon (discrimination), and internalization of negative stereotypes, known as self-stigma (Corrigan & Watson, 2002). Theories of stigma have proposed that harm to self-esteem is a possible consequence (Corrigan et al., 2009; Link, Cullen, Struening, Shrout, & Dohrenwend, 1989; Major & O’Brien, 2005). Low self-esteem means a low opinion of oneself and lack of self-worth. Lower self-esteem has been observed in stigmatized groups, for example people with minority sexual orientation (Bridge, Smith, & Rimes, 2019), higher body weight (Sikorski et al., 2015), and psychosis (Bradshaw & Brekke, 1999). Cross-sectional studies have found associations between self-stigma and low self-esteem (Austin & Goodman, 2017; Corrigan et al., 2006; Durso et al., 2012). Longitudinal studies have found evidence for a causal role for stigma processes in the reduction of self-esteem (Greene, Way, & Pahl, 2006; Link, Struening, Neese-Todd, Asmussen, & Phelan, 2001). Discrimination experiences in adolescence have been associated with lower self-esteem in adulthood (Yang, Chen, Choi, & Kurtulus, 2019), suggesting that stigma-related experiences in early life may have lasting effects on self-esteem. Low self-esteem is typically viewed as a transdiagnostic construct in relation to mental illness. Not only might it be a symptom, consequence, or maintaining factor for mental illness but also there is evidence that low self-esteem can be a vulnerability factor for depression (Orth, Robins, & Widaman, 2012; Ju & Lee, 2018), psychosis (Krabbendam et al., 2002), eating disorders (Cervera et al., 2003), social anxiety (van Tuijl, de Jong, Sportel, de Hullu, & Nauta, 2014), and other psychosocial problems (Trzesniewski et al., 2006). Low self-esteem is associated with negative outcomes in other life domains including relationship and job satisfaction (Orth, Robins, & Widaman, 2012).

Research suggests that responses and coping styles may increase vulnerability or buffer against effects of stigma (Miller & Kaiser, 2001). Group attachment and external attribution of blame can be protective for self-esteem (Bourguignon, Seron, Yzerbyt, & Herman, 2006; Crocker & Major, 1989). Poorer outcomes have been associated with avoidance and withdrawal, compared to using social support and proactive coping (Aristegui, Radusky, Zalazar, Lucas, & Sued, 2018; McDermott, Umaña-Taylor, & Zeiders, 2019). Rumination has been found to mediate the relationship between stigma-related stress and psychological distress (Hatzenbuehler, Nolen-Hoeksema, & Dovidio, 2009).

Interventions for low self-esteem include cognitive behavioural therapy (CBT) based on Fennell’s (1997) model, which aims to address factors that maintain low self-esteem, such as self-critical thinking and avoidance, and underlying core beliefs and unhelpful attitudes (Fennell, 1998). A meta-analysis of interventions based on Fennell’s model found strong effect sizes for improvements in self-esteem and reductions in depressive symptoms (Kolubinski, Frings, Nikčević, Lawrence, & Spada, 2018). Other forms of cognitive behavioural interventions may also be helpful. Rose, McIntyre, and Rimes (2018) found significant improvements in self-esteem following a six-session compassion-focussed intervention for highly self-critical university students. As far as the authors are aware, there have been no investigations of cognitive behavioural (or other) interventions targeting low self-esteem in people with a range of stigmatized characteristics.

Aims

The current study evaluated a new cognitive behavioural intervention for low self-esteem, tailored for people aged 16–24 years who had experienced stigma, prejudice, and/or discrimination. Young adults were targeted as a relatively early intervention approach with individuals who had sufficient life experience to be aware of the impact of stigma and low self-esteem on their life.

The following research questions were explored: Is the intervention feasible to deliver in terms of recruitment, retention, and protocol-adherence? Do participants find the study process and intervention acceptable? Compared to pre-intervention, at post-intervention and follow-up, do participants report higher self-esteem, lower functional impairments related to self-esteem, and reductions in depression and anxiety? Do participants show improvements in measures of potential processes of change; self-criticism, higher self-compassion, and coping responses related to stigma? Are changes in self-esteem associated with improvements in these process measures? Method Ethical statement

Ethical approval was obtained from the local university’s Research Ethics Subcommittee.

