Amplification of Positivity Therapy for Co-occurring Alcohol Use Disorder with Depression and Anxiety Symptoms: Pilot Feasibility Study and Case Series

Abstract

Positive valence system dysregulation is a relatively unexplored transdiagnostic mechanism and potential treatment target underpinning alcohol use and anxiety and depression symptoms. The current study examined the feasibility and potential benefit of a behavioral intervention focused on amplification of positivity (AMP) with eight adults (five female) diagnosed with alcohol use disorder and clinically significant depression or anxiety (ClinicalTrials.gov: NCT04278365). AMP for alcohol use (AMP-A) was delivered in 11 individual sessions involving positive activity interventions integrated alongside psychoeducation and alcohol use monitoring. Case descriptions are provided to illustrate treatment implementation. Treatment credibility and acceptability, participant endorsement of the therapy, and homework compliance were rated moderate to high. Exploratory, intent-to-treat analyses suggested medium to large effect sizes for post-treatment improvements in alcohol use, depression, anxiety, and positive affect. Results provide initial evidence of feasibility and acceptability of AMP-A and will be useful for informing future randomized clinical trials to examine clinical efficacy.

Introduction

Alcohol use disorder (AUD) is one of the most prevalent and escalating mental health problems in the United States, exhibiting a 50% increase in prevalence from 2001 to 2017 (Grant et al., 2017). AUD contributes an estimated 5% to global disease burden (World Health Organization, 2019) and 18.4 million years of life lost to disability (Rehm, 2011). Only 20% to 30% of patients exhibit long-term benefits with leading pharmacologic and psychosocial interventions (Carroll & Kiluk, 2017; Mann et al., 2009). The high prevalence of AUD, combined with the negative health consequences and modest treatment efficacy, results in high economic burden, that is, $235 million per year in the U.S. (Rehm et al., 2009).

Approximately 40% to 50% of treatment-seeking AUD patients have co-occurring anxiety or depressive disorders, and 10% to 20% of individuals with anxiety or mood disorders are diagnosed with AUD (Boschloo et al., 2011; Lai et al., 2015). The existence of co-occurring AUD and anxiety or depression has been associated with greater disability, higher healthcare utilization (Teesson et al., 2009), worse prognosis and lower likelihood of responding to treatment (Magidson et al., 2012). This suggests a critical, unmet clinical need and provides an opportunity for novel intervention development for individuals with co-occurring AUD and anxiety/depression symptoms (AUD + ANX/DEP). Most empirically-supported treatments for AUD or ANX/DEP have been developed and tested with a focal (i.e., principal) disorder, often ignoring secondary or co-occurring conditions. Providers are often left trying to target AUD or ANX/DEP separately (e.g., contingency management for AUD; cognitive-behavioral therapy [CBT] for ANX/DEP), often within completely separate service units (Morley et al., 2016). A few integrated treatments have been developed, most of which combine approaches used in AUD and ANX/DEP treatments together (i.e., integrating CBT approaches or motivational interviewing for both conditions). This previous work suggests that integrating treatment approaches into one intervention (Baker et al., 2010; Morley et al., 2016; Sannibale et al., 2013), as opposed to conducting them in parallel (Randall et al., 2001; Riper et al., 2014), may be particularly useful for increasing both efficiency and effectiveness of treatment for co-occurring AUD + ANX/DEP.

One potential avenue for integrated treatment of AUD + ANX/DEP would be to identify overlapping psychological, behavioral, and/or neurobiological treatment targets. Dysregulation of the positive valence system may represent one such transdiagnostic mechanism underpinning AUD and ANX/DEP (Dillon et al., 2014; Koob & Volkow, 2016; Volkow et al., 2010). The positive valence system is conceptualized as involving responses to contexts and situations that are positively motivating or associated with positive affect, including anticipation of and reactivity to rewarding or pleasurable events (Morris & Cuthbert, 2012). This system facilitates acquisition of psychosocial resources promoting resilience to stress and overall health and well-being (Fredrickson, 2013; Lyubomirsky et al., 2005). One hallmark characteristic of AUD is increased neural (e.g., within striatum, orbitofrontal cortex, and amygdala), behavioral, and affective sensitivity to alcohol combined with decreased neural, behavioral, and affective sensitivity to non-alcohol rewards (Diekhof et al., 2008; Kalivas & Volkow, 2005). Current theories of AUD highlight the importance of positive and negative reinforcement, that is, drinking alcohol to enhance positive emotions (e.g., pleasant feelings, social enhancement) or decrease negative emotions or the physical discomfort of withdrawal (Cho et al., 2019; Cooper et al., 1995). Thus, appropriate valuation of and response to reinforcers is likely important for supporting recovery from addiction. Accordingly, reward-related behaviors (e.g., increases in delay discounting, imbalanced monetary allocations) predict worse treatment outcome for AUD (MacKillop & Kahler, 2009; Tucker et al., 2009) and in animal models, the offering of non-drug rewards (e.g., social interaction) reduces drug self-administration (Venniro et al., 2018).

