Development of a Storytelling Video Self-Help Intervention based on Acceptance and Commitment Therapy for Major Depression: Open Trial Results

Abstract

This study involved the initial development and testing of a video self-help intervention called LifeStories, which features real patients describing their use of coping strategies for depression based on Acceptance and Commitment Therapy. We conducted a baseline-controlled open trial (AB design) of 11 individuals diagnosed with major depressive disorder. Participants reported high levels of satisfaction and transportation (i.e., engagement) after watching LifeStories. No significant changes were observed during the 4 week baseline period in terms of interviewer-rated depression severity (primary outcome), but a significant and large effect size improvement was observed at week 8 post-intervention. The majority of participants (54.5%) showed a reliable and clinically significant post-treatment response. Significant improvements also were observed during the intervention period only for self-reported depressive symptoms and aspects of mindfulness (nonreactivity). Qualitative data analysis of participant interviews identified additional areas for improvement and refinement. Future testing in a randomized trial is warranted based on these encouraging results.

Introduction

According to recent U.S. estimates, 6.6% of adults have had a recent major depressive episode (Kessler, Berglund, Demler, Jin, Koretz et al., 2003), resulting in $2.1 billion in direct and $4.2 in indirect costs in the U.S. alone (Jones & Cockrum, 2000). A variety of effective treatments currently exist (Davidson, 2010), but up to two thirds of patients do not respond to initial treatment and one third will not respond to multiple treatments (American Psychiatric Association, 2010). Although there are efficacious psychosocial treatments for depression, including interpersonal therapy, traditional cognitive behavioral therapies (CBT), and newer approaches such as Acceptance and Commitment Therapy (ACT) (Hayes, Strosahl, & Wilson, 2012), many people do not have access to empirically-supported psychotherapies because of difficulties with cost/insurance coverage, inability to attend appointments during business hours, and the lack of providers in their area who are trained to provide these treatments with fidelity (Mohr, Hart, Howard, Julian, Vella et al., 2006). Also, 20% of depressed patients endorse emotional barriers to psychotherapy, including privacy concerns and stigma (Mohr et al., 2006). Antidepressants may be somewhat easier to disseminate with fidelity than empirically supported psychotherapies; however, they are not without limitations, including unacceptability, nonadherence, and intolerable side effects (Van Schaik, Klijn, Van Hout, Van Marwijk, Beekman et al., 2004). Further, only 30–45% of patients who take an antidepressant remit after the first line of therapy; the remainder either improve but still experience clinically significant symptoms, or do not improve much at all (Carvalho, Cavalcante, Castelo, & Lima, 2007). There is evidence that combined treatment (e.g., psychotherapy plus medication) is superior to monotherapy for groups such as treatment-resistant primary care patients and those with severe depression (Wiles, Thomas, Abel, Ridgway, Turner et al., 2013). Although patients may benefit, combined treatment with psychotherapy and medication brings with it the same limits, barriers, and concerns discussed above.

A total of 8–12% of all primary care visits are made by people with a depression diagnosis (Hing & Uddin, 2010), where the provision of evidence-based psychotherapy is limited (Uebelacker, Wang, Berglund, & Kessler, 2006). There are many advantages to locating treatment for depression in primary care settings: the primary care team can take into account physical illnesses; stigma may be reduced; and patients prefer going to a familiar setting (Fleury, Imboua, Aube, Farand, & Lambert, 2012). However, there are also difficulties with treating mental health problems in primary care, including not having the time needed to treat both mental and physical health problems (Fleury et al., 2012). Collaborative care models that address some of these problems involve optimization of antidepressant medications (Katon, Unutzer, Wells, & Jones, 2010) and some include psychotherapy, but it continues to be difficult to have resources to provide psychotherapy in primary care.

In an effort to address dissemination of treatments for depression, there has been a recent proliferation of self-help forms of CBT. Whereas traditional self-help has relied heavily on the bibliotherapy format, newer technologies are increasingly being used in an effort to improve dissemination. Computerized, internet-delivered, or mobile self-help interventions, mostly based on CBT, have shown superiority to wait list control or treatment as usual for treating depression and anxiety in randomized controlled trials (Andrews, Cuijpers, Craske, McEvoy, & Titov, 2010). A meta-analysis of 5 trials comparing computerized CBT to face-to-face CBT across different psychiatric disorders failed to find differences in treatment efficacy or satisfaction (Andrews et al., 2010). However, self-help CBT interventions are not without limitations, including uptake, adherence, and completion rates that can be very low (Waller & Gilbody, 2009). Reasons cited for low adherence include time constraints, lack of motivation, technical or computer access problems, as well as perceived lack of effectiveness, improvement in symptoms, and program burden (Christensen, Griffiths, & Farrer, 2009). Waller and Gilbody also noted that in the studies of computerized CBT, the average education and employment level was higher than in the general population, and some individuals perceived these interventions as “cold” and unappealing. Therefore, although computerized CBT might improve reach beyond traditional psychotherapies, there is still a need to develop other ways of disseminating effective behavior change principles, particularly for more diverse populations.

Self-help CBT approaches typically emphasize scientific/logic-based ways of knowing and psychoeducation. An alternative approach that shows promise for improving intervention acceptability and engagement is to capitalize on so-called “narrative” ways of learning that rely more on storytelling and experience instead of the didactic presentation of material (Hinyard & Kreuter, 2007). Narrative communication has been defined as “any cohesive and coherent story with an identifiable beginning, middle, and end that provides information about scene, characters, and conflict; raises unanswered questions or unresolved conflict; and provides resolution” (p. 778, Hinyard & Kreuter). In addition, Green (2004) defines transportation as “an integrative melding of attention, imagery, and feelings, focused on story events” (p. 247). The use of personal narratives that are vivid and absorbing helps the viewer to be “transported” into the narrative world of the storyteller, where persuasion can more easily occur and thus influence a person’s behavior (Green et al., 2000). For transportation to occur, the viewer should perceive the narrative as authentic and relevant to the real world and the storyteller as similar to the viewer (Kreuter, Green, Cappella, Slater, Wise et al., 2007). This is consistent with social learning theory in that narrative approaches rely on observing and modeling the behavior of similar others (Bandura, 1977).