Design

This was an uncontrolled study with repeated measures at baseline (at least 2 weeks prior to session 1), pre-intervention (session 1), mid-intervention (after session 3), post-intervention (after session 6), and two-month follow-up. Qualitative and quantitative data were collected to investigate feasibility and acceptability of the intervention, study process, and assessment methods. Standardized self-report measures were collected to provide preliminary information about pre- to post-treatment changes. Randomization was not included because the purpose was to undertake primary data collection relating to feasibility and acceptability of the novel intervention, as the first stage of piloting within the process of intervention development, informed by MRC recommendations framework for development of complex interventions (O'Cathain et al., 2019).

Participants

People aged 16–24 years, who identified that stigma, prejudice, or discrimination had negatively impacted their self-esteem, were recruited from the general population. To participate they had to score under 25 on the Rosenberg Self-Esteem Scale (Isomaa, Väänänen, Fröjd, Kaltiala-Heino, & Marttunen, 2012; Rosenberg, 1965), report that low self-esteem was causing significant impairment in daily functioning as indicated by a score of 10 or above on the Work and Social Adjustment Scale (Mundt, Marks, Shear, & Greist, 2002), and have sufficient proficiency in English. They were excluded if they were receiving another psychological intervention, had started or changed dose of psychoactive medication in the past three months, met criteria for a current serious mental health problem that would be more appropriately treated in NHS services (e.g. schizophrenia, bipolar disorder, or anorexia nervosa), reported suicidal ideation with intent or plan to act on suicidal thoughts, or reported recent self-harm. The target sample size was 16–25 participants, in line with recommendations for pilot intervention studies (Hertzog, 2008).

Procedure

Recruitment took place from April to November 2019, with two researchers working 1–2 days per week on the study; follow-ups were completed in April 2020. The study was advertised primarily online, using social media, research recruitment registers, and circular emails inviting volunteers for university research projects. Potential participants accessed the information sheet online, registered their interest, and were provided with screening questions to complete online. If eligible based on screening questions, they were invited for telephone screening. If not eligible, participants were signposted to sources of support where appropriate. Following telephone screening, eligible participants completed the consent form and baseline questionnaires and were offered an appointment for session 1 at least two weeks later. Links to questionnaires were sent to participants from a study email account, and email reminders prompted participants if they were uncompleted.

Intervention

Two trainee clinical psychologists delivered the intervention, supervised by two consultant clinical psychologists in weekly group supervision. Therapists were in their second and third years of training, had CBT experience, had learned about compassion-focussed therapy (CFT), and were involved in intervention development. Six 1-hour individual sessions were delivered weekly, with booklets to supplement learning (see Table 1 for intervention content). The intervention was informed by stigma theory and research. It was based on CBT principles including Fennell’s (1997) model of low self-esteem and Gilbert’s Compassionate Mind approach (Gilbert, 2009). The CFT ‘threat/safety strategy’ formulation (Gilbert, 2010) was used to identify experiences contributing to key fears (known as ‘core beliefs’ in traditional CBT) about self and others, such as discrimination experiences, current coping strategies, and their intended and unintended consequences. Gilbert’s ‘three systems’ model of emotion regulation (Gilbert, 2009) was used throughout the intervention. These existing approaches were adapted to be tailored to people with stigmatized characteristics. For example, psychoeducation was provided about the activation of the ‘Threat’ system in response to stigma, and internalization of negative attitudes (self-stigma) in development and maintenance of low self-esteem. Each session incorporated stigma-specific explanations and examples. The first two sessions were adapted from Rose et al. (2018) to address self-criticism. Session 4 focussed on developing more helpful beliefs as alternatives to key fears about the self. Optional ‘modules’ for sessions 3 and 5 corresponded to unhelpful processes that were found to be associated with low self-esteem reported by stigmatized individuals in previous research, for example rumination and avoidance. The modular approach had the advantage of tailoring sessions to participants’ formulations. Every session was audio-recorded and listened to by the therapists’ supervisor, to assess fidelity and for supervision purposes.