Among the hallmark characteristics of ANX/DEP are exaggerated negative affect, reductions in positive affect, and heightened social withdrawal or avoidance (Dillon et al., 2014; Taylor, Pearlstein, & Stein, 2020)—all of which can contribute to positive and negative reinforcement motivations for alcohol use (Hussong et al., 2011). Thus, positive valence system dysregulation represents a core pathophysiological characteristic of AUD and ANX/DEP, yet is not a focal target of extant treatments. Behavioral interventions specifically targeting positive affect and reward processing in ANX/DEP have recently been developed and shown initial efficacy (Craske et al., 2019; Dunn et al., 2019; Taylor et al., 2017). By upregulating the positive valence system, these interventions also downregulate negative valence processes (e.g., leading to decreases in depression and anxiety)—consistent with the established role of positive emotions in “undoing” the adverse effects of negative emotions (Fredrickson, 2013). The current study builds on the treatment developed by Taylor et al. (2017)—Amplification of Positivity (AMP)—grounded in research demonstrating that nonclinical populations can increase positive thinking, emotions, and behavior by engaging in simple, intentional and repeated activities (e.g., acts of kindness, expressing gratitude) (Lyubomirsky et al., 2005). A waitlist controlled trial revealed that AMP significantly increased positive valence (i.e., positive affect, social connectedness) and decreased negative valence outcomes (i.e., negative affect and clinical symptoms) in patients with ANX/DEP (Taylor et al., 2017; Taylor, Pearlstein, Kakaria, et al., 2020). While some activities in AMP (e.g., planning of pleasurable or meaningful activities, affirming core values) overlap with that of established treatments (e.g., Behavioral Activation [Jacobson et al., 2001], Acceptance and Commitment Therapy [Hayes et al., 2006]) or other more recently tested interventions (e.g., compassion-focused mindfulness interventions [Koszycki et al., 2016]), the majority of activities are unique and focus more specifically on positive affect (e.g., counting blessings, amplifying positive emotions, acts of kindness). Thus, positive affect treatments may hold particular promise for co-occurring AUD and ANX/DEP by enhancing adaptive engagement in social and other meaningful activities and disrupting positive and negative reinforcement cycles.

The purpose of the present work is to describe the initial development, feasibility, and potential clinical utility of an integrated intervention focused on amplification of positivity for AUD (AMP-A) with co-occurring ANX/DEP. For this purpose, eight individuals with AUD + ANX/DEP were enrolled in a study involving completion of 11 individual AMP-A therapy sessions. For feasibility and treatment engagement, we summarize treatment completion rates, patient-reported feedback related to distress, endorsement, and acceptability of the intervention and both patient- and therapist-rated homework compliance. To characterize clinical response, we describe trajectories of alcohol use, anxiety and depression symptoms, positive affect, and social and daily functioning over the course of treatment.

Methods

Participants

The present study utilized a non-randomized design in which eight participants were offered the intervention. Participants were recruited from a database of individuals who had been screened at the research institution for other studies related to ANX/DEP, but who also reported symptoms consistent with AUD. Eligible participants were between 18 and 55 years of age, met DSM-5 criteria for AUD per the Mini International Neuropsychiatric Interview (MINI 7.0) (Sheehan, 2014), reported clinically significant depression or anxiety symptoms on the Patient Health Questionnaire (PHQ-9 [Martin et al., 2006]) ≥10 and/or Overall Anxiety Severity and Impairment Scale (OASIS [Norman et al., 2006]) ≥8, and had sufficient proficiency in the English language to understand and complete study procedures. Exclusion criteria included (1) diagnosis of psychotic or bipolar I disorders, diagnosis of substance use disorder other than AUD or mild cannabis use disorder (latter included due to legality of cannabis use in the state and the frequency of occurrence), or report of plan and intent to attempt suicide within the next month, (2) significant and unstable medical diagnoses or history of moderate to severe traumatic brain injury, (3) magnetic resonance imaging (MRI) contraindications (e.g., metal in body, cardiac pacemaker), (4) positive at baseline for alcohol or illicit drugs of abuse (other than cannabis, given the long half-life and inability to distinguish recent use), determined by urine drug test and alcohol breath test, (5) severity of AUD requiring more intensive treatment (i.e., intensive outpatient or residential), determined by baseline assessments and in reference to American Society for Addiction Medicine (ASAM) criteria (Mee-Lee, 2013), (6) use of psychotropic medications (e.g., stimulants, mood stabilizers, etc.) within 6 weeks prior to enrollment, though use of antidepressants (e.g., selective serotonin reuptake inhibitors) were accepted as long as the dose had remained consistent for 6 weeks, and (7) concurrent engagement in psychosocial treatments, which began within 12 weeks of study enrollment and specifically target AUD or depression and anxiety. Individuals receiving psychosocial treatments for other symptoms, or that were not designed to reduce specific symptoms (e.g., ongoing support groups) were not excluded as long as the frequency of sessions had not changed significantly within 6 weeks prior to enrollment. Many of the inclusion/exclusion criteria listed (e.g., age range, MRI contraindications) were established to allow for functional MRI scans (results not within the scope of the current report). Reasons for exclusion due to a positive alcohol and/or drug test included (1) ensuring that scans were not conducted while the individual was under the influence and (2) the intervention being delivered in an outpatient context and thus wanting to ensure that participants were able to abstain for at least a brief period of time.

