Outcomes of a Character Strengths–Based Intervention on Self-Esteem and Self-Efficacy of Psychiatrically Hospitalized Youths

Institutional review board approval of the study was obtained from the University of Michigan Medical School. Participants were recruited on admission to the child and adolescent inpatient psychiatric unit at the University of Michigan Hospital between May and October 2012. Informed assent and consent were obtained from all participants and their legal guardians, respectively. Exclusion criteria included intellectual disability, absence of a legal guardian, or psychosis. Four recruits were excluded because they were prematurely discharged from the hospital. The remaining 81 adolescents, of whom 49 were female and 32 male and whose ages ranged from 12 to 17 (mean±SD age=15.3±1.4), were randomly assigned to a comparison (N=40) or intervention (N=41) group. The average length of hospitalization was 8.4±4.2 days (range 4–29). [A table in the online supplement to this report shows participants’ demographic and descriptive statistics.]

We used a longitudinal randomized controlled design with repeated measures at four time points. Participants completed the Rosenberg Self-Esteem Scale (SES) and General Self-Efficacy Scale (GSE) on their day of admission (baseline), two days after the intervention or comparison activity (day 4 of hospitalization or postintervention), and by mail or e-mail at two weeks and three months. Completion rate at the two-week follow-up was 75% (N=61), and seven persons were inpatients at the time of completion (four youths were not yet discharged, and three were readmitted). Completion rate at three months was 60% (N=49), with five individuals readmitted at that time.

During the second inpatient day, participants in the intervention group completed the VIA-Youth online (https://www.viame.org). The instrument consists of 198 statements that probe for 24 character strengths; items are rated on a 5-point Likert scale (13). Scores for seven to nine statements per strength subscale are averaged, and the application returns, by subscale score, a list of each participant’s top, or “signature,” strengths. Multiple studies, including one with 250 middle school students, found the measure to be reliable and valid, with each subscale having good internal consistency (13). Participants finished the survey in approximately 20 minutes without investigator interaction other than to monitor Web use and clarify questions. Participants were then instructed to write their top three character strengths on index cards. The researcher and participant discussed, in approximately 24 minutes, ways to use these strengths as coping skills. Youths recorded an average of two to four coping mechanisms per strength [examples are provided in the online supplement]. Participants were instructed to keep the index cards with them and use their identified coping skills throughout the hospitalization and after discharge. Comparison group participants engaged in a one-on-one activity with the researcher (for example, a board game) for a period approximating the interactive component (about 24 minutes) of the VIA-Youth intervention.

The GSE is a ten-item self-report measure of perceived self-efficacy, rated on a 4-point Likert scale (14). Internal consistency for the scale was high at baseline (Cronbach’s α=.86), postintervention, and follow-ups.

The SES also has ten items rated on a 4-point Likert scale (15). Internal consistency was high at baseline (α=.91) and at all other time points.

To measure the degree to which treatment group participants were incorporating their character strengths into coping skills, we asked one additional question at postintervention, two weeks, and three months to determine whether (and how) youths had been using their strengths as coping mechanisms.

Acceptability was assessed with a three-question survey immediately after the intervention. Participants were asked whether they agreed with the VIA-Youth survey results, were previously aware that they possessed these strengths, and enjoyed the intervention and believed it was a positive experience.

Two-tailed paired t tests were used for self-esteem and self-efficacy outcomes after the intervention, and two-tailed independent t tests were used for all other comparisons. Because not all participants were retained at each time point, differences at each follow-up were analyzed with respect to the individual’s baseline. A mixed-model approach was not used, given that it was not possible to assume that data were missing at random. All analyses were performed with IBM SPSS Statistics 20 software.

There were no significant group differences in demographic characteristics (race, ethnicity, and gender) or self-esteem and self-efficacy at baseline. There was a significant increase in self-efficacy and self-esteem from baseline to postintervention for both groups, but only the experimental group showed sustained benefits in self-efficacy at two weeks and both self-efficacy and self-esteem at three months (Table 1; additional results are shown in a figure in the online supplement). Among participants in the experimental group, those who reported incorporating their strengths into coping skills (N=25), compared with those who did not (N=15), showed greater improvements in self-esteem (t=2.28, df=38, p=.028) and self-efficacy (t=2.91, df=38, p=.006) at postintervention (second figure in online supplement). This difference persisted at two weeks for self-esteem (24 youths incorporating versus eight youths not incorporating strengths; t=2.07, df=30, p<.05) and showed a similar, although not a statistically significant, trend for self-efficacy. Both differences were lost by three months. In addition, at the two-week follow-up, participants unaware of their strengths before the intervention (N=16), compared with those endorsing previous awareness of their strengths (N=18), showed greater improvement in self-esteem (t=2.45, df=32, p<.05). No such difference was found for self-efficacy or at any additional time point.

