A novel technology-enhanced internalized stigma and shame intervention for HIV-positive persons with substance use disorders

Abstract

Internalized stigma, shame, and other negative self-conscious emotions are inadequately addressed barriers to HIV-related self-care, particularly among people actively using substances. Innovative approaches are needed to optimize antiretroviral treatment (ART) adherence as well as engagement in HIV care among people living with HIV and substance use disorders. Based on qualitative feedback from providers and patients, we iteratively developed and conducted a proof-of-concept study of a relatively brief transdiagnostic emotion regulation intervention designed to improve ART adherence care by addressing behavioral and psychological barriers, including internalized stigma and shame, among people living with HIV and active substance use disorders. The final intervention included 5 individual sessions focused on metacognitive awareness of emotions and thoughts, cognitive reframing of dysfunctional thoughts about the self using concepts such as self-compassion, and identifying and reaching the participants’ personalized HIV-self-care goal(s). All participants received daily texts querying current emotion and weekly texts querying ART adherence and substance use. To extend the effects of the intervention, we developed a personalized bi-directional text component through which participants received their personalized compassionate self-statements, informed by the intervention content, in response to their answers to emotion queries for 8 weeks after the 5 sessions. The texts modeled using compassionate self-statements as a form of cognitive reframing, consistent with cognitive restructuring of distorted core beliefs. We consented 10 participants living with HIV and problematic substance use in the proof-of-concept pilot. Of the 8 participants who completed all intervention sessions, participants replied to 70% of all text messages sent. All 8 reported strong acceptability of the intervention content. This emotion-focused, technology-enhanced intervention demonstrated proof-of-concept, in that this patient population would participate in this intervention. A larger randomized controlled pilot is needed to determine feasibility and acceptability among people living with HIV and substance use disorders, a hard-to-reach and underserved population.

Introduction

In this era of HIV treatment as prevention (TasP), addressing barriers to engagement in HIV care is a public health priority (Gardner, McLees, Steiner, Del Rio, & Burman, 2011; Horstmann, Brown, Islam, Buck, & Agins, 2010; Mayer, 2011), as attrition over the HIV care continuum is associated with delayed viral load suppression and increased risk of onward HIV transmission (Cohen et al., 2016; NIH HIV/AIDS Guidelines, 2016). Multiple studies have identified associations between internalized stigma and antiretroviral therapy (ART) nonadherence and poor engagement in care (Blake Helms et al., 2016; Diiorio et al., 2009; Rao et al., 2012; Sayles, Ryan, Silver, Sarkisian, & Cunningham, 2007; Sayles, Wong, Kinsler, Martins, & Cunningham, 2009; Turan et al., 2016; Gwadz et al., 2016). Further, HIV-related shame has been associated with inconsistent ART nonadherence (Konkle-Parker, Erlen, & Dubbert, 2008), engagement in clinical trials (Zúñiga, Blanco, Martínez, Strathdee, & Gifford, 2007), and sexual HIV transmission risk behavior (Sikkema, Hansen, Meade, Kochman, & Fox, 2009). Qualitative research has also identified shame as a perceived barrier to self-care and substance use reduction among people living with HIV and co-occurring illnesses such as HCV (Batchelder et al., 2013; Batchelder, Peyser, Nahvi, Arnsten, & Litwin, 2015). Emerging evidence indicates that internalized stigma and related emotions, such as shame, may act as barriers to optimal engagement in HIV care (Dickerson, Gruenewald, & Kemeny, 2004; Earnshaw, Bogart, Dovidio, & Williams, 2013; Levi-Minzi & Surratt, 2014). However, these barriers have not been the primary focus of extant psychological interventions (Bennett, Traub, Mace, Juarascio, & O’Hayer, 2016).

Focusing on shame and other stigma-related emotions, also referred to as self-conscious emotions, may be a unique point of psychological leverage for optimal engagement in HIV care (Lickel, Kushlev, Savalei, Matta, & Schmader, 2014). Self-conscious emotions, such as shame, are considered social emotions as they result from the appraisal of how an experience pertains to one’s self in relation to others (e.g., “I am less valuable than others because I use drugs”; Tagney & Tracy, 2012). Shame specifically has been associated with avoidance and, paradoxically, a desire to change one’s self or behavior (Bennett, Hersh, Herres, & Foster, 2015; Lickel et al., 2014; Tagney & Tracy, 2012). This is consistent with the divergent findings that higher levels of self-reported shame are associated with both higher likelihood of relapse among people recovering from alcohol use disorder (Randles & Tracy, 2013) and longer duration in substance use treatment (Luoma, Kohlenberg, Hayes, & Fletcher, 2012). It may be that cognitive and behavioral responses to self-conscious emotions, like shame, are crucial determinants of whether and how individuals will effectively engage in self-care behaviors. For some, motivation to avoid shame may prevent the benefit of action-triggering aspects of this emotion (Lickel et al., 2014). Literature informed by Stress and Coping Theory suggests that identity-related stressors and related threat appraisals lead to chronic shame or other self-conscious emotions such as embarrassment or guilt, which exacerbate maladaptive avoidance of a stressor (Friedman & Silver, 2006). Among individuals living with HIV and other stigmatized identities, such as being a person with substance use disorders, the shame related to internalized stigma may lead to avoidance of HIV care or continued problematic substance use (Berjot & Gillet, 2011; Folkman & Lazarus, 1988).

Evidence indicates that self-care can be challenging for people living with HIV and substance use disorders in part due to emotional dysregulation (Blake Helms et al., 2016; Carrico et al., 2007; Luoma, Kulesza, Hayes, Kohlenberg, & Larimer, 2014; Pachankis et al., 2015), or the ability to influence the experience or expression of emotions (Gross, 2002). Higher levels of HIV-related discrimination, substance use, and ART nonadherence have all been associated with difficulties regulating negative emotions (Cheetham, Allen, Yücel, & Lubman, 2010; Pachankis et al., 2015). Emerging evidence indicates that enhancement of emotional and cognitive awareness coupled with reframing, or reappraisal, strategies is associated with better adjustment to HIV-related negative emotions, retention in HIV-care, antiretroviral adherence, and reductions in substance use (Carrico et al., 2007).