Such narrative or storytelling interventions have been used to change important and complex health behaviors (Larkey, Lopez, Minnal, & Gonzalez, 2009). For example, in a randomized controlled trial (RCT), Houston et al. (2011) demonstrated significant improvements in blood pressure amongst African Americans patients with hypertension who viewed videos in which African American individuals told their own stories about health behavior changes they made related to improving their hypertension based on recommended guidelines. For individuals with uncontrolled hyptertension at baseline, the difference between the two groups in change in blood pressure after 3 months was similar in magnitude to that seen in studies of antihypertensive medication.

Acceptance and Commitment Therapy (ACT) (Hayes et al., 2012) is a newer “third wave” CBT approach found to be effective in self-help administration that may be more naturally aligned with a narrative approach to treatment. ACT fosters people’s ability to engage in functional activities by clarifying their personal values, developing values-consistent behavioral goals, accepting and being willing to experience previously avoided thoughts and feelings in the service of these values, and fostering present moment, nonjudgmental awareness of internal experiences (i.e., mindfulness). To achieve its aims, ACT uses a variety of strategies, including the presentation of vivid metaphors/stories to communicate key treatment principles. These narrative-based strategies make the treatment more understandable, engaging, and memorable. Furthermore, ACT is listed as an empirically-supported therapy by the American Psychological Association for various conditions, including depression, mixed anxiety, and chronic pain (American Psychological Association, n.d.). Meta-analyses have found that ACT is at least as effective as more traditional forms of CBT for a wide variety of conditions, including depression (A-Tjak, Davis, Morina, Powers, Smits et al., 2015; Powers, Zum Vörde Sive Vörding, & Emmelkamp, 2009). Furthermore, clinical trials have demonstrated that ACT is efficacious for changing health behaviors and decreasing depression when provided in unguided, self-help administration (Buhrman, Skoglund, Husell, Bergstrom, Gordh et al., 2013; Fledderus, Bohlmeijer, Pieterse, & Schreurs, 2012).

In the current article, we describe an iterative and participatory research approach in which we used key behavior change principles from an evidence-based approach for depression (i.e., ACT) to develop a brief, video-based narrative/storytelling intervention designed to be especially appropriate for primary care settings, where need for psychosocial treatment is high but access low. As noted above, we believe that a narrative-based approach has the potential to increase patients’ acceptability of and engagement with traditional self-help interventions, which could help to decrease drop-out and even “prime the pump” for future engagement in more traditional psychological interventions when needed. Thus, we used storytelling to communicate the following key processes from ACT that are thought to be associated with optimal health and well-being: 1) values clarification: improving consistency between personal values and daily actions; 2) acceptance: being willing to experience negative thoughts/feelings in the service of personal values; 3) awareness of the transience of mental events: understanding that negative thoughts/feelings change over time; and 4) present moment awareness: being more fully present (or mindful) in everyday life. In partnership with a video production company, we produced four half hour episodes (i.e., 2 hours) of a storytelling intervention we call LifeStories. LifeStories features real people who discuss ACT-consistent behavior change strategies that have improved their depression. We describe our process for developing the LifeStories intervention, including: 1) initial interviewing and selection of patients to appear in the videos; 2) professional videorecording and editing of patients’ narratives; 3) initial focus group testing of storytelling video clips; 4) finalizing of the complete video series with professional commentary; and 5) completion of an initial baseline-controlled (AB) open trial to examine the feasibility, acceptability, and preliminary effects of LifeStories on the primary outcome (depressive symptoms) and potential target mechanisms (acceptance, mindfulness, values, and behavioral activation).

Method

Initial interviews about people’s experiences with depression

All of the following procedures were approved by the institutional review board of the hospital. First, we interviewed participants about their personal experience with depression to identify potential candidates to be included in our video intervention. We recruited participants from primary care sites and other local clinical settings. Eligibility criteria were: (a) currently under the care of a primary care provider (at least one appointment in past two years); (b) 18 years or older; (c) some lifetime experience of depression, based on self-report; and (d) had had some success in coping with depression using ACT-consistent strategies.

We recruited a total of 33 individuals to participate in an initial, videotaped interview at the research clinic. After consent was obtained, a member of the research team administered an interview (approximately 45 minutes) discussing: (a) the participant’s history of depression and treatment; and (b) the ACT-relevant strategies the participant used to cope successfully with depression (e.g., mindfulness, psychological acceptance, values-consistent actions). Patients could have used these ACT strategies for coping with depression at any time in their life. Sample questions from the interview include: “Tell me about your experiences coping with depression.” “Have you discovered these strategies through formal treatment, self-help, talking to important others in your life, readings, or other ways?” “Have you used mindfulness, meditation, yoga, or similar strategies to cope with depression?” “Have you found ways of actively embracing your feelings, even if sometimes they are painful?” “What has helped you to live a life that is consistent with your values?” Interviews were conducted either by one of the study investigators or an advanced research assistant. We compensated participants $20 for completing the interview.

Professional video recording of participants and initial intervention development

From the initial 33 participants interviewed, we chose 12 individuals reporting successful use of ACT-based coping strategies to take part in a professional video shoot. These participants had to be willing to share their personal stories of coping with depression with others on video. Among the 12 individuals selected for the professional videotaping, the average age was 50.2 (SD=15.1), 8 were female, 8 were non-Latino white, 1 was African American, 1 was American Indian, and 2 were Latino. The average education level was 14.6 years. All participants had a history of being prescribed antidepressant medication and of engaging in psychotherapy. We compensated participants of the professional videotaping $75.

