Circadian rhythm instabilities precipitate depressive and manic symptoms of bipolar disorder (
1), underscoring the need for interventions that target stabilization of circadian rhythms for individuals with the disorder. The widely accepted social zeitgeber theory of mood disorders (
2) purports that disruptions to circadian rhythms are caused by disruptions in individuals’ social routines, which, in turn, lead to mood episodes. Accordingly, a robust pathway to stabilize circadian and mood patterns is by regularizing social rhythms from awakening, through mealtimes and physical and social activities, to bedtimes.
Interpersonal and social rhythm therapy (IPSRT) is an evidence-based psychotherapy designed to stabilize social rhythms, provide psychoeducation about bipolar disorder, and resolve an interpersonal problem related to mood symptoms (
3). IPSRT has been shown in randomized controlled trials of adults with bipolar disorder to be effective in reducing mood symptoms (
4,
5). It has also shown promise as an adjunctive psychosocial intervention among adolescents and young adults (
6–
8). Social rhythm therapy (SRT) (
9), the component of IPSRT that focuses on teaching behavioral strategies to maintain healthy routines, especially sleep, has been tested separately from the full intervention among individuals with major depressive disorder (
10) and bipolar disorder (
11). Among persons with bipolar disorder (ages 15–35 years), the SRT component of IPSRT has been shown to be most helpful in staying well over 5 years (
12). The inclusion of adolescents in these trials suggests benefits of SRT for younger persons.
Circadian rhythm instabilities are also potent risk factors for suicidal thoughts and behaviors (
13) and likely contribute to the high suicide risk among individuals with bipolar disorder. Adolescence and young adulthood constitute the primary epoch when fully syndromal bipolar disorder develops and when suicidal thoughts and behaviors often emerge. Adolescence and young adulthood may therefore be a crucial developmental window when interventions that target circadian rhythm disruptions could not only relieve acute mood symptoms but could also have lifelong benefits, including minimizing illness progression and reducing suicide risk.
In the present study, we adapted SRT to include remote delivery, with nine of 12 sessions delivered via a secure video platform to adolescents and young adults with bipolar disorder. SRT delivered by telehealth has not been previously tested within the adolescent and young adult population. The benefits of remote delivery of interventions, or telehealth, transcend cost and time savings (
14,
15). Indeed, these may be only second-order gains because a key impact of telehealth is improved access to treatment. Telehealth exponentially expands the reach of an intervention, beyond both geographical and socioeconomic boundaries (
16). It allows evidence-based interventions to reach individuals who are in need and who have limited access to mental health services in their communities. In-person visits demand greater initiative and motivation from individuals, whereas telehealth poses a lower barrier to inertia. In a vicious cycle, the highest barriers to in-person visits exist for those who are severely ill, including those at risk for suicide, and telehealth can ease access to treatment for such individuals (
17,
18). The past decade has witnessed a steady rise in telehealth efforts, and the emergence of COVID-19 has necessitated an almost overnight shift in mental health provision (
19,
20). In a period when social distancing and lockdowns are common, telemedicine can ensure continuity of treatment, which is important to the long-term successful management of bipolar disorder and suicidal thoughts and behaviors. We expected in this trial that SRT delivered by telehealth would contribute to daily routine regularity, because we anticipated that remote delivery would facilitate sessions being administered on the same day of the week and at the same time.
The goals of this study were twofold. The primary aim was to determine the feasibility and acceptability of a 12-week SRT delivered predominantly via telehealth to a preliminary sample of adolescents and young adults with bipolar disorder. The secondary aim was to explore whether the intervention would be beneficial in improving social rhythm regularity and reducing mood symptoms and suicide propensity among these individuals. We hypothesized that SRT delivered predominantly by telehealth would be feasible and acceptable for adolescents and young adults with bipolar disorder, as demonstrated by a high retention rate, excellent client satisfaction, and a strong therapeutic working alliance. We also expected that individuals receiving the intervention would show improved regularity of social rhythms and reductions in depressive and manic symptoms and propensity for suicidal behavior. Because evidence suggests that treatments shown to decrease suicide risk among individuals with bipolar disorder may do so via mechanisms independent of reduction in mood symptoms (
21), we explored whether reductions in suicide propensity were dependent on mood symptom reductions.