Several promising strategies have been used to address emotion regulation challenges among people living with HIV and substance use disorders. Interventions have employed metacognitive approaches such as mindfulness, more traditional cognitive strategies such as cognitive reframing (Carrico & Antoni, 2008; Creswell, Myers, Cole, & Irwin, 2009), as well as behavioral strategies such as coping skills training and problem solving (Carrico & Antoni, 2008; Safren et al., 2012), commonly used in cognitive-behavioral therapy. However, most of these interventions have been tested in HIV-positive persons without substance use disorders.

Metacognitive skills can enable nonjudgmental present-centered awareness of cognitive, emotional, and behavioral barriers to self-care. Mindfulness-based interventions have been shown to improve emotion regulation, reduce avoidance behaviors, increase self-compassion, and improve self-care (Black, 2014; Bowen et al., 2014; Hoppes, 2006; Khanna & Greeson, 2013). While there is mixed evidence for the effectiveness of mindfulness strategies on relapse reduction (Grant et al., 2017), some evidence indicates that Mindfulness-Based Relapse Prevention is associated with decreased substance use (Witkiewitz & Bowen, 2010). It may be that components of these strategies can be used to enhance metacognitive awareness, enabling the identification of shame and related emotions and cognitions. Further, this may be a necessary first step in reframing negative thoughts and beliefs about the self that act as barriers to HIV self-care. Greater metacognitive awareness, rather than nonjudgmental acceptance, may also facilitate engagement in cognitive reframing, coping skills training, and problem solving to enhance HIV-related self-care.

Metacognitive awareness of dysfunctional core beliefs may enable cognitive reframing of such beliefs. By providing classic cognitive behavioral strategies for challenging cognitive errors related to the self, combined with new ways of conceptualizing the self (e.g., using self-compassion; Neff & Germer, 2013), individuals may be able to reframe distorted core beliefs. People with higher self-compassion report that emotions such as shame are less of a barrier to seeking medical care, practicing safe sex, and disclosing HIV-status (Dawson Rose et al., 2014). Additionally, individuals with higher self-compassion report less HIV-related sexual risk behaviors in the presence of substance use (Terry, Leary, Mehta, & Henderson, 2013) and more health promoting behaviors (Terry et al., 2013; Ngô, 2013). Taken together, metacognitive awareness may be important for acknowledging distorted core beliefs in order to utilize self-compassion to reframe those beliefs that contribute to problematic substance use and serve as key obstacles to HIV self-care.

In addition to metacognitive awareness and cognitive reframing, setting attainable HIV self-care goals is a useful strategy for improving HIV health outcomes (Moskowitz et al., 2017; Safren et al., 2001). This involves identifying attainable goals, barriers to achieving those goals, problem-solving strategies to address the barriers, and adjusting the goals as needed. Cognitive behavioral goal setting and problem-solving strategies have strong evidence for improving HIV-related self-care, such as Life-Steps (Safren et al., 2001) and Cognitive Behavioral Therapy for Adherence and Depression (CBT-AD; Safren et al., 2009).

While there are several existing intervention strategies with promising results for improving HIV adherence and engagement in care for people with substance use disorders, most of these interventions are resource intensive, involve those in substance use recovery, and do not focus on negative self-conscious emotions or internalized stigma (e.g., Carrico et al., 2014; Landovitz et al., 2014; Esposito-Smythers et al., 2014). Rather, existing interventions to improve HIV adherence and engagement in care among people with active substance use disorders leverage substance use treatment methods, such as contingency management, to reduce or eliminate substance use in order to improve HIV outcomes. Other existing interventions target depression as a barrier to HIV care, which may overlap with internalized stigma and shame, but do necessarily co-occur (e.g., Safren et al., 2009). While this intervention embraces harm reduction and may be beneficial for people living with depression, it explicitly aims to improve emotion regulation related to internalized stigma and shame, differing from existing interventions addressing substance use or depression as barriers to HIV-care. To our knowledge, there are no current evidence-based interventions that specifically treat internalized stigma, shame, and related emotions to improve engagement in HIV-care among individuals with active substance use disorders.

Consistent with the ORBIT Model’s iterative nature of intervention development, we present the iterative design of the intervention (consistent with Phase I’s define and refine components) as a proof-of-concept pilot (consistent with Phase II’s preliminary testing component; Czajkowski et al., 2015). Following this process, we iteratively developed a relatively brief intervention, informed by evidence-based content including Life-steps, CBT for Adherence and Depression (CBT-AD), Mindful Self-Compassion, and Mindfulness Based Relapse Prevention (Neff & Germer, 2013; Safren et al., 2001; Safren et al. 2009; Witkiewitz & Bowen, 2010), as well as stakeholder feedback. Based on stakeholder feedback, this intervention included only 5 sessions with a clinician, supplemented with a bidirectional text messaging component for an additional 8 weeks, to extend the impact of the intervention. The duration was informed by feedback related to structural barriers (e.g., limited access to transportation) and the availability of costly clinician-time in low-resource clinical settings.

Methods

Intervention Development

Leveraging current theory, existing literature, and evidence-based intervention components, as well as iterative interviews with providers and individuals who met inclusion criteria, we developed an intervention to address internalized stigma and negative self-conscious emotion to improve HIV-related self-care. This was based on the hypothesized pathway of 1) experienced stigma related to stigmatized identities (i.e., being a person with HIV and substance use disorders) → 2) internalization of that stigma including negative cognitions about the self → 3) the affective experience of shame, which evokes a desire to avoid reminders of the impetus → 4) behavioral avoidance of reminders of HIV, such as substance use and ART non-adherence. We leveraged the Revised Stress and Coping Theory, an appraisal-based model that includes evaluating the significance of stressful events and coping options (i.e., problem, emotion, and meaning-based coping; Breines & Chen, 2012; Folkman & Lazarus, 1988; Pecoraro et al., 2013; Price, Diana, Smith-Dijulio, & Voss, 2013). Notably, in this work appraisals include both accurate and distorted perceptions.