The video recording was overseen by a local video production company. Using predetermined questions based on the initial interview, the video director conducted the on-camera interview with each participant with the researchers present. Over a series of three separate shooting days, a total of 9 hours of video footage was collected from the 12 participants. The video production team, incorporating feedback from the researchers, edited the raw footage into a series of vignettes, featuring several personal stories from each participant touching on a variety of ACT-consistent themes.

Focus groups of videos clips and intervention refinement

We pilot tested a selection of edited video clips in two separate focus groups consisting of 12 total participants (see description in Results section) recruited through community advertisements. Eligibility criteria for focus group participation were: (a) currently seeing a primary care provider; (b) 18 years or older; and (c) some lifetime experience of depression, based on self-report. Participants were compensated $25 for attending the focus group.

Study staff informed focus group participants that they would be shown short videos depicting individuals describing their experience coping with depression to be used in a future video intervention. After each clip was screened, participants completed a 14-item self-report survey consisting of Likert-scale questions assessing acceptability and transportation (i.e., degree of engagement with the videos) adapted from similar measures (Green et al., 2000; Ondersma, Chase, Svikis, & Schuster, 2005). Afterward, research staff led a brief group discussion to obtain further feedback on the videos. Based on focus group feedback and additional discussions between the research team and video production crew, the final video intervention was developed.

Description of intervention

The video intervention, titled LifeStories, consisted of four two-part episodes, each focusing on different key ACT concepts. In addition to the participant stories, two psychologists acted as “hosts” of the series who related the coping strategies described in the stories to key principles of ACT. Each part of an episode included 3–4 storytelling vignettes from patients with lived experience interspersed with commentary from the hosts. Each of the 4 video episodes was 20–25 minutes in length. LifeStories series content is outlined in Table 1.

Table 1.

Outline of LifeStories intervention

Episode 1A: Changing Experiences of Depression
Participant Stories	Workbook activities
Diane describes a time when she contemplated suicide and then realized that her life could change if she took action to move forward with her life. Melissa recounts the challenge of going through a divorce and learning how to cope with change.	Mood monitoring chart Telling a story (writing, drawing, audio recording) of a time when your mood changed unexpectedly
Episode 1B: Finding New Ways of Coping with Depression
Participant Stories	Workbook activities
Mary describes going to work every day despite feeling depressed and using hobbies to cope. Justin talks about playing online video games as a way to connect with others.	Steps for choosing, implementing, and evaluating a new coping strategy for depression
Episode 2A: Defining your Values
Participant Stories	Workbook activities
David defines values and what they mean to him. Tami recounts how her values of taking care of her children and pursuing her own education guide her through difficult life decisions	Answering values clarification questions (“What do you want your life to be about?”)
Episode 2B: Living a Life that is Consistent with your Values
Participant Stories	Workbook activities
David explains how focusing on his values helped him recover from childhood trauma. Michael reflects on his decision to work in child protective services given his own abuse history.	Telling a story about engaging in a difficult behavior because it was values-consistent Values into Action Worksheet: Identifying values, goals, barriers, and an action plan
Episode 3A: Developing Acceptance
Participant Stories	Workbook activities
Ruth describes learning to accept the difficult feelings associated with caring for her mother with dementia. Carolyn remembers learning to “recognize, feel, and deal” with her difficult thoughts and feelings.	Making a list of costs and benefits of avoidance of a difficult thought Writing difficult thoughts on notecards and carrying them during the day
Episode 3B: Self-Compassion
Participant Stories	Workbook activities
Mary talks about learning to “make friends” with her depression. Jeannette emphasizes the importance of forgiving oneself.	Writing down one kind thing to do for oneself each day Writing a self-compassionate letter to oneself
Episode 4A: Living More Fully in the Present Moment
Participant Stories	Workbook activities
David describes the “fleetingness” of all thoughts and emotions. Jeannette discusses the benefits of deep breathing. Practice “Leaves on the Stream” meditation exercise.	Mindfulness meditation script for home practice Additional suggestions for cultivating mindfulness (apps, forms of yoga)
Episode 4B: Advice on Finding a Treatment Provider
Participant Stories	Workbook activities
Jasmine says she looks for “someone who is open-minded” when choosing a therapist. David explains how it’s okay to try out different providers to find the right one.	Explanation of different kinds of mental health providers and treatment approaches Advice on finding a treatment provider that fits your particular needs

Note. Each episode included more participant stories; we describe only a subset here for illustrative purposes.

A 54-page workbook also was developed to accompany the video series. Included in the workbook were summaries of concepts presented in each video episode, pictures and notable quotes from the storytellers featured in the series, and activities intended to encourage users to practice applying coping strategies learned in the video series to their own lives. At the end of each video episode, participants were asked to refer to the section of the workbook that related to what they had just learned and to complete the designated exercises. We recognized that viewing a video is largely a passive experience. Therefore, the rationale for including the supplemental workbook was to provide people with further opportunities to implement the skills and concepts that they were learning from watching the videos to their own lives. In other words, the workbook was used as a way to both reinforce what was being taught in the videos and to encourage participants to transfer these concepts into action to cope with the depression they were experiencing.

Open trial of video intervention

Sample.