Table 1. Core and optional modules for each session of the cognitive behavioural intervention for low self-esteem in young people with stigmatized characteristics Core modules Description of module Formulation of low self-esteem in the context of stigma and identifying self-critical thinking (Session 1) Collaborative formulation of how low self-esteem developed and is maintained in the context of stigmatized characteristics (e.g. highlight role of discrimination experiences and self-stigma in development of negative beliefs about self, world, and others) Psychoeducation regarding self-esteem, ‘three systems’ model of emotion regulation; how stigma activates the threat system; self-compassionate approach (Gilbert, 2009) Introduction to self-criticism and monitoring of self-critical thinking for homework Self-compassion as an alternative to self-criticism (Session 2) Introduction of self-compassion as an alternative to self-criticism, including in response to experiences of stigma Completion of a self-compassionate thought record in session and for homework Introduction of compassionate imagery exercise for homework Addressing Key Fears (Session 4) Psychoeducation about link between stigma, low self-esteem, and overgeneralized negative beliefs about the self and/or others (core beliefs, called ‘key fears’ in this intervention) Updating key fear with new information Identifying more helpful belief as alternative to key fear Set up positive data log and behavioural experiment Therapy summary/goal setting for follow-up period (Session 6) Review of previous sessions and techniques Therapy summary including overview of main learning points and planning for situations where self-esteem might be more vulnerable or stigmatizing experiences may occur Goal setting for between Session 6 and two-month follow-up Follow-up telephone session (2 months after session 6) Review current level of self-esteem, goals/plan and each intervention strategy Option to discuss one strategy in more detail Relapse prevention plan Signposting to further support if necessary Optional modules (Sessions 3 & 5) Description of module Dealing with avoidancea Collaborative discussion about role of avoidance in maintenance of low self-esteem Psychoeducation about avoidance, including as a response to stigma, prejudice, or discrimination Identify alternatives (approach behaviours); set up behavioural experiment/graded exposure Reducing overthinkinga Collaborative discussion about role of overthinking in maintenance of low self-esteem Psychoeducation about overthinking (worry/rumination), including in response to stigma, prejudice, or discrimination Identifying alternatives (distraction, present moment focus, move towards action, e.g. group connection/activism) and set up practice for homework Problem solving (optional for participant) Letting go of very high standards Identify perfectionist standards as coping strategy for having a stigmatized characteristic and how this maintains low self-esteem Identify pros/cons of perfectionist standards and behaviours Explore current perfectionist standard and identify alternative Set up behavioural experiment If time: plan to expand another valued area of life Social comparisons, social media use, and role modelsa Collaborative discussion about social comparisons and social media use, including in relation to stigma, and how these factors can maintain of low self-esteem Psychoeducation about consequences of social comparison Identify positive social media use (e.g. connecting with others who share stigma experiences; engaging with organizations that empower marginalized voices) Make plan for more helpful social media use Finding role models (optional) Assertivenessa Collaborative discussion about low assertiveness, including in the face of stigma, and how this contributes to maintenance of low self-esteem Psychoeducation regarding communication styles Self-compassionate thought record in relation to assertiveness (optional) Set up behavioural experiment to be assertive Role play assertive communication (optional to client) Coping with unpleasant feelingsa Collaborative discussion about current distress tolerance strategies in maintenance of low self-esteem Psychoeducation about distress tolerance and impact of avoiding emotions Strategies to accept/tolerate unpleasant feelings Set up behavioural experiment to try new strategy Hiding part of ourselvesb Collaborative discussion about concealment of stigmatized characteristics in maintenance of low self-esteem Psychoeducation about hiding stigmatized characteristics Cost/benefit analysis Hierarchy for safe disclosure or behaviour change Set up behavioural experiment to test feared disclosure or behaviour change Building a support networka,b, a,b Collaborative discussion about reduced social support in the context of stigma, and maintenance of low self-esteem Psychoeducation regarding purpose of support network Identification of barriers to accessing support Identification of at least one new source of support and set up plan for homework Working with early memoriesb,c, b,c Identify negative/distressing memories linked to key fear and provide rationale for memory-focussed techniques Discrimination training: identify differences between ‘then’ and ‘now’; practice for homework Imagery re-scripting: ‘re-live’ the event from participant’s perspective and insert new meaning or information (using updates completed in key fears module) Note a Module can be shortened and combined with another module in sessions 3 and 5 b Modules were prepared and offered but not delivered c Only to be delivered at Session 5 if negative images and memories from the past are problematic and linked to key fear. Feasibility and acceptability objectives

Feasibility was assessed by examining rates of recruitment, suitability of eligibility criteria, retention, and protocol adherence. Acceptability was assessed through participant feedback.

Fidelity

Treatment fidelity was assessed by therapists’ primary supervisor. Session recordings were rated a dichotomous yes/no for whether therapists adhered to session protocol. Protocol violations were recorded and fed back to therapists.

Measures

Standardized questionnaires were completed online. Clinical measures were administered at screening. Clinical and process measures were completed at baseline, pre-intervention, mid-intervention, post-intervention, and two-month follow-up.