Initial materials were drafted based on theory, existing literature, and previous work focused on emotion regulation and ART adherence (Carrico et al., 2015, 2016; Moskowitz et al., 2017; Safren et al., 2009). To ensure that the protocols were relevant for this population, qualitative interviews were conducted with ten providers who serve people living with HIV and substance use disorders. Inclusion criteria for these interviews included working with people living with HIV and substance use disorders. We then conducted condensed visits with two HIV positive adults with substance use disorders to obtain feedback on the manual and rectify any challenges. Informed consent was obtained before each interview. Qualitative feedback was iteratively integrated into the protocols and intervention materials, resulting in a final manual, protocols, and materials with altered organization, language, examples, including reduction of the literacy level of exercises initially used in other interventions (described in the results).

Based on stakeholder feedback, we developed a bidirectional text message system for this project. We built the internet backend using Drupal™, a widely used content management system with strong security features. The system included participant phone numbers, but not names, email addresses, or any other identifying information. All internet communications were encrypted with SSL, a standard encryption technology used for business, research, and financial transactions. The software included a customizable, automated, bidirectional text messaging system with a data export dashboard that enabled study staff to program automated text messages and monitor participants’ responses. The system and procedures were approved by the University of California, San Francisco’s institutional review board (IRB) and were deemed HIPAA compliant.

Intervention Design

Enrolled participants were asked to attend five in person intervention sessions that lasted 60 minutes. All participants received two emotion query text messages daily through the five weeks of in-person sessions and for approximately two months after the last session. In addition to the session content described below, participants developed compassionate self-statements at the end of each session, which involved reframing core beliefs about the self, based on intervention content. In the final session, participants selected the compassionate self-statements they felt would be most helpful when experiencing each of the self-conscious emotion response options from the daily text queries. For two months following the in-person sessions, we sent participants their chosen compassionate self-statements in response to the endorsement of the linked emotion response (e.g., shame) to model using compassionate self-reframing when experiencing self-conscious emotions. See Figure 1 for text message outline.

Figure 1.

Text component of the intervention

Intervention Content

Individual session content.

The final individual session content involved three domains addressed in each of the five sessions, including metacognitive awareness of emotions and cognitions, cognitive reframing (e.g., compassionate self-statements), and identifying and refining self-care goals (e.g., goal setting and problem-solving skill development). See Table 1 for session content.

Table 1:

Intervention Content

	Session 1	Session 2	Session 3	Session 4	Session 5
Metacognitive Awareness: Emotions &Thoughts	Review emotion glossary, focusing on differentiating self-conscious emotions. Begin labeling emotions (introduce daily text queries).	Introduce metacognitive awareness of emotions and cognitions. Discuss and practice informal mindfulness.	Continue to practice and discuss metacognitive awareness.	Reflect on how emotions and thoughts impact behavior. Practice awareness about oneself in relation to self-care.	Review content covered over all sessions and practice exercises related to challenging concepts.
Cognitive Reframing	Introduce self-compassion and compassionate self-statements. Write 2–3 self-statements in log in study binder.	Introduce concept of identifying and challenging dysfunctional thoughts about the self. Write 2–3 self-statements.	Introduce cognitive reframing and practice using self-compassion to challenge dysfunctional thoughts about the self. Write 2–3 self-statements.	Revisit cognitive reframing, practice ‘self-reframing.’ Write 2–3 self-statements.	Write 2–3 more self-statements. Pair self-statements with text response options.
Behavioral Goal Setting and Problem Solving	Identify attainable self-care goal, identify barriers and a stepped plan, update self-care goal(s) as needed, and discuss continuation plan for pursuing self-care goal(s).

Metacognitive awareness of cognitions and emotions.

These included informal mindfulness approaches (Hoppes, 2006; Black, 2014) coupled with psychoeducation related to emotions to create awareness of emotions, particularly self-conscious emotions and related cognitions (e.g., “I am a bad person” or “I am not worthy of self-care because I use drugs”). The psychoeducation related to emotion identification, differentiation, and labeling involved adapted content from other emotion regulation interventions (Carrico et al., 2015; Moskowitz et al., 2017), as well as an emotion glossary (see Supplemental Material 1). This glossary was iteratively developed during the qualitative interview phase, to include emoticon images (Keltner et al., n.d.), given the expressed concern in preliminary interviews that existing emotion identification tools appear child-oriented and, therefore, condescending to adult participants.

Cognitive reframing.

As metacognitive awareness may not be enough to change internalized stigma, shame, and other self-conscious emotions, we combined awareness with cognitive reframing strategies, including alternative conceptualizations of the self, such as those consistent with self-compassion (Neff & Germer, 2013). Self-compassion involves developing a caring and compassionate attitude toward oneself, feeling connected to the larger human experience, and like mindfulness, involves engaging in nonjudgmental awareness of one’s experiences (Brion, Leary, & Drabkin, 2014). Self-compassion may inform the reframing of distorted core beliefs, or self-reappraisals, by reducing self-judgment, and thereby promote emotion regulation and engagement in health behaviors in the context of HIV and substance use. We differentiated cognitive reframing from metacognitive awareness by emphasizing that reframing involves identifying and potentially challenging maladaptive thoughts, whereas metacognitive awareness involves acknowledging but not changing emotions and thoughts. We also leveraged existing evidence-based cognitive behavioral therapy (CBT) interventions that emphasize cognitive reframing, to reduce the effect maladaptive cognitions have on HIV self-care: Moskowitz’s IRISS intervention (Moskowitz et al., 2017) and Carrico’s ARTEMIS intervention (Carrico et al., 2015).

Self-care goal setting & problem solving.

Self-care goal setting included identifying HIV self-care goal(s) and then breaking the goal(s) into attainable components, identifying potential barriers, and problem solving ways to handle those barriers. This domain leveraged Life-Steps (Safren et al., 2009; Safren, Otto, & Worth, 1999), a one-session cognitive behavioral intervention for improving HIV medication adherence that uses education, problem solving, and rehearsal strategies. This component also involved adapting several resources from CBT-AD (Safren et al., 2009). Successfully accomplishing attainable goals served as a behavioral form of challenging negative core beliefs. By accomplishing these goals, participants behaviorally demonstrated their deservingness of self-care while acknowledging negative self-related cognitions and emotions.