We recruited participants for a pilot study of the effectiveness of the LifeStories intervention at a local primary care site as well as through online advertisements (e.g. Craigslist, Facebook). Potential study participants completed an initial phone screen and were invited to the research clinic for an intake interview to determine if they met the following criteria: (a) 18 years or older; (b) currently under the care of a primary care provider (at least one appointment in the past two years); (c) current or lifetime major depressive disorder, assessed via the Structured Clinical Interview adapted for DSM-5 Axis I Disorders (SCID-5) (First, Williams, Karg, & Spitzer, 2015); (d) score of 11 or greater (at least moderate severity) on the Quick Inventory of Depression Symptoms (QIDS) (Rush, Trivedi, Ibrahim, Carmody, Arnow et al., 2003); (e) not currently receiving psychotherapy/counseling; (f) not currently receiving psychopharmacological treatment from a psychiatrist; (g) if taking antidepressant medication prescribed by their primary care provider, no changes in the past 6 weeks; (h) a score of 9 or above (at least moderate severity) on the Patient Health Questionnaire (PHQ-9; Kroenke, Spitzer, & Williams, 2001); (i) no lifetime diagnosis of a psychotic disorder or bipolar disorder as assessed by the SCID-5; (j) below-cutoff scores for hazardous substance use on the Alcohol Use Disorders Identification Test (AUDIT) and Drug Use Disorders Identification Test (DUDIT) (Babor, Higgins-Biddle, Saunders, & Monteiro, 2001; Berman, Bergman, Palmstierna, & Schlyter, 2003); (k) if pregnant, less than 24 weeks; (l) no suicidal ideation or behavior requiring immediate attention; and (m) ability to transport self to research clinic to watch the videos. We required that participants not be receiving current pharmacotherapy provided by a psychiatrist because our ultimate target for the intervention is depressed primary care patients who may be receiving non-specialist depression care. However, if participants were receiving pharmacotherapy for depression prescribed by their primary care provider, we specified that they had to be stable on their type and dose of medication for at least 6 weeks, so that medication effects would be less likely to confound the impact of our intervention.

Study Assessments.

At baseline, we administered the mood module and psychotic screen of the SCID-5 to diagnose MDD, bipolar disorder, and psychotic disorders. As our primary outcome measure, we administered the reliable and valid Quick Inventory of Depression Symptomatology-Clinician Version (QIDS-C) (Rush et al., 2003) to assess depression symptom severity at baseline, week 4, and week 8. Severity ranges on the QIDS-C are as follows: none = 0–5, mild = 6–10, moderate = 11–15, severe = 16–20, and very severe = 21–27. Assessors were trained to initial interrater reliability (> .80) on interviewer-administered measures (i.e., QIDS, SCID) with ongoing training to prevent drift. As a secondary outcome measure of depression, participants completed the PHQ-9 weekly during the baseline and treatment periods. The PHQ-9 is a self-report measure of depression frequently used in primary care which has established reliability and validity.

Additionally, the following measures were administered at baseline, week 4, and week 8. As secondary outcomes, participants completed the self-report PROMIS Anxiety Scale (Pilkonis, Choi, Reise, Stover, Riley et al., 2011) and the 12-item World Health Organization Disability Assessment Schedule-II (WHODAS 2.0) (World Health Organization, 2012) to assess overall functioning. The following potential target mechanisms were assessed because they were designed to tap into the processes affected by ACT for depression based on prior research in related samples. The Acceptance and Action Questionnaire-II (AAQ-II) was chosen because it is a validated measure of experiential avoidance/psychological inflexibility (Bond, Hayes, Baer, Carpenter, Guenole et al., 2011) which is a central target of ACT theorized to result in behavior change. Similarly, the Valued Living Questionnaire (VLQ) was used because it is a validated measure of the consistency between values and daily actions (Wilson, Sandoz, Kitchens, & Roberts, 2010), which also is emphasized in ACT. In addition, we selected the Behavioral Activation in Depression Scale (BADS), which assesses overall level of behavioral activation (Kanter, Mulick, Busch, Berlin, & Martell, 2006), because this is consistent with the proposed behavioral effects of ACT on depression. Finally, the Five Facet Mindfulness Questionnaire (FFMQ; 24-item version) (Baer, Smith, Hopkins, Krietemeyer, & Toney, 2006), which assesses various aspects of mindfulness (observation, description, awareness, non-judgment, and non-reactiveness), was chosen because LifeStories included instruction in mindfulness and encouraged meditation practice.

We administered the Credibility and Expectancy Questionnaire (CEQ) at pre-treatment (week 4), which is a reliable and valid self-report measure of initial expectancies for improvement from treatment (Devilly and Borkovec, 2000). At post-treatment (week 8), participants also completed the Client Satisfaction Questionnaire-8 (CSQ-8), a reliable and valid self-report measure that evaluates satisfaction with treatment (Larsen et al., 1979). Finally, participants completed the same self-report measure of acceptability and transportation previously used in the focus groups (Green et al., 2000; Ondersma et al., 2005). This measure yielded a mean acceptability score and transportation score, each ranging from 1 = “not at all” to 5 = “very much.”

Procedure.

Individuals who met criteria for enrollment participated in the study for 8 weeks total. A quasi-experimental, AB design was used for the open trial (Kazdin, 1998), meaning that participants completed several weeks of assessment prior to receiving the LifeStories intervention. During the first 4 weeks of enrollment, participants completed self-report questionnaires of symptoms and functioning via a secure, online web portal. Beginning at week 4, participants returned to the clinic weekly to watch one episode of the LifeStories program. Participants were asked to watch the videos at the clinic to ensure that they viewed all the episodes and to obtain detailed feedback about their experience to assist in intervention refinement. Prior to viewing each video, participants completed the same self-report questionnaires as they did at home during the first 4 weeks. After viewing each episode, a member of the research staff conducted a brief qualitative interview with the participant to illicit additional feedback. Qualitative interview questions included: “What was the main point of this video for you?” (i.e., message), “What did you like/not like about this video?” (i.e., acceptability), and “What aspects of the video were engaging/not engaging?” (i.e., transportation). Participants viewed one LifeStories episode per week over the course of 4 weeks. One week after the final episode, participants returned to the research clinic for a final appointment (week 8).