Screening tool for mental health problems Mini International Neuropsychiatric Interview (MINI, version 7.0.2)

The MINI, a brief structured interview for major disorders in DSM-5 and ICD 10, was administered at telephone screening to assess current mental health diagnoses (Sheehan et al., 1998).

Clinical measures Rosenberg Self-Esteem Scale (RSES)

The RSES has ten items assessing global self-esteem; responses are on a 4-point scale from 1 (‘strongly disagree’) to 4 (‘strongly agree’; Rosenberg, 1965). Five items provide a negative statement and are reverse scored. Total scores range from 10 to 40; higher scores reflect higher self-esteem. Cronbach’s alphas were .55 to .91 (<70 at baseline and pre-intervention).

Work and Social Adjustment Scale (WSAS)

The WSAS assesses day-to-day functioning in five domains: work, home management, social leisure activities, private leisure activities, and relationships (Mundt et al., 2002). Usual wording of the scale’s instructions was adapted (‘your problems’ replaced with ‘low self-esteem’) to assess impact of low self-esteem. Responses are on an 8-point scale from 0 (‘not at all’) to 8 (‘very severely’). Total scores range from 0 to 40; greater scores indicate poorer functioning. Cronbach’s alphas were .69 to .89.

Patient Health Questionnaire (PHQ-9)

The PHQ-9 has nine items measuring depressive symptoms (Kroenke, Spitzer, & Williams, 2001). Response options range from 0 (‘not at all’) to 3 (‘nearly every day’). Higher scores indicate greater severity of depressive symptoms. Cronbach’s alphas were .72 to .90.

Generalized Anxiety Disorder (GAD-7)

The GAD-7 has seven items measuring symptoms of anxiety, on a scale from 0 (‘not at all’) to 3 (‘nearly every day’; Spitzer, Kroenke, Williams, & Löwe, 2006). Higher scores indicate greater severity of anxiety-related symptoms. Cronbach’s alphas were .82 to .90.

Process measures Forms of Criticism/Self-Attacking and Self-Reassuring Scale (FSCRS)

The FSCRS is a 22-item instrument with a three-factor structure (Castilho, Pinto-Gouveia, & Duarte, 2013; Gilbert, Clarke, Hempel, Miles, & Irons, 2004). The ‘Inadequate Self’ subscale (nine items) measures self-criticism related to perceived inadequacy and disappointment with oneself; ‘Hated Self’ (five items) measures self-criticism relating to self-hatred and self-punishment; ‘Reassured Self’ (eight items) measures the ability to reassure and sooth oneself. Responses range from 0 (‘not at all like me’) to 4 (‘extremely like me’). Cronbach’s alphas: FSCRS-IS.56 (baseline) −.96; FSCRS-HS.31 (pre-intervention) −.91; FSCRS-RS .68 to .96.

Self-Compassion Scale (SCS)

The SCS has 26 items; responses range from 1 (‘almost never’) to 5 (‘almost always’; Neff, 2003). A total SCS score can be used as an overall measure of self-compassion (Neff, Whittaker, & Karl, 2017). Higher scores indicate greater self-compassion. Cronbach’s alphas ranged from .88 to .96.

Discrimination and Prejudice Responses Scale (DAPR)

The DAPR assesses coping responses associated with prejudice and discrimination, with 11 components representing different categories of responses, each with four items (Armstrong, Henderson, & Rimes, 2019). Subscales are Preparation (e.g. ‘ready myself for encountering prejudice or discrimination’), Raise Awareness (‘educate people about the characteristic(s) to increase their understanding’), Avoidance (‘avoid people who I know to be prejudiced’), Enjoyable Activity (‘do an activity that makes me feel good’), Group Attachment (‘identify more closely with other people with the characteristic(s)’), Secrecy (‘hide the characteristic(s) from people’), Self-reliance (‘rely on myself more than others’), Distancing (‘change my behaviour to avoid being stereotyped by people’), Rumination (‘go over and over what I could have done differently during the event’), Resignation (‘put up with the way I am treated’), and Blame (‘blame the individual(s) involved for their behaviour rather than myself’). Responses range from 1 (‘never’) to 5 (‘always’). Cronbach’s alphas were Resignation: .32 to .72 (<70 at all time points except baseline); Avoidance: .60 (mid-intervention) −.81; Distancing: .68 to .92; Rumination: .70 to .97; Blame: .72 to .92; Raise Awareness: .81 to .95; Enjoyable Activity: .83 to .96; Group Attachment: .87 to .92; Self-reliance: .78 to .86; Preparation: .85 to .97; Secrecy: .90 to .98.