Text message component.

Throughout the intervention, participants received weekly and daily texts (see Figure 1). Text messages included truncated words to ensure texts were within the character limit of most phone plans, thereby avoiding texts being split into two messages, which are often received out of order. Participants were sent weekly adherence texts “How much of your medication did you take this week? (0–100%)” and weekly substance use texts, coded to minimize unintentional disclosure, “How were the skies this week? Txt back: clear, rainy, snowy, hail, cloudy, smoggy, other” adopted from Ingersoll (Ingersoll et al., 2014; Ingersoll et al., 2015). Response options included coded weather words for substances: “clear” (no substance use), “rainy” (alcohol), “snowy” (crack or cocaine), “hail” (methamphetamine), “cloudy” (marijuana), “smoggy” (marijuana), and “other” (other illicit substance use). Additionally, participants received two daily emotion query texts: “How are you feeling right now? Text back 1 letter: (A)nger, An(X)ious, (C)ontent, (F)ear, (J)oy, (S)adness, (W)orried or (O)ther, please specify” and “How are you feeling about yrself? Text back 1 ltr: (A)shamed, (C)onfident, (E)mbarrassed, (G)uilty, (P)roud, (S)elf-compassionate, or (O)ther, plse specify”. By asking participants to label their emotions we aimed to reinforce the metacognitive awareness of self-conscious emotions and related thoughts.

In the last session, participants worked with the interventionist to select compassionate self-statements from their generated list that would be most helpful when experiencing each of the self-conscious emotions (see Figure 1). For example, one participant chose to pair “I am more than just a junky” with the “(A)shamed” response because he felt this self-statement may be helpful when he felt ashamed. While the selection process was straightforward for some participants, others were helped by the interventionist asking the participant to imagine past situations that elicited each of the self-conscious emotions and to consider which of their self-statements may have been useful in that situation. Study staff then programmed the selected statement into the electronic platform, which automatically sent the selected statement when the participant responded with the linked emotion. This was intended to model the use of self-statements, or self-reappraisal, as an emotion regulation strategy. Finally, after receiving the individualized compassionate self-statement, participants received one additional text: “How useful was this? (0=extremely unhelpful to 5= extremely helpful)” to confirm the self-statement message was read.

Research Visits

The study involved three research visits (30–90 minutes), including baseline and two follow up visits for which participants were compensated. The first and last research visit involved blood draws for viral load.

Proof-of-Concept Pilot Inclusion Criteria

To be eligible for the study, participants were required to be eighteen years of age or older, HIV-positive, able to complete informed consent, consent to release their medical records, and own a cell phone with text message capacity, which could be government subsidized. Participants had the option of working with study staff to obtain a government-subsidized phone and/or add minutes to their existing cell phone plan, at no cost. The substance use criteria for the study was based on the results from the Alcohol, Smoking and Substance Involvement Screening Test (ASSIST; Humeniuk et al., 2008) and required that participants meet the criteria for moderate substance use in the last three months for two or more substances including alcohol, excluding cannabis and tobacco. Participants who endorsed active suicidality or psychosis on the Mini International Neuropsychiatric Interview (Pinninti, Madison, Musser, & Rissmiller, 2003) were excluded from the intervention and were provided a list of resources and treatment options, including psychiatric emergency services if needed. Being virally detectable and non-adherent to ART were not inclusion criteria.

Recruitment and setting.

Participants were recruited through the HIV and substance use clinics at a public hospital and a community health center focused on sexual and gender minorities and people living with HIV. Additionally, flyers were posted at local community health settings (e.g., outpatient substance use treatment locations, emergency rooms, and community-based organizations). Visits were conducted in private rooms in the substance use treatment clinic at a public hospital and at a community health center focused on serving sexual and gender minorities and people living with HIV.

Open Pilot Implementation

Our primary recruitment source was clinicians’ familiar with the recruitment criteria, resulting in screening only 27 individuals. Ten were deemed ineligible due to not owning or being willing to obtain a cell phone (n=2), not meeting the substance use criteria (n=7), and both (n=1). An additional 6 were eligible but did not attend the required baseline visit. One individual was deemed ineligible after completing the initial baseline assessment visit due to psychiatric instability, including florid psychosis, which was assessed to be a barrier to completing the study. Of the ten participants who were recruited and enrolled, eight completed the full intervention, including the face-to-face sessions and the text component (see Table 2). The two who did not complete the intervention sessions did respond to text messages and, notably, both reported housing challenges that interfered with attendance.

Table 2:

Participant Demographics

N=10	N(%)
Age (Mean(SD))	43.2 (9.6) Range (28–60)
Race & Ethnicity
Black/African American	4 (40%)
White Non-Hispanic	4 (40%)
Hispanic	2 (20%)
Education
Some high school	1 (10%)
High school graduate	7 (70%)
>High School	2 (20%)
Sexual Orientation
Exclusively Homosexual	3 (30%)
Predominantly Homosexual	1 (10%)
Predominantly Heterosexual	1 (10%)
Exclusively Heterosexual	5 (50%)
Ever Homeless	5 (50%)
Polysubstance use disorders
Alcohol + amphetamine	2 (20%)
Alcohol + cocaine	1 (10%)
Alcohol + amphetamine + cocaine	2 (20%)
Alcohol + amphetamine +cocaine + opioids	2 (10%)
Alcohol + cocaine + opioids	1 (10%)
Amphetamine + opioids	1 (10%)
Amphetamine + benzodiazepines + opioids	1 (10%)

This study was carried out in accordance with the Code of Ethics of the World Medical Association for experiments involving humans with all procedures approved by the institutional review board at the University of California, San Francisco. All participants provided informed consent and a Certificate of Confidentiality was obtained.

Results

We present the feedback from providers followed by a description of the small proof-of-concept pilot including preliminary feasibility and acceptability of the intervention, self-report results, and qualitative feedback from participants’ exit interviews. All provider and participant interviews were recorded, transcribed, and analyzed using Thematic Analysis (Braun & Clarke, 2006).