Statistical Analyses

Changes during the baseline (weeks 1–4) versus intervention (weeks 4–8) periods were analyzed separately. Paired-samples t-tests were used for assessing within-subject change over time for study variables using pre- and post-period assessments. The only exception to this was in the case of the PHQ-9, which was analyzed using repeated-measures analyses of variance to examine weekly changes during the baseline versus intervention periods. Effect sizes are also reported in terms of Cohen’s d statistics (along with their corresponding confidence intervals): 0.2 = small, 0.5 = medium, and 0.8 representing large effects. Clinical significance criteria (Jacobson & Truax, 1991) were used to assess the percentage of participants showing reliable improvement/worsening on our primary outcome (QIDS-C) at post-treatment. As all participants exhibited at least moderate depression severity at the start of the intervention, we also report the percentages of patients in the mild severity (defined as treatment response) and no symptom (remission) ranges at post-treatment, based on QIDS-C scores of ≤ 10 or ≤ 5, respectively. Given the small sample, we were not able to conduct formal mediation analyses. However, we examined the associations between changes in our target mechanisms and outcomes in an exploratory fashion to inform future work. Pearson correlations were conducted between mechanisms and outcomes using pre- to post-intervention residualized change scores. Residualized change scores control for error related to repeated measurement by standardizing scores and accounting for the correlations between time points (Steketee & Chambless, 1992). Given the pilot nature of the study and the small sample size, all tests were two-tailed and alpha was set at .05. Analyses were intent-to-treat with the last observation carried forward for missing data to be more conservative in interpretation of effects. Finally, we conducted qualitative analyses of interviews conducted with participants following each video viewing. Three members of the research team reviewed all responses to the qualitative questions. Using ATLAS.ti (Friese & Ringmayr, 2011) they collaboratively developed a coding system for responses to questions. Codes were developed for the following categories: acceptability and transportation of the video, message of the video, and workbook. The coding system was iteratively updated until all transcripts and notes were coded.

Results

Focus Groups

The average age of the 12 focus group participants was 46.5 years (SD=17); 1 was non-white; 10 were female; and the average education level was 13.5 years. Focus group members reviewed 12 video clips total. On a scale from 1 (“not at all”) to 5 (“very much”), average item acceptability rating for videos was 4.08 (SD = 0.24) and the average item transportability rating 3.40 (SD = 0.34), suggesting good overall satisfaction and engagement with the video clips.

Open Trial Outcomes

Sample Characteristics.

A total of 41 participants completed the initial phone screen, 12 were assessed in person to determine study eligibility, and 11 were eligible and started the study. Twenty-nine of the original 41 participants screened for the study were deemed ineligible for the following reasons: not interested in participating (n = 5), receiving psychiatric treatment (n = 10), no primary care physician (n = 2), not meeting diagnostic or depression severity criteria (n = 10), or inability to return to office (n = 2). The final sample was comprised of 8 women (73%) and 3 men, 2 (18%) people identifying as Hispanic/Latino and 9 (82%) as White. A total of 4 (36.5%) participants were currently married and 6 (54.5%) were employed full-time. The sample had a mean age of 49.8 (SD = 9.5) and mean education level of 14.8 years (SD = 2.3). In terms of diagnoses, 10 (91.7%) met DSM-5 criteria for current major depressive disorder, with 6 (54.5%) also meeting criteria for current persistent depressive disorder. The average length of the current depressive episode was 17.6 weeks (SD = 22.1) and 10 participants (91.7%) reported more than 1 lifetime depressive episode. At study entry, 5 (45.5%) participants were in the moderate range and 6 (55.5%) were in the severe range for depressive symptoms, according to the QIDS-C. In terms of medication, 5 (45%) participants were taking an antidepressant prescribed by their primary care provider at baseline. Antidepressants prescribed included sertraline (n=2), venlafaxine (n=1), fluoxetine (n=1), and citalopram (n=1). No patients were in psychotherapy or counseling at study start.

Participant retention and adherence.

Eleven participants completed the baseline period (i.e., all baseline assessments), and 10 completed all assessments during the intervention period. Only 1 person dropped out of the study at week 6 due to purported difficulty returning to the clinic to view the remaining videos. Ten of the 11 participants viewed all 4 video episodes. During the intervention phase, 2 patients reported discontinuing their antidepressant medication under the supervision of their primary care physician, and 2 patients began individual psychotherapy in the community to continue treatment post-study.

Credibility, Acceptability, and Satisfaction Ratings.

The CEQ was used to assess beliefs related to the intervention at the start of the treatment phase. The mean CEQ Total score was 32.5 (SD = 9.9), the mean CEQ Credibility score was 18.3 (SD = 4.8), and the mean CEQ Expectancy score was 14.2 (SD = 5.8). CEQ scores indicated overall acceptable levels of credibility/expectancy that was similar to scores reported in previous studies of novel psychosocial interventions for mood disorders (Wenze, Gaudiano, Weinstock, Tezanos, & Miller, 2015). In addition, participants rated the acceptability of the video episodes highly, with a mean of 4.57 (SD = 0.13) on a 1–5 scale. On the same 5-point scale, participants gave the video series a mean rating of 3.74 (SD = 0.21) for “transportation” (i.e., how engaged they were while watching the videos). Furthermore, the mean post-treatment satisfaction according to the CSQ-8 in the sample was 28.1 (out of 32 max; SD = 17.7), which was comparable to ratings found in previous studies for novel psychosocial interventions for mood disorders (Wenze et al., 2015).

Primary Outcomes.

See Table 1 for descriptive statistic for study measures and effect size changes. According to paired-samples t-tests for the interviewer-rated QIDS-C, depressive symptom severity did not change significantly between weeks 1 and 4 of baseline assessment (t = 1.64, df = 10, p = .13, Cohen’s d = 0.47). After participants received the LifeStories intervention from weeks 4 to 8, QIDS scores were reduced significantly at week 8 when compared to week 4 (t = 3.93, p = 0.003, Cohen’s d = 1.28). See Figure 1.

Figure 1.

Changes during Baseline and Intervention Periods on Interviewer-Rated Depression Severity

Secondary Outcomes.

According to the self-report PHQ-9, change in depressive symptom severity was only approaching significance over weeks 1 and 4 of baseline assessment, F (df = 4, 40) = 2.12, p = 0.096, Cohen’s d = 0.60. After participants received the LifeStories intervention, PHQ-9 scores were reduced significantly from weeks 4 to 8 (F = 3.56, df = 4, 40, p = 0.014, Cohen’s d = 0.91). We did not find significant changes pre- to post-assessment for either the baseline or intervention periods on the PROMIS-Anxiety or WHODAS (ps = ns).