Methods
Participants
The participants for this ongoing study (Clinicaltrials.gov identifier: NCT03183388) were enrolled from November 2017 to February 2020 and were recruited from a large, urban, university-affiliated, academic medical center and the surrounding community. Thirteen participants (mean±SD age=20.5±1.9 years, range 17–24 years) met criteria for bipolar I or II disorder according to the
DSM-5 (
22) and had scores of ≥15 on the 29-item Hamilton Depression Rating Scale (HDRS-29) (
23) and/or ≥12 on the Young Mania Rating Scale (YMRS) (
24). Exclusion criteria were major medical disorders (except treated hypothyroidism); central nervous system conditions, including loss of consciousness ≥5 minutes; substance or alcohol use disorders within the prior 6 months; IQ <70; currently receiving structured interpersonal and social rhythm, cognitive-behavioral, dialectical behavior, or family therapies; current psychosis; YMRS score > 25 or being too symptomatic to participate according to the clinical judgment of a licensed clinician; and having active suicidal intent or homicidal ideation.
Participants were enrolled in a program of study, entitled Brain Emotion circuitry-targeted Self-Monitoring And Regulation Therapy (BE-SMART), during which they also participated in magnetic resonance imaging sessions at the three in-person visits. Within the BE-SMART study, the SRT was delivered as BE-SMART-DR, whereby DR stands for “daily rhythm” regularization. Participants were randomly assigned in a 1:1 allocation to either BE-SMART-DR or to a therapy focused on explicit emotion regulation skills without an SRT component (BE-SMART-ER). BE-SMART-ER was not designed to be a comparator for BE-SMART-DR, but rather to investigate variations in therapy approaches—bottom-up (by regularizing daily rhythms with BE-SMART-DR) and top-down (by increasing executive control of emotion regulation with BE-SMART-ER)—which were hypothesized to converge in improving the functioning of emotion regulation brain circuitry. Therapists were randomly assigned participants across the BE-SMART-DR and BE-SMART-ER therapy conditions. Therapists were trained by three clinicians (W.K.S., H.A.S., H.P.B.) and attended weekly supervision meetings. During training and supervision, emphasis was placed on ensuring that the therapists understood the important distinctions of each therapy and that there was no “bleeding” between the two therapies in order to preserve integrity and fidelity. Results from the BE-SMART-ER arm and neuroimaging analyses will be reported separately after completion of the registered clinical trial.
Written informed consent from participants ages ≥18 years, or assent from minors and permission from their guardians, was obtained in accordance with the Yale School of Medicine Human Investigation Committee Institutional Review Board.
BE-SMART-DR
BE-SMART-DR is a 12-week manualized adaptation of SRT consisting of one-to-one weekly sessions and between-session assignments. The 12-week length was based on prior SRT studies (
11) and was chosen to include time to attend to nonspecific factors inherent to any bipolar-specific psychotherapy (
25), such as establishing therapeutic alliance and providing psychoeducation about the disorder. The adaptation also included remote session delivery. First, middle, and last sessions were conducted in person, whereas remaining sessions were offered on a secure video platform. The need for in-person sessions was driven primarily by the demands of the BE-SMART protocol, which required completion of on-site neuroimaging. The intervention adaptation additionally incorporated developmental considerations, such as explicitly evaluating and addressing social media use, developing handouts with engaging graphics, and including examples relevant to this age group (e.g., in the context of school). All sessions were video or audio recorded and used for training and supervision. Although fidelity was not formally measured, an investigator (H.A.S.) reviewed all sessions for fidelity to the manual. Each participant had a licensed clinician in the community who was notified of their client’s participation in the study; the study intervention was delivered adjunctive to the community treatment.
The intervention focused on SRT’s main goals: to provide psychoeducation about bipolar disorder and circadian rhythms (sessions 1–2); to develop participants’ understanding of the relationship between disruptions to their SRs and mood disturbances (sessions 3–5); to regularize sleep and other daily rhythms that affect circadian rhythms (sessions 6–8); to develop self-awareness of symptoms, zeitstörers (rhythm disruptors), and early warning signs of relapse (session 9–10); and to discuss termination and develop relapse prevention plans (sessions 11–12). Each week, participants completed the five-item version of the Social Rhythm Metric (
26), which was used throughout as a therapeutic tool.