Initial Qualitative Feedback from Providers and Case Managers

Prior to implementing the proof-of-concept pilot, we conducted key informant interviews with ten providers including physicians, psychologists, social workers, and case managers who work with people living with HIV and substance use disorders. During these interviews, we obtained feedback on the perceived need for this type of intervention and the iteratively refined intervention content, duration, and materials (i.e., exercises, handouts, and study binders).

Perceived need.

All ten providers agreed that internalized stigma and shame were under-addressed barriers to HIV self-care among a subset of their patients living with HIV and active substance use disorders. One provider conveyed how this unaddressed challenge interferes with patient care.

“There’s a lot of internalized stuff... some days when I’m in clinic all I feel I’m doing is de-shaming people. They come in- not able to do anything because they’re feeling so bad about themselves.” - HIV primary care and addiction medicine internist

Another provider commented on the linkage between patients’ psychological experiences and non-adherence to antiretroviral medication.

“Why don’t our patients have this! ... I had a patient who was not taking her medication and I didn’t understand why- it was just a few pills every day. I finally asked and she said it was like getting diagnosed again every time she took them.” - Addiction medicine physician

Specific feedback.

Providers also commented on the intervention content and duration. One psychologist commented on the benefits of reframing, or reappraisal, in addition to metacognitive awareness, for patients with trauma histories and who experience lack of safety.

“These clients are feeling so unsafe and hopeless about getting out of wherever they are stuck with their drug use. It is helpful to start questioning some of that - just acceptance is like ‘I’m trying to accept that I’m a miserable failure.’” -Psychologist

Providers and case managers also commented on the literacy level of the content. This feedback informed the iterative refinement of the content, and resulted in reducing the overall literacy level of the worksheets and exercises (Flesch-Kincaid grade level of 6.4, with a Flesch Reading Ease score of 72.3) and the manual (Flesch-Kincaid grade level of 8.7, with a Flesch Reading Ease score of 60.1).

Providers also emphasized the many structural barriers this patient population faces, including limited access to transportation, housing, as well as food insecurity. This feedback influenced the number of in-person sessions and the text component, which we designed to extend the effect of the intervention. This feedback also influenced our decision to reimburse for travel and provide snacks during every session.

Open Pilot

Sample.

Ten participants’ completed the baseline interview. Their age ranged from 28–60 years old (mean, standard deviation (SD) = 43.2 (9.6). Eight identified as male and 2 as female. Four participants (40%) identified as African American and 2 (20%) as Hispanic or Latino. Eight (80%) reported having a high school education or less and 50% reported being gay or bisexual. All participants screened positive on the ASSIST for at least two substance use disorders. Nine screened positive for an alcohol use disorder and nine screened positive for a stimulant use disorder. Additionally, five participants screened positive for an opioid use disorder and one screened positive for a sedative, hypnotic, or other anxiolytic use disorder. Further, five (50%) reported ever having been homeless (Table 2).

Feasibility and acceptability.

Of the 10 participants who completed baseline, eight participants attended all five individual intervention sessions. Text response rates varied from 46–98% with participants responding to an average of 77% (SD=16%) of the queries during the individual sessions. During the 8-weeks following the in-person sessions, participants responded to an average of 70% of queries with a range of 9–100% (SD 28%). Notably, one participant initiated an in-patient substance use treatment program during the follow up portion of the study during which time she did not have phone access, which affected these results. Only five participants completed the two follow-up research visits, which may be partially attributable to a break in staffing during the follow up period. However, the three who did not complete the follow-up research visits did continue to respond to text-messages and did complete exit interviews over the phone. In exit interviews, participants were asked about their anticipated frequency of engaging in the skills covered in the intervention (0= not at all, 1=once a month, 2=once a week, 3=2–3 times/week, 4= daily). When asked about likelihood of practicing HIV self-care strategies the mean (SD; range) was 2.40 (0.55; 2–3), mindfulness was 3.40 (0.55; 3–4), self-compassion was 2.80 (1.10; 1–4); nonjudgmental acceptance was 2.80 (1.10; 1–4); and compassionate self-statements was 2.50 (1.30; 1–4). The mean likelihood of recommending the study to a friend on a scale of 0–10 (0 = being not at all–10 = definitely) was 9.3 (SD= 2.0) and a median of 10 (IQR:10–10). The likelihood of recommending to a friend who uses substances was comparable with a mean of 9.7 (SD=0.8) and a median of 10 (IQR: 10–10).

Self-report results.

While the aim of this proof-of-concept study was to assess preliminary feasibility and acceptability of this intervention in this hard-to-reach population, we obtained several pre and post self-report measures of HIV self-care (e.g., viral load and ART adherence). At baseline, all but one participant had undetectable viral load, all of which were maintained at post assessment. One participant had a detectable viral load at baseline but did not return for the follow up viral load assessment. Participants’ reported mean (SD) baseline adherence of 94.1% (5.3) and 89.2%. (14.3) after the intervention. However, in self-reported assessments collected during the intervention participants self-reported adherence rates ranged from 50%−100%. Similarly, in self-reported adherence rates collected via text message over the full course of intervention, responses ranged from 0–100%. However, in qualitative exit interviews, over half of participants reported improvements in their ART adherence. Given the small sample size and the variability in self-reported ART adherence rates, we are cautious regarding interpretation of these results.

Additionally, we conducted assessments of substance use and several psychological constructs including: the ASSIST was used to assess substance use (Humeniuk, R et al., 2008); the HIV and Abuse Related Shame (HARSI) was used to assess HIV-related shame (Neufeld, Sikkema, Lee, Kochman, & Hanson, 2012); the Self-Compassion Scale was used to assess self-compassion (Neff, 2003). Additionally, the CES-D was used to assess depression (Hann, Winter, & Jacobsen, 1999). Notably, participants reported finding the self-compassion assessments difficult to understand, and therefore, we are cautious regarding interpretation of these results. (See Table 3).