Study procedures were approved by Western Institutional Review Board (WIRB). Participants provided written informed consent prior to participation and were compensated for their time completing assessments but not for time spent completing therapy. Participants provided additional written informed consent to publish case series information. Research was conducted ethically in accordance with the World Medical Association Declaration of Helsinki and registered at the US National Institutes of Health (ClinicalTrials.gov: #NCT04278365).

Measures

Participants were asked to complete self-report measures at baseline (within 2 weeks of beginning the intervention), weekly during the intervention, and at post-treatment. Alcohol breathalyzer tests were administered weekly for safety monitoring. See Supplemental Appendix for further description and timing of all measures administered. Below, we describe the main variables of interest for the current report. The primary analyses focused on intervention feasibility and acceptability, with secondary analyses focusing on change in clinical symptoms.

Intervention feasibility and acceptability.

The main variables of interest related to feasibility and acceptability included (1) the Treatment Adherence and Acceptability Scale (TAAS) (Milosevic et al., 2015) administered at Week 1 of AMP-A; (2) the Distress/Endorsement Validation Scale (DEVS) (Devilly, 2004) administered at post-treatment; and (3) completion rate (i.e., completion of all 11 sessions). The TAAS total score reflects greater acceptability and anticipated ability to adhere to the treatment. The DEVS consists of two factors in which higher scores indicate greater distress (e.g., experienced in response to the intervention) and endorsement (e.g., likelihood of recommending the treatment to others), respectively. Exploratory measures related to feasibility of the treatment include the 12-item self-report Homework Rating Scale (HRS) (Kazantis et al., 2004), a 5-point single-item therapist homework compliance rating (HCR), and expectancy and credibility ratings from the Credibility/Expectancy Questionnaire (CEQ) (Devilly & Borkovec, 2000).

Clinical outcomes.

The main clinical variables of interest related to alcohol use included the (1) Timeline Followback (TLFB) (Sobell & Sobell, 1992) and (2) National Institute of Health (Banyard et al., 2004) Patient Reported Outcome Information System (PROMIS; [Broderick et al., 2013; Cella et al., 2010]) measure related to Alcohol Use. The main clinical variables related to affective symptoms included: (1) PROMIS measures related to Depression and Anxiety; (2) Neuro-QoL (Quality of Life in Neurological Disorders) Positive Affect and Well-being (Salsman et al., 2013); and (3) Positive and Negative Affect Schedule—Extended (PANAS-X) (Watson & Clark, 1999). The Sheehan Disability Scale (SDS) (Sheehan et al., 1996) was also included as a measure of general functioning.

The TLFB is a calendar-based self-report measure designed to retrospectively assess daily drinking (i.e., number of standard drinks per day), including abstinent days. For the present study, the TLFB was used to assess drinking over the past 7 days at baseline and post-treatment, as well as at each weekly visit. The NIH PROMIS^® and Neuro-QOL measures were delivered with computer adaptive testing (CAT) at baseline and post-treatment, as well as all 11 weekly sessions. The PANAS-X was administered at baseline and post-treatment and assesses 11 specific affects (e.g., fear, sadness, guilt, excitement, etc.), combined for measures of positive and negative affect. The SDS assesses functional impairment in three major life domains: work, social life/leisure activities, and family life/home responsibilities, with greater scores relating to greater disruption in activities. Due to technical error, the item related to work dysfunction was not administered. We therefore focused instead on the SDS item assessing the number of days lost at school or work, thus providing an indication of the level of work functioning.

Exploratory measures included: (1) Alcohol Craving Questionnaire-Short Form-Revised (ACQ-SF-R) (Singleton, 1997); (2) NIH PROMIS measures for Positive and Negative Expectancies of Alcohol and General Life Satisfaction (Broderick et al., 2013; Cella et al., 2010); (3) Social Connectedness Scale—Revised (SCS-R) (Lee et al., 2001); (4) Snaith-Hamilton Pleasure Scale (SHAPS) (Franken et al., 2007), and (5) a modified version of the Michigan Nicotine Reinforcement Questionnaire (MNRQ) (Pomerleau et al., 2003). Descriptions and results for these measures are reported in Supplemental Appendix A and Table A1