Potential Mechanisms of Action.

For the FFMQ, the Non-Reactivity Subscale showed a significant improvement from intervention weeks 4 to 8 (t = 2.28, df = 10, p = 0.046, Cohen’s d = 0.49), whereas there was no significant change observed during the baseline period (t = 0.89, p = ns, Cohen’s d = 0.29). Change on the FFMQ Non-Judgment Subscale was approaching significance during the intervention period only (t = 1.90, df = 10, p = .087). No other significant differences were found pre- to post-assessment for either the baseline or intervention periods on the BADS, AAQ, or VLQ (ps = ns).

Clinically Significant Changes.

We also examined reliable change and the percentages of patients showing response/remission on the QIDS-C (primary outcome). We used the reliability (Cronbach’s alpha = .86) of the QIDS-C as reported by Rush et al. (2003) in our calculations. Reliable change analyses revealed that 4 participants (36.4%) showed no change, 0 showed clinical deterioration, and 7 (63.6%) showed improvement. In addition, 6 participants (54.5%) achieved clinically significant change, defined as showing reliable change plus a change from the moderate/severe range at intervention baseline to the minimal/none range on the QIDS-C at post-treatment. In addition, 2 participants (18.2%) were in the remitted range (≤ 5) according to the QIDS-C at post-treatment.

Change Score Correlations between Mechanisms and Outcomes.

Pearson correlations were computed based on residualized change scores for intervention weeks 4 to 8 between our primary outcome (QIDS-C) and potential mechanisms of action (BADS, FFMQ, AAQ, VLQ). The QIDS-C and BADS change scores were significantly correlated with each other with a large effect (r = .62, p < .05), indicating that decreased depression was associated with increased behavioral activation. No other significant correlations were detected (ps = ns).

Open Trial Qualitative Feedback

As described above, a coding system was developed by the investigators by extracting common themes based on the detailed review of each participants’ recorded feedback obtained after they viewed the videos. Themes were coded based on their theoretical relevance to the intervention and content acceptability/unacceptability. The major categories coded from the interviews included: messages derived from each video, transportation/engagement in each video and the relatability of the storytellers, acceptability/unacceptability of the content presented, and workbook-related use and comments. Coding of interviews was done with multiple investigators present and any disagreements were resolved through discussion among the team to ensure reliability. Participants’ feedback statements were individually categorized according to the above themes in an iterative fashion and refined into subcategorizes until all interviews were rated. Major findings from this qualitative analysis are summarized below.

Message.

In response to questions about the message of video 1, many participants articulated the idea that one needs to change one’s behavior to improve one’s life. For example, one participant said: “try to find something to help yourself. Set a little goal for yourself every day to do.” Participants also articulated the message of impermanence of thoughts and feelings (e.g., “depression comes and goes”). After watching video 2, participants identified that the main message was about identifying values and turning values into action. For example, in response to a question about the main point of the video, one participant said “That if you pay attention to your values and live true to them, mindfully, then you feel much better about yourself over time.” For video 3, participants identified the main message as the importance of identifying avoidance and increasing acceptance of emotions. One participant said: “The message is you’re okay, you just have to stop letting all the emotions and the negatives make you feel like you’re not okay because you have them.” Many participants also identified self-compassion as an important message, for example, saying “you need to forgive yourself.” Participants identified mindfulness as a major message in video 4. For example, one person said the video was “about mindfulness, being in the present moment, awareness of thoughts.” Finally, participants mentioned the discussion of treatment options, and some expressed specific interest in ACT.

Acceptability and transportation.

In response to all of the videos, participants said that the storytellers seemed authentic. One participant said “The people are very real. Feel like you’re in the room with them…I can imagine getting to know them.” Participants also gave feedback about specific stories to which they related well, with occasional feedback about storytellers or aspects of stories that they did not relate to (e.g., drug and alcohol use, history of abuse as a child, depression due to a temporary stressor). The only message about which participants had some negative comments was mindfulness (in the 4^th video). Although generally participants liked the focus on mindfulness, three mentioned that it can be difficult to practice. Two of those three liked the focus on mindfulness despite the fact it could be difficult: “mindfulness I liked a lot…although it is hard sometimes.”

Workbook.

Across all four weeks of the intervention, participants spoke about liking the workbook and the exercises, planning to use it, and believing that the workbook could be helpful. Specific exercises that participants tried and found helpful included a daily mood log, defining and prioritizing values, and doing one kind thing for oneself per day. However, some participants did not use the workbook at all. Barriers to using it included finding time and remembering to use it, finding some of the activities overwhelming, not wanting to be vulnerable, and not liking the concept of homework.

Discussion

In this article, we described our iterative and participatory research approach for developing a narrative, video-based self-help intervention for depression using behavior change principles from ACT. After interviewing patients about their experiences coping with depression, and identifying a subgroup who were willing to be professionally-recorded discussing their experiences on camera, we conducted focus groups to obtain initial feedback on our storytellers’ video clips. Next, we developed a professionally produced self-help video series, called LifeStories, featuring people describing their personal success using ACT-based strategies for coping with depression. We then conducted an open trial (AB design) to assess the acceptability, feasibility, and potential effects of LifeStories. Results from the open trial suggested that LifeStories was acceptable and the use of real patient vignettes was considered engaging by participants. We also found significant changes during the intervention period (but not the baseline period) on interviewer-rated and self-report depressive symptoms, as well in our target mechanism related to mindfulness (non-reactivity to inner experiences). The experience and data obtained from this project will inform the further refinement of the self-help videos, and we plan to test the intervention in a pilot randomized controlled trial as the next step in our development process. We plan to test LifeStories against a comparison condition in future trials to better understand the unique effects of the intervention and its components. Possible control conditions that we believe would be useful to consider include an attentional video featuring health education information or a possible comparison condition with a more traditional mobile/internet-delivered CBT intervention.