Table 3:

Self-Report Results

Mean(SD);range	Baseline (n=8)	Post Face-to-Face Sessions (n=5)	Post Intervention (n=5)
HIV-related shame subscale a
Impact of HIV-related shame on behavior a	2.16 (1.19); 0.62–4.00	1.82 (1.03); 0.31–3.08	2.02 (1.53); 0.31–3.62
	1.10 (1.28); 0.00–3.60	0.72 (0.68); 0.10–1.60	0.26 (0.26); 0.00–0.60
Self-Compassion b	3.08 (0.84); 1.73–4.08	2.99 (0.86); 1.77–3.81	3.15 (0.66); 2.23–3.88
Depression c	26.50 (18.17); 2.00–56.00	21.40 (13.65); 7.00–42.00	28.40 (18.28); 12–52.00
Mindfulnessd
Observing	3.23 (0.91); 1.50–4.88	2.95 (0.47); 2.50–3.63	3.62 (1.22); 1.71–5.00
Describing	3.15 (0.75); 1.71–4.25	2.97 (0.64); 2.25–3.71	3.50 (0.72); 2.86–4.50
Acting with Awareness	3.20 (0.65); 2.63–4.38	2.98 (0.44); 2.38–3.50	3.05 (1.37); 1.00–4.50
Non-judging of Inner Experience	3.06 (1.11); 1.50–5.00	3.30 (0.62); 2.50–4.13	3.28 (1.31); 1.00–4.25
Non-reactivity to Inner Experience	2.99 (0.91); 1.00–4.29	3.00 (0.45); 2.57–3.71	3.06 (1.39); 1.43–5.00
Substance Usee
Tobacco	16.50 (7.56); 3.00–27.00	13.20 (8.04); 4.00–21.00	10.40 (9.10); 2.00–23.00
Alcohol	16.75 (8.97); 4.00–35.00	15.20 (11.75); 3.00–32.00	12.40 (10.55); 1.00–24.00
Cannabis	8.88 (4.94); 4.00–18.00	8.60 (7.40); 2.00–20.00	4.80 (3.42); 1.00–10.00
Cocaine	15.25 (8.66); 4.00–27.00	11.20 (10.33); 4.00–29.00	9.80 (11.34); 1.00–29.00
Amphetamine	15.63 (11.56); 4.00–30.00	15.40 (14.10); 2.00–31.00	10.40 (8.96); 1.00–24.00
Sedative	8.63 (9.24); 4.00–29.00	6.60 (6.99); 2.00–19.00	3.20 (2.17); 1.00–6.00
Hallucinogens	4.13 (0.35); 4.00–5.00	3.60 (0.89); 2.00–4.00	3.40 (2.51); 1.00–7.00
Inhalants	4.00 (0.00); 4.00–4.00	3.60 (0.89); 2.00–4.00	3.60 (2.88); 1.00–8.00
Opioids	10.63 (10.35); 4.00– 30.00	12.80 (12.76); 3.00–30.00	4.40 (3.05); 1.00–8.00
Other	4.00 (0.93); 3.00–6.00	4.40 (2.70); 2.00–9.00	4.00 (3.08); 1.00–8.00
Total Substance Involvement Score	104.38 (34.64); 75.00–180.00	94.60 (42.65); 26.00–131.00	46.75); 16.00–119.00

^a.HIV Shame Inventory (HARSI), responses range from 0 (not at all) – 4 (very much), higher scores indicates greater shame;

^b.Self-Compassion Scale, responses range from 1 (almost never) – 5 (almost always), higher scores indicate greater self-compassion;

^c.CES-D, response options range from 0–60, higher scores indicate greater depressive symptoms;

^d.Five Facet Mindfulness Questionnaire, scores ranges from 1= never or very rarely true to 5= very often or always true, higher scores indicate greater levels of mindfulness;

^e.ASSIST substance scores (0–3= low risk, 4–26= moderate risk, 27+ = high risk; alcohol 0–10= low risk, 11–26= moderate risk, 27+= high risk).

Participant feedback.

In the exit interviews, all eight participants interviewed reported finding the intervention helpful including improving self-care behaviors, learning about and utilizing compassionate self-talk, and impacting the way they thought about themselves including experiencing less self-judgment, consistent with reductions in internalized stigma. The majority of participants reported improvements in their HIV-related self-care, including improving ART adherence or HIV-related appointment attendance. One participant described the changes he experienced over the course of the intervention.

“It kind of keeps me on track… about the problem, the HIV … making sure I take care of myself, making sure I go to every doctor’s appointment because I know how important it is, making sure I take my medicine, that’s important. So, just more thinking of all of these things, about which I need to care for myself.”

Several participants described specified benefits experienced by implementing compassionate self-talk. One participant explained her experience.

“I was taught to talk to myself like a best friend or a mother would talk to me… makes me feel real positive, I’m actually inspired… it’s a feeling of accomplishment, of being worthy. You know, the opposite of not wanting and not being able to complete things.”

Half of the participants explained that they experienced general changes in the way they think about themselves. One participant remarked how the intervention elicited thinking more about himself and his context.

“The whole program just really makes you think about yourself and putting more thought about certain things that are going on.”

Relationship to substance use.

Over half of the participants reported that the intervention influenced their substance use. One participant initiated a residential rehabilitation program during the course of the intervention, one reported no substance use at the final visit, and three reported reductions in substance use, including abstaining from specific drugs (i.e., abstaining from methamphetamine but continuing to drink alcohol). One participant described her newfound conceptualization of how her substance use related to her emotions and behavior.

“In order to be compassionate to yourself, you have to realize what you’re putting into your body and how it affects your emotions and behavior. It creates a situation where you allow negative forces to come into your world. It’s a seedy world in the drug use. There’s no way not to get grimy… There’s nothing you can get out of using drugs except feeling more depressed. So, it actually made me understand that. Hey, the less I self-medicate, the more I can deal with life on life’s terms.”

Several participants reported changes in their perceptions of the necessity of their substance use. One participant exemplified this testament, “It helped me to realize, hey, every day I don’t have to be high.”

Feedback on components of the face-to-face intervention.