Intervention

Following completion of baseline assessments, participants were asked to complete 11 AMP-A sessions, conducted once or twice weekly based on participant preference. Sessions were scheduled for 1 hour with the exception of Sessions 1 and 2, which were scheduled for 1.5 hours each due to the amount of content to be covered. Each session was completed one-on-one with a therapist. Table 1 shows an outline of AMP-A content, which involved positive emotion enhancement exercises established in prior studies (Taylor et al., 2017). We used the protocol developed by Taylor et al. (2017) with modifications to specifically address alcohol use, based on previous work (Wolitzky-Taylor et al., 2018). The intervention took a harm reduction (rather than an abstinence) perspective (Marlatt & Witkiewitz, 2002), with each client developing their own individualized treatment goals. The general structure of each session followed standard CBT regimens: (1) meet with clinician to review the prior week’s exercises (e.g., self-monitoring forms of emotion, alcohol use, and positive activity exercises); (2) identify and troubleshoot issues that arose during exercise completion; (3) introduce material about a new positive emotion enhancement activity; (4) identify concrete exercises to implement for the upcoming week. Treatment implementation was integrative, such that positive activity exercises were presented and discussed alongside the participant’s ability to manage alcohol cravings and behaviors. The intervention was delivered by licensed doctoral or master’s level clinicians or a therapist in training (i.e., clinical psychology post-doctoral fellows or graduate students). Each therapist completed a full day in-person or video-recorded workshop led by CTT, followed by a half-day workshop led by CTT. Each therapy session was video and audio recorded for review by the on-site supervisor (RLA) and to support weekly consultations with the developers of AMP-A and an expert in CBT with co-occurring substance use disorders and ANX/DEP (KWT).

Table 1.

Summary of AMP-A Session Content.

Session	Summary of content covered
1	Overview of intervention, enhance motivation for change. Elicit treatment goals related to alcohol use, positive affect, and/or anxiety/depression.
Throughout all remaining sessions, homework is reviewed and client is expected to maintain completion of alcohol diaries, mood self-monitoring forms, and monitoring of positive events.
2	Downward spirals of negative emotion; upward spirals of positive emotions. Role of alcohol use in the emotion spirals. Use of non-judgmental observation of cravings.
3	Capitalizing on positive events (e.g., savoring, telling someone about it; writing about it) and counting one’s blessings.
4	Counting one’s blessings and acts of kindness
5	Acts of kindness; three paths to positivity (pleasurable, engaging, meaningful activities).
6	Three paths to positivity, noticing and using personal strengths
7	Using your strengths and combining this with other positive affect activities; imagining best possible future
8	Imagining best possible future; focus on making someone else happier
9	Using strengths to make someone else happier; live like it is your last few weeks in this area.
10	Live like this is the last few weeks in this area; writing a gratitude letter to someone. Develop a personalized treatment plan through discussing person-activity fit
11	Review progress through treatment and personalized treatment plan; prepare for continuation of change

Data Analysis

In addition to individual case series information, we provide descriptive statistics concerning reports on the CEQ (collected at baseline), DEVS (post-treatment or at withdrawal), AAS (week 1 of treatment), HRS, and HCR (averaged across weeks 2–11 of the intervention). All statistical analyses were conducted using R statistical package (R Core Team, 2013). Quantitative analyses with clinical outcome measures were conducted to explore clinical potential and provide estimates of effect sizes for future work. For each variable administered weekly (TLFB, PROMIS Alcohol, Depression, Anxiety; Neuro-QOL Positive Affect and Well-being, SDS, ACQ), we conducted linear mixed effects models using R package lme4 (Bates et al., 2015) examining time effects (including baseline, weekly, and post-treatment time points), with subject as a random effect. For all variables, we conducted Wilcoxon signed-rank tests using R package stats (R Core Team, 2013) comparing baseline and post-treatment scores for (a) treatment completers (those who completed all 11 sessions) and (b) the intent-to-treat sample, with the last observation carried forward (LOCF), providing 95% confidence intervals. Estimates and 95% confidence intervals for effect size (r) were also calculated for each Wilcoxon signed-rank test (R package rcompanion) (Mangiafico, 2020).

Results

Treatment Feasibility and Acceptability

A total of eight participants with AUD enrolled in the study. Table 2 provides information concerning demographics and co-occurring diagnoses, as well as means and standard deviations for variables related to feasibility and intervention acceptability. Of the eight participants who began AMP-A, one withdrew after the first session (citing family issues as reason for withdrawal) and another withdrew after Session 6 (citing that she was no longer symptomatic), and the remaining six participants completed all 11 sessions. One participant completed Session 2 over two meetings with the therapist due to not completing all content within one session; this was due to a combination of the participant starting the session approximately 10 to 15 minutes late and due to difficulties covering the amount of information included in that session. For the six completers, average duration for completion of all 11 sessions was 13.7 weeks. As shown in Table 2, all participants reported relatively moderate levels of acceptability and likelihood of adherence on the TAAS (M = 40.14; possible range: 10–70), which was similar to that reported in previous work (M = 42.94 for a two-session exposure intervention [Milosevic et al., 2015]). Participants also reported little distress (M = 22.71; possible range: 7–63) and moderate to high levels of endorsement (M = 23.43; possible range: 3–27) on the DEVS, which were similar or slightly favorable compared to previous work (M = 26.50–30.31 for distress and M = 22.25–22.41 for endorsement (Devilly, 2004); possible range: 10–90). Homework compliance was rated moderate to high, as reported by participants (HRS, Quantity of homework completed; M = 2.89; possible range 0–4) and therapists (M = 3.89; possible range 0–5). On the CEQ, participants’ ratings of intervention credibility (M = 6.50; possible range: 3–27) and expected benefit from the intervention (M = 57.50%; possible range: 0%–100%) were similar to that previously reported for cognitive-behavioral therapy (M = 7.47 for credibility, M = 67.90 for expectancy; (Thompson-Hollands et al., 2014).