Participants viewed LifeStories as acceptable and credible. They also provided mostly positive feedback about their experience. In fact, we received consistent feedback that participants found the storytellers authentic and relatable. Participants also reported a generally high degree of engagement while watching the video, which is considered important in narrative interventions because the degree of transportation has been found to predict changes in knowledge, attitudes, and behavioral intentions related to health behaviors (Murphy, Frank, Chatterjee, & Baezconde-Garbanati, 2013). We also would note that there was a very low attrition rate in the study. All participants completed the baseline period and only one person dropped out during the intervention period. One interesting observation was that two participants reported that they decided to discontinue their current antidepressant medication in consultation with their primary care provider due to lack of effectiveness, and two other participants decided to pursue formal psychotherapy in the community by the end of the study. The two participants who discontinued their antidepressant treatment additionally expressed an interest in trying psychotherapy at the end of the study. This suggests that LifeStories may be helpful for creating increased interest and direct requests by consumers for evidence-based psychotherapies after getting introduced to these topics through the video series (Santucci et al., 2012).

It is important to note that almost all participants met criteria for current major depression at the start of the study, and all exhibited moderate to severe levels of depressive symptoms. Despite the relatively severe and chronic depression in the sample, participants showed large effect size improvements in depressive symptoms over the 4 week intervention phase of the study. We also examined the clinical significance of depression outcomes, with 64% of our sample showing reliable symptom improvement, 54% showing reliable response, and 18% showing reliable remission according to QIDS-C criteria at the end of treatment. Our findings are consistent with previous research showing that relatively brief self-help interventions are efficacious for treating depression. For example, Cuijpers et al. (Cuijpers, Donker, Johansson, Mohr, Straten et al., 2011) reported a meta-analysis of self-guided psychological interventions without therapist assistance and found that they had a small but significant effect on depressive symptoms at post-treatment that was maintained through 12 month follow-up. Our results are also consistent with other self-help interventions featuring ACT. For example, Fledderus et al. (2012) conducted a randomized controlled trial of an ACT self-help program with email support vs ACT self-help without email support vs a waitlist control. Both ACT self-help conditions showed moderate to large effect size improvements in depression, anxiety, mindfulness, experiential avoidance, and positive mental health compared with the waitlist control, with no differences between active interventions. Effect size changes on depression outcomes reported for ACT in the Fledderus et al. study were both large in magnitude and in a similar range as those found in the current trial of LifeStories. For example, pre-post change in the primary outcome of depressive symptoms was d = 1.28 (QIDS-C) in the current study compared with d = 1.41 (Center for Epidemiological Studies Depression Scale) in the Fledderus et al. study in their self-help ACT with email support condition (as calculated from Table 2, p. 491). A potential advantage for LifeStories is that this intervention is shorter than other typical self-help programs (e.g., 1 month compared to 2 months) and does not require formal therapist support.

Table 2.

Open Trial: Descriptive Statistics of Study Measures

	M (SD)			Cohen’s d effect size change with 95% confidence intervals		t (df=10)		P
	Baseline	Pre-Int	Post-Int	Baseline to Pre-Int	Pre-Int to Post-Int	Baseline to Pre-Int	Pre-Int to Post-Int	Baseline to Pre-Int	Pre-Int to Post-Int
QIDS-C	15.55 (1.92)	14.45 (2.70)	9.45 (4.80)	−0.47 ( −0.65, 2.12)	−1.28 (−0.25, 4.18)	1.64	3.93	0.13	0.003
PHQ-9	15.36 (4.06)	13.27 (2.87)	10.0 (4.22)	−0.60 (−1.77, 2.32)	−0.91 (−0.74, 3.44)	1.24	2.03	0.24	0.07
PROMIS – Anxiety	15.91 (5.87)	16.91 (4.76)	13.36 (4.25)	0.20 (−3.67, 2.62)	−0.83 (−1.99, 3.34)	−0.60	1.78	0.56	0.12
WHODAS	15.45 (4.08)	13.27 (4.0)	12.09 (4.57)	−0.57 (−1.85, 2.93)	−0.29 (−2.08, 2.99)	1.47	0.87	0.17	0.40
AAQ	27.0 (9.53)	29.73 (7.59)	28.55 (6.68)	0.33 ( −5.96, 4.15)	−0.17 (−4.31, 4.12)	1.75	−0.91	0.11	0.39
BADS	71.36 (12.90)	73.82 (10.48)	78.36 (11.85)	0.22 (−7.84, 5.97)	0.43 (−6.62, 6.58)	−0.56	−1.42	0.59	0.19
FFMQ
Non-reactivity	13.09 (2.22)	12.36 (2.77)	13.91 (3.45)	−0.31 (−1.01, 1.94)	0.52 (−2.16, 1.52)	0.87	−2.28	0.40	0.05
Observing	12.82 (4.51)	13.36 (4.32)	13.55 (4.55)	0.13 (−2.79, 2.42)	0.04 (−2.60, 2.64)	−0.76	−0.40	0.47	0.70
Acting with awareness	14.0 (3.97)	13.27 (2.94)	13.18 (3.49)	−0.22 (−2.13, 1.96)	−0.03 (−1.71, 2.09)	1.23	0.14	0.25	0.89
Describing	17.0 (4.90)	17.18 (4.81)	17.73 (4.29)	0.04 (−2.93, 2.80)	0.13 (−2.97, 2.41)	−0.27	−1.75	0.79	0.11
Non-judging	13.91 (3.05)	13.82 (3.34)	14.55 (3.21)	−0.01 (−8.16, 1.98)	0.23 (−2.21, 1.66)	0.15	−1.90	0.88	0.09
VLQ	39.55 (14.84)	41.46 (11.21)	38.94 (11.25)	0.15 (−8.92, 6.47)	−0.24 (−6.39, 6.88)	−0.87	0.84	0.40	0.42