In the exit interviews, participants were asked for their perspective on the specific components of the intervention including: metacognitive awareness, cognitive reframing, and goal setting.

Metacognitive awareness.

Several participants commented on the importance of metacognitive awareness and how being aware of one’s emotions can positively impact their self-care and interpersonal relationships.

“If I get frantic and edgy, it’s intolerable. It’s hard to express my feelings at that point.. that’s a majority of the time I forget about self-care. You’re more or less acting them out once you get to a certain stress level, so you have to definitely be mindful.”

“I realize that unless you actually can understand how you’re feeling and relate it to another person, they’ll never know. There are so many different words for different feelings, it’s very important to clarify those feelings.”

All participants described liking and utilizing at least one of the informal mindfulness exercises after the intervention. One participant conveyed how his continued use of one of the breathing exercises helped him regulate his anger after the intervention.

“Certain breathing exercises helped me calm [down] when I usually would just go fly off the handle.”

Cognitive reframing.

While most participants liked the concept of self-compassion initially, three participants conveyed initially feeling uncomfortable with the concept. Several participants first reported thinking that self-compassion sounded “selfish.” Notably, those who reported feeling initially uncomfortable all expressed that they grew to appreciate the concept over the course of the intervention. All participants reported currently liking some aspect of the concept in the exit interviews. One participant described how her perception of self-compassion changed.

“Now, I’m getting the same gratification another person would [from helping myself], I’ve learned to understand that that’s good, that’s important, that’s beneficial, that’s wonderful.”

Another participant reported how the study helped him care more about himself.

“[The study] taught me a little bit about myself… about caring about me.”

One participant described that while he struggles with being self-compassionate he now has integrated the concept into his daily life, including referencing it frequently with friends.

“A lot of times I have a hard time being self-compassionate. But it is kind of a coin word with me and my friends now.”

Goal setting.

The majority of participants reported benefitting from identifying and adjusting self-care goals, as well as the information related to identifying attainable goals and breaking them down into achievable steps. One participant explained how setting attainable goals was new for him and how he now applies these skills in different aspects of his life.

“I liked making goals for myself, that part was cool. I don’t normally do that sort of thing. I just go with the easiest course- and if you set a goal for yourself, then you can achieve more. So, that’s what I’m trying to do.”

Feedback on text component.

All participants reported appreciating the text messages and liking that they received their selected compassionate self-statements in response to their self-reported emotion. Further, almost all of the participants reported wanting to continue to receive the daily text messages.

The majority of the participants reported liking the text messages in general. Several participants commented on what receiving the texts meant to them.

“I looked forward to seeing [the texts], like somebody cares about me. God sending me a message, hey, are you alright today?”

“It was nice to get them because I don’t talk to too many people.”

Another participant commented on how identifying his emotions daily helped him become more aware of his emotions and ultimately led to caring more about himself.

“It’s hard to really get in touch with your feelings, especially if you haven’t practiced for a long time. That’s why the text messages made it better overall. It’s hard to care for yourself- it really is. It’s so easy to help everyone else and forget about yourself…”

Several participants commented on how receiving their selected compassionate self-statements in response to their reported emotion reminded them of the session content.

“I like getting the texts to remind me to just think positive and everything. They reminded me of when I wrote down everything. And, it reiterated the goals that I set for myself.”

Almost all participants reported liking that they wrote their own compassionate self-statements. One participant commented on how receiving his own words inspired him to be more responsible.

“You make a statement about how you were going to change your life, and then that statement is thrown right back in your face. You don’t know how to handle it. I mean, it’s coming back over your phone. You got to be more responsible, that’s the way I look at it. I mean, I’m a winner, not a loser.”

Another participants conveyed that the use of their language led to a feeling of empowerment.

“I like the fact that the study is giving me ownership of what I want to create- my own positive, rather than somebody telling me well try this… because sometimes we have our own internal dialogue and we can talk ourselves out of situations.”

Feedback regarding challenges.

Several participants reported feeling confused or put-off by the self-compassion and mindfulness content initially. However, all participants reported appreciating the majority of the content in the exit interviews. In particular, two participants reported being initially confused by informal mindfulness. One participant captured this sentiment.

“At first it was confusing because I didn’t really understand the way you guys were doing things and asking questions… After a while it was okay, because I started to understand it… it’s about feelings and certain things you just don’t think about, like the wind blowing and a smell when you walk… I never actually thought about that until here.. So, now I do that. I really learned about that in this program.”

The majority of participants described the self-report questionnaires as repetitive and difficult to understand. Notably, many of the available self-report questionnaires that are designed to assess constructs involved in the intervention (e.g., self-compassion, mindfulness, internalized stigma) are above an 8^th grade reading level.

Additional feedback.

Several participants reported other life changes after the intervention including increased health-related behaviors.

“I’m eating better, I’m living better, and I’m considering my options. And, I’ve realized that I’m doing nothing but tearing my body up.. You want to stay healthy, and mindful, and consider yourself important, just as important as the next person.”

The majority of participants reported liking the length of the intervention.

“I liked that it was very succinct. It was really super easy to complete.”

Discussion

This proof-of concept study describes the iterative development and preliminary testing of the first intentionally brief text-enhanced intervention we are aware of designed to improve engagement in HIV treatment among people with active substance use disorders by addressing internalized stigma, shame, and other self-conscious emotions. The qualitative interviews with providers who work with this population echoed the existing literature that internalized stigma and shame are under-addressed barriers to engagement in HIV care, particularly among people living with HIV and substance use disorders (Bennett, Traub, Mace, Juarascio, & O’Hayer, 2016). After the intervention development and refinement, participants’ attendance and engagement indicated that this intervention is viable for this hard-to-reach population, with preliminary indications of acceptability and feasibility. Through this proof-of-concept pilot, we demonstrated that this hard to reach population of HIV-positive people with active substance use disorders engaged in the intervention and reported finding it meaningful and useful.