Table 2.

Individual Case Information Relevant for Demographics, Diagnoses, and Treatment Compliance, Credibility, Expectancy, Acceptability, and Endorsement.

	Case 1	Case 2	Case 3	Case 4	Case 5	Case 6	Case 7	Case 8	Mean (SD)
Age	25	29	27	42	28	26	30	34	30.72 (5.59)
Sex	M	M	F	F	F	M	F	F
MINI diagnoses	AUD	AUD	AUD	AUD	AUD	AUD	AUD	AUD
	MDD	CUD	MDD	MDD	CUD	CUD	MDD	MDD
		MDD		GAD	MDD	MDD	SAD
		Panic		Panic	SAD	SAD
		GAD
		SAD
Sessions completed	11	11	11	11	11	1	11	6	9.25 (3.41)
Therapist homework compliance rating	3.33 (0.49)	4.0 (0.71)	3.6 (0.84)	4.0 (0.0)	3.5 (0.53)	—	4.42 (0.51)	4.0 (0.71)	3.78 (0.67)
Participant Homework Rating Scale	3.1 (0.57)	3.2 (0.92)	1.6 (1.17)	3.4 (0.84)	3.3 (0.67)	—	3.0 (0.47)	3.2 (0.92)	2.89 (0.97)
CEQ credibility	5.67	9	7	5.67	5.67	5.67	6.33	7	6.5 (1.17)
CEQ expectancy	30%	100%	70%	50%	40%	30%	70%	70%	57.5% (24.35)
DEVS distress	11	18	27	11	45	—	17	30	22.71 (12.23)
DEVS endorsement	27	27	14	27	25	—	27	17	23.43 (5.53)
TAAS	40	37	43	42	35	41	42	48	40.14 (3.93)

Note. MINI = MINI International Neuropsychiatric Inventory; CEQ = Credibility and Expectancy Questionnaire; DEVS = Distress/Endorsement Validation Scale; TAAS = Treatment Adherence and Acceptability Scale; AUD = alcohol use disorder; CUD = cannabis use disorder; MDD = major depressive disorder; Panic = panic disorder; GAD = generalized anxiety disorder; SAD = social anxiety disorder.

Clinical Outcomes

Results from linear mixed effects analyses including all eight participants suggested significant improvements over time in TLFB average drinks per day (F[12, 67] = 2.20, p = .020) and alcohol craving (ACQ; F[12, 67] = 4.44, p < .001), but not the number of drinking days (F[12, 67] = 1.53, p = .136). Improvements were observed across PROMIS Alcohol Use (F[12, 67] = 5.70, p < .001), Depression (F[12, 67] = 4.59, p < .001), and Anxiety (F[12, 67] = 4.79, p < .001), Neuro-QOL Positive Affect and Well-being (F[12, 67] = 2.65, p = .006), and SDS social functioning (F[12, 68] = 6.00, p < .001), family functioning (F[12, 67] = 4.25, p < .001), days lost at work (F[12, 67] = 4.09, p < .001), and days unproductive at work (F[12, 68] = 2.22, p = .020). Average scores over time are displayed in Figure 1. Baseline and post-treatment means and results from the associated Wilcoxon signed-rank tests (for both completer and intent-to-treat samples) are shown in Table 3. Medium to large effects were identified for all outcomes.

Figure 1.

Group averages on measures relating to alcohol use (as measured by the Timeline Followback [TLFB], left panel), depression and anxiety (as measured by NIH PROMIS measures, middle panel), and positive affect (measured by NIH PROMIS measures) and social functioning (measured by the Sheehan Disability Scale, right panel).

Table 3.

Means, Confidence Intervals, Effect Sizes, and Main Analyses of Change from Baseline to Post-Treatment for Treatment Completers (N = 6) and the Intent-to-Treat (ITT; N = 8) Samples.