A majority (88%, N=36) of participants who received the intervention reported that it was a positive experience, 10% (N=4) had neutral feelings, and one participant reported that the experience was not positive. In addition, 88% (N=36) of the intervention participants reported that they agreed with all of the VIA-Youth results (in other words, they thought their signature strengths were accurate), whereas 12% (N=5) agreed with only some.

This study demonstrated that a simple and brief character strengths–based intervention used with a broad spectrum of adolescents admitted to a psychiatric hospital for children resulted in measurable gains in self-esteem and self-efficacy up to three months later. Such improvements may in turn increase confidence and enhance motivation and coping behaviors, leading to increased engagement in psychiatric treatment (11). The gains were largely independent of participants’ prior knowledge of their strengths and self-reported compliance with the therapeutic assignment, although short-term gains were greater for individuals who reported being previously unaware of their character strengths and for those who reported having incorporated their strengths into coping skills. Participants overwhelmingly viewed the experience as positive—not insignificant for a psychotherapeutic exercise with youths (online supplement). The intervention is easily incorporated into inpatient programming and takes about 45 minutes, of which about 20 minutes is completing a publicly available, self-administered, Web-based tool. To our knowledge, this is the first application of a positive psychology intervention in a natural population of youths who were psychiatric inpatients.

Several observations are noteworthy. Self-esteem and self-efficacy in the comparison group also significantly increased from baseline to postintervention on day 4 of hospitalization. However, this improvement was short-lived, likely reflecting therapeutic benefits of inpatient treatment as usual; no significant gains were apparent in the comparison group at two-week and three-month follow-ups. The weak correlation of outcomes with prior knowledge of character strengths suggests that the latter may not be a reliable predictor of benefit; even those who claimed no new learning of their strengths did better over time. This may indicate that knowledge of one’s strengths is not equivalent to understanding that those strengths are valued and can be utilized as coping skills. However, the significant difference at the two-week follow-up indicates that further research is needed to gain a more nuanced understanding of how learning about character strengths may influence intervention success.

The hypothesis that youths who report continuing to use their strengths as coping skills would show a more positive change in self-esteem and self-efficacy was partially supported, with significant differences at postintervention for both outcomes and at the two-week follow-up for self-esteem. The difference, however, faded by three months. In contrast, adults who continued using their strengths reported greater happiness and less depressive symptoms at every follow-up assessment (8). The difference in the tenacity of the intervention may be attributable to different outcome measures or to developmental differences between youths and adults. Future studies should use a variety of measures to facilitate comparison.

This preliminary study tested whether an easily administered intervention based on positive psychology principles delivered in a real-world setting could produce a measurable, sustained benefit. A number of factors limit our ability to attribute the benefit specifically to positive psychology. To avoid contamination effects, we did not expose the comparison group to the VIA-Youth survey; hence we cannot know if there were differences in character strengths at baseline. Only the treatment group was prompted to generate coping strategies; thus some benefit may be attributable to the exercise, expectancy, and related visual cues (index cards) rather than specifically to the positive psychology content. The nature of the interaction with the researcher in the comparison group—engaging in a leisure activity versus a discussion of personal relevance—could have contributed to differences. This study was also unable to implement an entirely blinded procedure. Inpatient psychiatric treatment encourages sharing of experiences. Although we do not know how much was shared or its impact, it seems unlikely that any contamination would magnify the difference between groups or explain the benefit seen at three months. Future studies should attempt to minimize or quantify this confound. Finally, potential biases include there being more nonresponders in the comparison group and that just one investigator carried out all study elements. To minimize possible bias from the latter condition, a strict protocol for intervention, administration, and survey distribution was followed.

The generalizability of the results is limited in that the study was conducted with primarily Caucasian patients in a relatively affluent area. The study should be replicated in various inpatient psychiatric treatment facilities and include a larger, more ethnically and socioeconomically diverse sample.

Our findings illustrate the potential benefits of incorporating a character strengths–based approach in the mental health treatment of youths. As Resnick and Rosenheck (10) observed, a “language of strengths” can be fostered and utilized throughout the recovery process, and once youths are able to utilize their top strengths confidently and effectively, they can leverage them to address lesser strengths and weaknesses (4).

The authors thank staff on the Child and Adolescent Psychiatric Unit at the University of Michigan for their support, and they acknowledge the late Christopher Peterson, Ph.D., for his contribution to the positive psychology field.