Findings from this proof-of-concept study provide preliminary support for a larger randomized controlled pilot study to assess the feasibility and acceptability of this integrated strategy that incorporates metacognitive awareness and cognitive reframing with HIV self-care goal setting strategies to address our hypothesized pathway (experienced stigma → internalization of that stigma → affective experience of shame, which evokes a desire to avoid reminders of the impetus → behavioral avoidance of reminders of HIV-status, such as substance use and non-adherence to ART), that leveraged the Revised Stress and Coping Theory (Folkman & Lazarus, 1988; Pecoraro et al., 2013). In this work, intervention components targeting metacognitive awareness are considered to be a crucial first step to cultivating insight into underlying cognitive and emotional mechanistic responses to stigma that may contribute to substance use as well as complicate HIV self-care. These metacognitive approaches laid the foundation for becoming aware of negative cognitions related to the self and related emotions, such as shame, that are known to elicit avoidance responses (Tagney & Tracy, 2012). Further, by combining metacognitive awareness with cognitive reframing strategies, informed by self-compassion, we aimed to help participants reframe cognitions related to internalized stigma, including dysfunctional core beliefs. Finally, by linking these skills to goal setting and problem solving, we worked with participants to address the behavioral sequelae of emotions such as shame, that are known to elicit avoidance responses (e.g., ART nonadherence or substance use). Whether this approach achieves meaningful reductions in substance use and improvements in HIV medical outcomes remains an important area for future clinical research.

In addition to the development and refinement of the intervention content and demonstrated proof of concept of the intervention, several lessons were identified that may be useful for future work. First, the innovative text component was reported to be acceptable and feasible via qualitative interviews by this hard to engage population. Two participants had low tech/text-literacy, which identified a need for additional time allocated for assisting participants with low tech/text-literacy in future work. Additionally, several lessons were learned in relation to maintaining participant engagement, including the need to obtain additional contact information from participants, including frequented community-based organizations, names and contact information of individuals likely to know participants’ whereabouts, and social media accounts. While active engagement and retention of this population is effortful, particularly getting participants initially engaged and to complete follow-up visits, all participants who completed the intervention reported finding it useful. Further, these lessons highlighted that more extensive resources are needed to facilitate engagement in the intervention activities and retention of this population in randomized controlled trials. Finally, in future studies, self-report questionnaires will need to be carefully chosen or possibly adjusted to reduce literacy level to be more understandable among this population.

This intervention was iteratively designed to address internalized stigma and shame as a barrier to HIV self-care among people living with HIV and active substance use and the proof-of-concept pilot indicated it was possible, with preliminary indications of acceptability and feasibility. This is the first intervention we are aware of that was iteratively designed based on current theory (Breines & Chen, 2012; Folkman & Lazarus, 1988; Pecoraro et al., 2013; Price, Diana, Smith-Dijulio, & Voss, 2013), evidence based content (Neff & Germer, 2013; Safren et al., 2001; Safren et al. 2009; Witkiewitz & Bowen, 2010), and then iteratively refined based on expert provider and participant feedback. The duration, inclusion of the text component, as well as literacy level and refinement of the content were designed to meet the patient population’s needs, including taking structural barriers into account. While people living with active substance use disorders are thought to be difficult to engage and retain, 8 out of 10 of our participants completed the intervention, participants responded to over 70% of the text messages, and all participants who completed the exit interviews reported finding it meaningful and useful. This study demonstrates that this population will engage in a 5-session, text-enhanced intervention to address internalized stigma and shame to improve HIV self-care and reported that it was useful, meaningful, and worth recommending to friends living with HIV and substance use disorders. Ultimately, stigma and shame remain under-addressed barriers to engagement in HIV care, particularly among people with substance use disorders; there is a need for additional clinical research to refine and test the feasibility and acceptability of this intervention in a randomized controlled trial as well as other novel approaches to optimize the health of this hard to reach and engage population.

Limitations

Our process of developing this intervention and conducting a proof-of-concept study offers an incremental step in leveraging evidence-based cognitive and behavioral strategies to address internalized stigma and shame as barriers to HIV-related self-care behaviors among people living with HIV and substance use disorders, a hard to reach and engage population. While our sample was small, we demonstrated that this population will engage in this integrative intervention, indicating preliminary feasibility. However, the sample size precluded formal pre- and post-comparison of outcome measures. Further, detectable viral load or ART nonadherence was not an inclusion criterion for this study, limiting our ability to see a change in these variables. It is also noteworthy that participants enrolled in the study screened positive for at least two substance use disorders (including alcohol). Although the majority of participants reporting problematic patterns of stimulant and alcohol use, the sample was heterogeneous in terms of types of substance use. However, the focus on cognitive and emotional responses to stigma, including substance use-related stigma, may address a common driver of substance use and difficulties with HIV self-care. Future studies will need to specifically address distinct patterns of polysubstance use and recruit participants who are poorly engaged in care to assess the impact of this intervention.

Conclusions

In this era of HIV treatment as prevention (TasP), addressing barriers to engagement in HIV care for hard to reach populations, such as people with active substance use disorders, is a public health priority. Internalized stigma and shame have been identified as under-addressed barriers to engagement in HIV care among people living with substance use disorders and HIV. The aforementioned intervention was iteratively designed, integrating stakeholder feedback, to address these challenging psychological barriers to engagement in HIV self-care experienced by people actively using substances. This integrative, technology-enhanced intervention was found to be preliminarily feasible and acceptable in a proof-of-concept study involving people living with HIV and substance use disorders, a hard-to-reach and engage population. Future work is needed to assess feasibility and acceptability more conclusively as well as whether this intervention improves HIV health outcomes, including viral load, using a randomized controlled trial design, among a sample with similar co-occurring stigmatized identities (e.g., men who have sex with men who are living with HIV and substance use disorders) who are poorly engaged in HIV self-care behaviors. Ultimately, by testing the intervention among people poorly engaged in HIV-related self-care, we will be able to assess our hypothesized pathway from internalized stigma to avoidance of self-care, which will lead to further intervention refinement and ideally significant improvements in HIV health outcomes and reductions in HIV transmission. Further, these results may lead to insights related to addressing psychological barriers to self-care among people living with other stigmatized identities expanding the clinical utility of this intervention.