Measure	Sample	Pre Mdn (range)	Post Mdn (range)	Mdn difference: 95% Cl	Z	p Value	r [95% Cl]
TLFB drinks/day	Completed	2.15 (0.71–3.86)	0.86 (0.57–2.29)	−1.00–2.86	1.42	0.156	0.64 [0.13–0.92]
	ITT	1.43 (0.43–3.86)	0.86 (0–2.29)	−0.43–2.15	1.60	0.110	0.59 [0.05–0.90]
TLFB number of drinking days	Completed	4.50 (1–7)	2.0 (1–5)	0.00–4.00‡	1.62	0.105	0.70 [0.14–0.95]
	ITT	3.00 (1–7)	2.0 (0–6)	−1.00–4.00	1.35	0.178	0.50 [0.05–0.91]
PROMIS alcohol use	Completed	61.20 (57.2–65.1)	56.15 (40.2–59.1)	1.20–21.20	2.15	.031	0.90 [0.90–0.92]
	ITT	60.65 (57.2–65.1)	56.15 (40.2–61.4)	1.20–15.10	2.06	0.039	0.74 [0.30–0.90]
PROMIS depression	Completed	63.25 (55.0–67.7)	52.53 (39.4–58.9)	8.80–16.0	2.15	0.031	0.90 [0.90–0.92]
	ITT	64.10 (55.0–67.7)	52.30 (39.4–67.7)	4.50–16.00	2.42	0.015	0.84 [0.60–0.90]
PROMIS anxiety	Completed	67.05 (59.8–69.4)	57.25 (42.1–65.4)	4.00–17.70	2.15	.031	0.90 [0.90–0.92]
	ITT	67.20 (59.8–70.9)	57.25 (42.1–71.6)	1.50–15.30	2.27	0.023	0.80 [0.45–0.90]
Neuro-QOL positive affect	Completed	48.4 (44.4–58.1)	53.80 (46.9–59.9)	−8.35–2.70§	−1.89	.059	0.86 [0.70–0.92]
	ITT	46.90 (40.9–58.1)	53.80 (41.1–59.9)	−9.38–0.56	−1.77	0.076	0.67 [0.15–0.90]
PANAS-X positive†	Completed	28 (22–34)	30.0 (22–43)	−7.00–2.00	−1.16	0.248	0.52 [0.04–0.91]
	ITT	26.50 (14–34)	—	—	—	—	—
PANAS-X negative†	Completed	20 (10–30)	14.5 (10–29)	−2.50–9.00‡	1.26	0.207	0.56 [0.13–0.93]
	ITT	19.75 (10–30)	—	—	—	—	—
SDS social	Completed	6.50 (3–7)	2 (0–5)	2.00–7.00‡	1.89	.058	0.86 [0.70–0.93]
	ITT	6.50 (3–7)	2.00 (0–7)	3.00–6.00‡	2.12	0.034	0.83 [0.70–0.91]
SDS family	Completed	5.50 (2–8)	2 (0–5)	2.00–5.00‡	1.90	.058	0.86 [0.70–0.93]
	ITT	5.00 (2–8)	2.00 (0–7)	2.00–4.00‡	2.11	.035	0.83 [0.70–0.91]
SDS days lost	Completed	1.0 (0–3)	0 (0–0)	1.00–2.00	1.92	.054	0.87 [0.70–0.95]
	ITT	1.50 (0–3)	0 (0–3)	0.00–2.50‡	1.62	0.105	0.63 [0.13–0.91]

Note. TLFB = Timeline Followback; PROMIS = patient reported measurement information system; Neuro-QOL = neurology quality of life; SDS = Sheehan Disability Scale; ACQ = Alcohol Craving Questionnaire; SHAPS = Snaith-Hamilton Pleasure Scale; PANAS = positive and negative affect schedule.

^†Measures only administered at baseline and post-treatment and thus, intent-to-treat analyses were not completed. For all ITT analyses, we used last observation carried forward (LOCF) for post-treatment scores.

^‡90% CI (95% confidence interval could not be computed).

^§80% CI (95% confidence interval could not be computed).

Individual Case Series

Trajectories for TLFB average drinks per day (assessed over the past week), PROMIS Depression, Anxiety, and Positive affect are provided in Figure 2 for all seven participants who completed more than one session of AMP-A. A qualitative description of Case 2 is provided herein, with the remaining cases described in Supplemental Appendix A.

Figure 2.

Trajectory of individual case responses on measures relevant for depression, anxiety, and positive affect and well-being (measured via NIH PROMIS scales) and alcohol use (as measured by the Timeline Followback [TLFB]).

Case 2 is a 29-year-old Caucasian who identifies as male with a high school education, who was currently enrolled in an electrician training program. He reported living with one roommate, working part time as a server at a local restaurant, and being in a long-term committed relationship. He described his girlfriend and friends as his main source of social support. He was currently diagnosed with major depressive disorder, single episode in partial remission; panic disorder, current; cannabis use disorder, mild; AUD, moderate; generalized anxiety disorder; and social anxiety disorder. Case 2 reported near daily use of alcohol for approximately 8-years prior to starting treatment, with pre-treatment use reported at approximately 1–3 drinks per day.

Case 2 described himself as a slightly pessimistic person and a realist, explaining that he viewed optimism as slightly naïve. He expressed being fairly skeptical of talk-therapies generally and was unsure how it would be beneficial. However, he indicated he was willing to engage in the protocol with an open mind. His stated therapy goals were: (1) only drinking once per week, (2) being able to stop drinking when he wants, (3) exercising more, including hiking or biking once per week, (4) making an adventurous plan with his significant other once per week, and (5) taking an hour a week to engage in drawing or another hobby. Case 2 reported daily cannabis use, which he indicated using for medical purposes, particularly for anxiety. He indicated that the was open to discussing his cannabis use but was not interested in reducing it.

Case 2 demonstrated quick success integrating positive events tracking into his routine. He reported that his favorite ways to amplify positive events were to reflect on them through journaling or to talk about the events with friends or his significant other. Specifically, he reported enjoying spending time on art-related hobbies and offering meals to individuals who were homeless. Early in treatment, he often described his positive emotions as “feeling pretty good,” but was open to expanding his vocabulary in this regard. Subsequently, he often identified feelings of happiness, pride, and fulfillment, among others.