Note. Pre-Int = Pre-Intervention; Post-Int = Post-Intervention; QIDS-C = Quick Inventory of Depression Symptoms; PHQ-9 = Patient Health Questionnaire; AAQ = Acceptance and Action Questionnaire-II; BADS = Behavioral Activation for Depression Scale; FFMQ = Five Facet Mindfulness Questionnaire; PROMIS = Patient-Reported Outcomes Measurement Information System; VLQ = Valued Living Questionnaire; WHODAS = World Health Organization Disability Assessment Schedule

Because LifeStories was an ACT-based intervention, we hypothesized potential target mechanisms that included mindfulness, acceptance, values, and behavioral activation. Effect size changes observed post-treatment in these target mechanisms were consistent with our theoretical rationale for the intervention; although, given the small sample, it was not surprising that most of these changes were not statistically significant. We did find significant improvement during the intervention period on the FFMQ-Non-Reactivity subscale, which measures people’s tendency to permit their thoughts and feelings to come and go, without getting unnecessarily attached or caught up in them. The previously mentioned Fledderus et al. (2012) self-help study of ACT also reported large effect sizes improvements on the FFMQ subscales of non-reactivity and non-judgment (which was approaching significant change post-treatment in the current study). Previous research shows that changes in non-reactivity to inner experiences as measured by the FFMQ are related to time spend doing certain mindfulness exercises, including the body scan and sitting meditation (Carmody & Baer, 2008). We also found that changes in behavioral activation pre- to post-intervention were associated with changes in depressive symptoms over the same period. This is consistent with prior treatment research documenting that increased behavioral activation according to the BADS is related to depression improvement (Manos, Kanter, & Busch, 2010). We plan to test the potential mechanisms further in future studies of LifeStories relative to a comparison condition using more formal mediation analyses in a larger sample.

There are several things that we learned from the open trial that will factor into our refinement of LifeStories for the next phase of testing. First, although participants noted that they liked the accompanying workbook and thought that it provided useful suggestions, few participants actually reported using the workbook frequently. We plan to refine the workbook so that the exercises are simpler and easier to complete to encourage increased adherence. Second, we plan to add supplemental video clips featuring our storytellers based on feedback from some participants who said they would learn more about their lives. We wanted to keep the overall length of each video episode manageable (less than 30 minutes total, to mimic a typical television episode), but we think it could be helpful to include this additional “bonus” content for those who interested in viewing more content. Third, we received feedback that our “leaves on the stream” meditation exercise in particular may have been too advanced, at least for some of our more meditation naïve participants. In this exercise, participants are encouraged to practice placing thoughts and feelings on leaves as they arise in the moment, and watching them gently float down an imagined stream. We plan to change this exercise to a more “beginner” meditation exercise that focuses simply on attention to the breath.

There are several limitations that should be considered when interpreting results. Although having a baseline lead-in period prior to the intervention helps to document that changes occurring afterward are more likely to be associated with the actual provision of the treatment, only a true parallel, randomized controlled design can rule out extraneous factors and determine cause-effect. Our next step is to conduct a pilot randomized trial to do just that. In addition, given the small sample size, we only had statistical power to detect larger changes on measures, and our confidence intervals around effect sizes were wide. Scholars recommend caution in attempting to estimate the true effects of an intervention based on results obtained in small samples due to error and variability (Kraemer, Mintz, Noda, Tinklenberg, & Yesavage, 2006). Given the exploratory nature of the study, we did not correct for multiple tests which increases the risk of Type I error, so it will be important to replicate the current findings in future larger-scale investigations of LifeStories. Furthermore, we did not have a follow-up period post-intervention, so we were not able to determine how long effects were maintained following the study. Finally, we chose to have participants watch LifeStories at our office because we wanted to ensure that the videos were being viewed completely and so that we could elicit detailed post-viewing feedback from participants to aid in our refinement and future testing of the intervention. However, it is possible that having participants view LifeStories at home would have resulted in differences in adherence or treatment effects. Thus, in future trials of LifeStories, we plan to have participants complete the intervention remotely to better examine the effects of the intervention during home use.

Results from the current study demonstrate that patient narratives may be an effective and engaging way to deliver evidence-based coping strategies for treating depression. In the future, we can envision adapting the content of LifeStories into a more widely-disseminable web/mobile-based format if further testing confirms that the intervention is safe and effective for treating depression. We envision that programs like LifeStories can be used for a number of different purposes in various settings, including primary care: 1) as a standalone self-help intervention for those with lower levels of depression; 2) as a brief intervention to possibly motivate people to engage in more intensive psychotherapeutic interventions when indicated; 3) as an adjunctive intervention for those taking antidepressant depression with suboptimal response; and 4) as an initial immediately-available intervention delivered to those who are currently waiting to start traditional treatment. LifeStories and other self-help interventions, especially considering their lower costs, hold promise for increasing the dissemination and impact of evidence-based behavior change strategies to larger numbers of people to improve public health.

Biographical Statements

Brandon A. Gaudiano, Ph.D. is Associate Professor (Research) at the Alpert Medical School of Brown University and Research Psychologist at Butler Hospital in Providence, Rhode Island. His research interests focus on the devepment and testing of novel psychological interventions for mood and psychotic disorders, particularly acceptance/mindfulness-based approaches.

Carter H. Davis, B. F. A. is a Research Assistant in the Psychosocial Research Program at Butler Hospital. His primary interests include mindfulness, acceptance and commitment therapy, and innovative treatment delivery methods (including web, mobile, and video).

Ivan W. Miller, Ph.D. is Professor at the Alpert Medical School of Brown University and Director of the Psychosocial Research Program at Butler Hospital. His current research interests pertain to the study of innovative methods for assessing and preventing suicide.

Lisa A. Uebelacker, Ph.D. is Associate Professor (Research) at the Alpert Medical School of Brown University and Research Psychologist at Butler Hospital. Her research focuses on integrating behavioral health into primary care, and on improving health in people with depression via mind-body programs such as hatha yoga or physical activity.