Feasibility and Acceptability
Retention rate was calculated as the percentage of participants who participated in all 12 sessions among those who engaged in therapy (defined as attending more than one session). Client satisfaction and therapeutic alliance of the participants and therapists were rated at the end of the intervention by using the Client Satisfaction Questionnaire (CSQ) (
27) and the Working Alliance Inventory (WAI) (
28). The CSQ consists of eight items rated by the participant on a 4-point scale, ranging from 1, low satisfaction, to 4, high satisfaction. Scores range from 8 to 32, with higher scores (
26–
32) indicating “excellent” satisfaction. The WAI consists of 36 items each rated, in separate administration to either the participant or the therapist, on a 7-point scale ranging from 1, never, to 7, always. The WAI measures three domains: agreement on the tasks of therapy (task subscale), agreement on the goals of therapy (goal subscale), and development of an affective bond (bond subscale). Scores for each of the three domains range from 12 to 84. Global scores (the sum of the domain scores) range from 36 to 252, with higher scores indicating a more positive rating of working alliance. The participants and therapists were blind to each other’s ratings on the WAI.
Clinical Outcomes
Prior to and again after the intervention, participant self-report measures included assessments of regularity of social rhythms and propensity for suicidal behavior. Regularity of social rhythms was assessed by using the Brief Social Rhythm Scale (BSRS) (
29), a 10-item scale where participants respond to each item on a 6-point scale ranging from 1, very regularly, to 6, very irregularly. Scores range from 6 to 60, with lower scores indicating more regularity. Propensity for suicidal behavior was assessed by using the suicide propensity subscale of the Concise Health Risk Tracking (CHRT) Scale (
30), which has been used previously to assess suicide propensity among adolescents (
31) and individuals with bipolar disorder (
30). The propensity subscale of the CHRT consists of nine items, and participants respond to each item on a 5-point scale ranging from 0, strongly disagree, to 4, strongly agree. Scores range from 0 to 36, with higher scores indicating greater suicide propensity. Clinicians blinded to intervention assignment administered the HDRS-29 (
23) and YMRS (
24). The 29-item HDRS includes the 21-item HDRS and eight additional items to assess atypical depressive symptoms. HDRS-29 items are scored on a 3-, 4-, or 5-point Likert scale. Total scores range from 0 to 90, with greater scores indicating greater severity of depression. The YMRS scale consists of 11 items and is scored on a 5-point Likert scale. Scores range from 0 to 60, with higher scores indicating greater severity of mania. The four scales (BSRS, CHRT, HDRS-29, YMRS) were also administered at the middle visit (week 7) to explore potential early changes.
Data Analyses
Analyses were conducted by using SPSS, version 26.0. The WAI was missing <1% of values, which were imputed by using the mean imputation method. One participant was excluded from the CSQ analyses for omitting 50% of responses. The data were checked for adherence to normal distribution by using normal probability plots and Kolmogorov-Smirnov tests. Potential differences from pre- to postintervention and from pre- to midintervention in normally (BSRS, HDRS-29) and nonnormally (YMRS, CHRT) distributed continuous variables were assessed by using paired t and Wilcoxon signed-rank tests, respectively. Effect sizes were calculated using Cohen’s d (
32) and r for normally and nonnormally distributed data, respectively. Analyses for associations between changes in BSRS and CHRT suicide propensity scores were performed by using partial regression, with HDRS-29 and YMRS change scores as covariates. Potential differences between therapists and participants in their working alliance (as assessed by the WAI) were analyzed by using the paired samples t test.
Results
Feasibility and Acceptability
Retention was 77%, with 10 of 13 participants having completed all study procedures. Of the three individuals who discontinued participation in the study, one (17-year-old female) did so after 11 sessions because she did not wish to return for the magnetic resonance imaging planned during the 12th visit and therefore missed the termination session focused on review of the participant’s gains and experience with the intervention and recommendations for follow-up care. The second participant (20-years-old male) discontinued after nine sessions because of clinical worsening requiring a higher level of care, and the third participant (19-year-old male) discontinued after two sessions, informing study staff that he did not think he had bipolar disorder so the therapy was not right for him. Treatment satisfaction among the 10 who completed treatment was excellent (CSQ, mean=29.4±2.7). Both therapists and participants had high WAI scores on the task subscale (client, mean=74.8±6.3; therapist, mean=70.2±10.3), bond subscale (client, mean=76.6±4.4; therapist, mean=72±8.2), and goal subscale (client, mean=79.9±3.3; therapist, mean=72.4±10.8). Participants’ assessments of the quality of the therapeutic alliance (WAI global score) were high (mean=231.3±8.1), did not significantly differ from those obtained from the therapists (mean=214.6±28.6, t=1.91, df=9, p=0.088), and were similar to those achieved by published telehealth psychotherapy trials (
33).