With respect to alcohol monitoring, Case 2 reported that it was helpful to track his expectations of drinking in an effort to boost positive affect, and for identifying the contexts in which he was likely to engage in excessive drinking. He indicated that this helped him to make plans ahead of time to mitigate the likelihood of drinking more than planned. He noted specific success with non-judgmentally observing the craving come and go. Because the monitoring of alcohol use seemed to be helpful for Case 2, the therapist suggested he could also track his use of cannabis. He was agreeable to doing so and noticed he was primarily using cannabis to cope with anxiety. However, he expressed no desire to change his cannabis use early in treatment.

Case 2 reported enjoying planning positive activities. In session 7, he reported that he had previously viewed positive activities as something he should do alone to relax. However, he came to realize through treatment that he enjoyed incorporating social aspects into his positive activities. Specifically, he reported enjoying game night with friends and becoming involved with a political advocacy group. He also reported individual positive activities, such as savoring the view of downtown and drawing in his sketch book. He reported that the gratitude exercise of counting his blessings seemed to lose potency after a few weeks, but that writing a gratitude letter was more powerful and beneficial for him. With respect to using his strengths, Case 2 indicated that two of his strengths were love of learning and curiosity. He reported his love of learning is the driving motivation behind pursuing his education as an electrician. On a related note he indicated that a specific instance of using his curiosity was building a solar powered windmill as a project at school. Case 2’s best possible future exercise focused on the social domain in line with his experience incorporating social elements to his positive activities reported above. His best possible future included spending time weekly with friends to play games and enjoy each other’s company and the exercise was a narrative of one such game night.

At the last session, Case 2’s goals were reviewed and discussed. He noted that he was able to meet his goal of drinking alcohol only once per week. Although he never specified a goal related to cannabis use, it was observed that he had reduced his use to once per week. He reported that due to experiencing less anxiety and negative affect, his need for using alcohol and cannabis had decreased. He expressed being pleasantly surprised by how helpful the treatment protocol had been for him. He noted that he still would not describe himself as an optimist, but that he was looking forward to positive aspects about his future. He stated he would like to continue (1) self-monitoring and writing in his alcohol diary, (2) savoring positive events, (3) incorporating others into pleasurable and meaningful activities, and (4) making others happier.

Discussion

The current study examined the feasibility and potential clinical utility of a psychosocial intervention focused on amplification of positivity for individuals suffering from AUD with co-occurring ANX/DEP. Results suggested that the intervention is both feasible and acceptable to patients. Further, there was evidence the intervention has potential for improving symptoms of alcohol use, depression, and anxiety, and for improving positive affect and daily function.

Feasibility is evidenced first and foremost by the fact that seven of the eight participants completed the majority of the intervention, which is encouraging given that intervention research on co-occurring psychiatric and substance use disorders tends to report higher rates of non-compliance and treatment dropout as compared to psychiatric disorders alone (Herbeck et al., 2005). Overall, participants rated the intervention favorably, indicating that it was credible prior to starting, reported experiencing very little distress during the intervention, and indicated they were likely to recommend the intervention to others. In addition, the treatment providers anecdotally cited the ease and enjoyment of implementing the intervention, specifically discussing how the intervention allowed them to focus on the values and strengths of each client. One potential modification to future uses of AMP-A would be to extend the intervention to 12 sessions, as both providers and participants provided feedback that Session 2 was rather content heavy and could be broken into two distinct sessions.

Quantitative analyses suggest large effect sizes for improvements from baseline to post-treatment in relation to alcohol use (e.g., average drinks per day, number of drinking days, problematic alcohol use), positive and negative affective symptoms (e.g., depression, anxiety, and positive affect) and overall functioning (e.g., social and family functioning), within both completer (N = 6) and intent-to-treat (N = 8) samples. The observed improvements in symptoms and alcohol use and the positive feedback from participants and providers, point to the clinical potential of this intervention. As such, the present findings support previous research suggesting the potential value of directly targeting the positive valence system for enhancing positive affect and decreasing anxiety/depression symptoms (Craske et al., 2019; Dunn et al., 2019; Taylor et al., 2017). Results extend this previous work to indicate that by doing so, such interventions may also impact negative and positive reinforcement cycles in order to reduce reliance on alcohol use.

It is important to recognize that effect sizes observed in the present study may be inflated due to the small sample size and the lack of a comparison intervention (Szucs & Ioannidis, 2017). Randomized clinical trials comparing this intervention to other interventions often used for AUD + DEP/ANX are warranted. Given that current empirically supported psychosocial treatments for AUD are associated with relatively small effects (Magill & Ray, 2009), it is crucial that we follow through on promising treatments that have the potential to improve outcomes. There are other interventions that have overlapping targets with AMP-A that could be considered for AUD. Specifically, Behavioral Activation (BA) targets engagement in pleasurable or meaningful activities to increase opportunity for positive reinforcement. BA is considered an efficacious treatment in depression (Dimidjian et al., 2006; Jacobson et al., 1996) and has also recently shown some promise in the treatment of substance use disorders (Daughters et al., 2018). However, the little evidence that exists suggests that BA has modest effects on positive affect and related clinical outcomes (i.e., anhedonia; (Dichter et al., 2009; Moore et al., 2013). In addition, unlike AMP-A, BA does not specifically incorporate empirically supported strategies from the positive psychology literature on enhancing positive affect (e.g., savoring, expressing gratitude, making someone else happier, etc.).