Feedback

Participant feedback was collected at post-intervention and follow-up, with quantitative rating scales and open-ended questions, adapted from Rose et al. (2018). Post-intervention feedback included general questions about the intervention and study process, and ratings of usefulness of intervention elements. Post-intervention, participants reported the average proportion of weekly booklets they had read, and average time spent on weekly homework. At follow-up, ratings of usefulness and frequency of use of elements of the intervention were collected.

Qualitative feedback was collected from therapists and supervisors after intervention delivery was completed.

Data analyses

Descriptive statistics were conducted to examine feasibility and quantitative feedback items. Open-ended feedback was analysed using brief content analysis (Mayring, 2019).

Descriptive statistics and within-participants effects (effect sizes and 95% confidence intervals) were conducted to examine changes in scores for each measure. Changes are presented for the eighteen treatment completers. Change scores were calculated for the following time points: baseline to pre-intervention; pre-intervention to post-intervention; pre-intervention to follow-up; post-intervention to follow-up. Effect sizes were calculated using Cohen’s dz for repeated measures and interpreted using the following cut-offs: ‘negligible’ effect <0.2; small effect ≥0.2, medium effect ≥0.5, large effect ≥0.8.

Reliable Change Index (RCI) was used to discover whether participants made reliable improvements and reliable recovery (Jacobson & Truax, 1991). The proportion of participants who made reliable improvements was calculated, and the proportion of participants who made reliable recovery was calculated for RSES and WSAS. Reliable recovery was not examined for PHQ-9 and GAD-7 because participants did not all begin the intervention in the clinical range.

Associations between pre- to post-intervention change scores and clinical and process measures were explored using Spearman’s correlations.

Results Participant characteristics

Twenty-two participants began the intervention. Their characteristics are shown in Table 2. Participants’ mean age was 21.9 years. Participants were mostly female (n = 20, 90.9%) and fulltime students (n = 16, 72.7%). Participants were ethnically diverse; the largest ethnic group was White (n = 7, 31.8%). Participants reported a range of stigmatized characteristics; most common were race/ethnicity and having experienced a mental health condition. Three participants (13.6%) reported one characteristic, most reported two or three (total n = 18, 81.8%). One participant reported four characteristics (4.5%). Most participants met criteria for at least one psychiatric disorder at screening and had received psychological therapy before.

Table 2. Characteristics of participants who started the intervention (n = 22) Characteristics Age, mean (SD, range), years 21.3 (1.9, 18–24) Sex, n (%) Female 20 (90.9) Male 2 (9.1) Ethnicity, n (%) Asian 4 (18.2) Black 3 (13.6) Mixed/multiple ethnic groups 5 (22.7) Other ethnic group 3 (13.6) White 7 (31.8) Highest education achievement A-levels 1 (4.5) Undergraduate/Bachelors (current) 14 (63.6) Undergraduate/Bachelors (obtained) 5 (22.7) Postgraduate/Masters (current) 2 (9.1) Self-identified stigmatized characteristics, n (%) Race/ethnicity 8 (36.4) Mental health condition 8 (36.4) Body weight 7 (31.8) Other physical appearance 7 (31.8) Sexual orientation 5 (22.7) Religion 4 (18.2) Sex (female) 5 (22.7) Physical health condition 3 (13.6) Class/socioeconomic status 2 (9.1) Learning difficulty (dyslexia or dyspraxia) 2 (9.1) Gender identity 1 (4.5) Psychiatric diagnoses, n (%) None 3 (13.6) Major depressive episode Current 14 (63.6) Past 2 (9.1) Panic disorder Current 2 (9.1) Lifetime 2 (9.1) Social anxiety disorder (Current) 7 (31.8) Alcohol use disorder (Past) 1 (4.5) Psychotic disorder (Lifetime) 1 (4.5) Bulimia nervosa Current 4 (18.1) Past 1 (4.5) Binge-eating disorder (Current) 1 (4.5) Generalized anxiety disorder (Current) 6 (27.3) Past psychological therapy, n (%) Cognitive behavioural therapy 5 (22.7) Counselling 4 (18.2) Group dialectical behavioural therapy 1 (4.5) Group therapy, unspecified type 1 (4.5) Other individual psychotherapy 3 (13.6) Past psychological therapy (any) 14 (63.6) Feasibility

Figure 1 summarizes details of recruitment, retention, and reasons for drop-out. One hundred and fifty-eight people responded to advertisements, 93 co