Limitations of the current study include the small sample size and lack of a comparison condition. In addition, the current study focused solely on an outpatient population. Future studies are needed in large randomized clinical trials to enable examination of clinical outcomes and the proposed mechanisms of action (e.g., including increased reactivity to non-drug rewards and/or diminished negative valence reactivity). The current study only focused on alcohol use and clinical symptoms immediately post-treatment. Longer-term follow-ups would be preferable in future clinical trials to ascertain maintenance of gains. In addition, there were a number of participants in the current sample that reported frequent cannabis use in addition to alcohol. While use of other recreational drugs was excluded, it was difficult to completely exclude for cannabis use given that use is legal and can be prescribed within the state in which this study was conducted. Future research would benefit from more specifically assessing the impact of AMP-A on cannabis use, particularly given that reductions in use were noted for the case example described herein. Lastly, the results of the present study are limited to individuals who were able to abstain from alcohol use for a brief period of time prior to intervention (required for study inclusion) and did not require medical or residential treatment for AUD-related symptoms. Thus, future studies may need to assess the feasibility of including AMP-A as an adjunctive treatment within other contexts and settings. However, the current sample is likely reflective of co-occurring AUD and ANX/DEP often presenting to community mental health clinics and thus, current results highlight the feasibility and potential for its utility in such settings.

Conclusions

AMP-A was feasible to implement, acceptable to participants, and associated with improvements in alcohol use, depression, anxiety, and positive affect, as well as social, family, and work functioning. Randomized clinical trials are warranted to further assess the potential clinical benefit and mechanisms of action for AMP-A.

Author Biographies

Elisabeth Akeman, M.S., LPC-C, is a research therapist and study coordinator at the Laureate Institute for Brain Research. Her research interests include enhancing treatment and prevention efforts to improve well-being, as well as better understanding factors related to risk and resilience.

Evan White, Dr. White is an Associate Investigator at Laureate Institute for Brain Research. His research interests include: (1) advancing neuroscientific understanding of cultural factors that are protective against poor mental health among American Indians and (2) employing neuroscience and psychophysiological methods to understand development and treatment of anxiety and mood disorders.

Kate Wolitzky-Taylor, Dr. Wolitzky-Taylor is Associate Faculty of the UCLA Anxiety and Depression Research Center. Her research interests include investigating mechanisms of change during behavioral treatment for anxiety disorders, increasing access to CBT for anxiety disorders in community settings, and understanding and treating comorbid anxiety and substance use disorders.

Jessica Santiago, M.Ed., LPC, is a doctoral student in Counseling Psychology at Oklahoma State University. Her interests include supporting minority clients in overcoming academic and professional obstacles, engaging in research to help increase understanding of the stressors impacting minority populations, and developing support for such populations through outreach and advocacy.

Timothy J. McDermott, M.A., is a doctoral student in Clinical Psychology at the University of Tulsa. His research interests include understanding the behavioral and neural correlates of cognitive control deficits and how these relate to emotion regulation in individuals with PTSD, depression, anxiety, or substance use disorders.

Danielle C. DeVille, is a doctoral candidate in clinical psychology under the mentorship of Dr. Robin Aupperle at the Laureate Institute for Brain Research and the University of Tulsa. Her research focuses on the neurobiological, interoceptive, and psychosocial correlates of suicidal thoughts and behaviors.

Jennifer L. Stewart, Dr. Stewart is a Principle Investigator at Laureate Institute for Brain Research. Her work employs subjective reports, behavioral methods, electro-encephalography, event-related potentials, and functional magnetic resonance imaging to investigate how brain patterns linked to cognition, emotion, and their interaction intersect with individual differences in substance use, depression, and anxiety disorders.

Martin Paulus, Dr. Paulus is the Scientific Director at the Laureate Institute Brain Research. His research includes using neuroimaging to develop predictive biomarkers for anxiety disorders and addictive disorders, as well as using computational psychiatry to better quantify the behavioral dysfunctions in individuals with mood, anxiety, and addictive disorders.

Charles T. Taylor, Dr. Taylor is an Associate Professor in the Department of Psychiatry at the University of California San Diego. His research interests include transdiagnostic dimensional approaches to understanding positive and negative valence system functioning, as well as the development of empirically supported interventions to enhance positive emotions, social connections, and well-being.

Robin L. Aupperle, Dr. Aupperle is a Principle Investigator at Laureate Institute for Brain Research. Her research focuses on (1) using neurocognitive methods to enhance understanding of the development and maintenance of anxiety and depression and (2) how that knowledge may be used to enhance treatment and prevention efforts.