Table 1 summarizes demographic and clinical information of the participants who completed all 12 sessions.
Clinical Outcomes
Participants demonstrated significant pre- to postintervention improvements across outcome measures (BSRS, t=2.97, df=9, p=0.016, d=0.94; HDRS-29, t=2.70, df=9, p=0.024, d=0.86; YMRS, Z=–2.20, df=9, p=0.028, r=0.49; CHRT suicide propensity, Z=2.20, df=9, p=0.028, r=0.49). Increases in BSRS scores were associated with reduced CHRT propensity scores when analyses were controlled for HRSD-29 and YMRS score reductions (r=0.90, p=0.002). Exploration of changes from pre- to midintervention showed significant improvements in BSRS (t=4.39, df=9, p=0.002, d=0.1.39) and CHRT suicide propensity (Z=2.67, df=9, p=0.008, r=0.60) scores and reductions at the trend level in HDRS-29 (t=2.11, df=9, p=0.064, d=0.67) scores.
Discussion
This study, to our knowledge, was the first to investigate the feasibility, acceptability, and effects on social rhythm regularity, mood symptoms, and suicide propensity of SRT delivered largely via telehealth to adolescents and young adults diagnosed as having bipolar disorder. High treatment retention, excellent client satisfaction, and strong working alliance scores support the feasibility and acceptability of this intervention for adolescents and young adults with bipolar disorder. Examination of changes from pre- to postintervention showed significant improvements in social rhythm regularity and reductions in depression, manic and hypomanic symptoms, and suicide propensity. Furthermore, associations between improvements in social rhythm regularity and reductions in suicide propensity were observed even when the analyses were controlled for mood symptom changes. At the midpoint (week 7), significant improvements in social rhythm regularity and suicide propensity and trend-level reductions in depression were observed, suggesting the potential for early benefits. Effect sizes observed were in the moderate-to-high range. The findings were consistent with a recent systematic review and network meta-analysis (
34) showing that individuals with bipolar disorder benefit from adjunctive skills-based psychosocial interventions combined with pharmacotherapy and that social rhythm regularization is one of the few strategies effective in reducing both depressive and manic symptoms.
Suicide risk among individuals with bipolar disorder is a critical clinical concern. More than half of youths with bipolar disorder have thought about suicide and will at some point make an attempt (
35). Studies examining the effectiveness of psychotherapies in reducing suicidal thoughts and behaviors among adolescents and young adults are limited. Dialectical behavior therapy (DBT) is the only evidence-based psychotherapeutic intervention that has demonstrated reductions in suicidal ideation among adolescents with bipolar disorder (
36). However, DBT is typically administered over longer periods of time than the 12 weeks of the BE-SMART study. For preadolescents (ages 7–13 years) with bipolar disorder, child- and family-focused cognitive-behavioral therapy, a psychosocial intervention developed as an adjunct to pharmacological treatment (
37), has been shown to reduce the intensity of suicidal ideation when administered over 12 weeks (
38). In the current study, reduction in suicide propensity among adolescents and young adults with bipolar disorder was observed as early as after seven sessions, suggesting that SRT may have a role in reducing more proximal suicide risk in still more scalable versions.
This study had some limitations. Absence of a comparator condition made it impossible to separate the effects of SRT from community care or regression to the mean. SRT was delivered adjunctive to treatment in the community, which included pharmacotherapy for all but one participant. The small sample size means that the findings, including the moderate-to-high effect sizes, should be viewed with caution. An adequately powered randomized controlled trial is needed to formally evaluate efficacy of the remotely delivered SRT intervention in reducing suicide propensity among this age group; however, the positive findings have important implications for early interventions to reduce suicide risk associated with bipolar disorder.
The administration of SRT by video teleconferencing enhances accessibility across geographical boundaries. Feasibility of implementing SRT by telehealth suggests that this intervention is well suited to the future of psychotherapy delivery, which will undoubtedly include remote treatment delivery (
39). To further extend the reach of SRT, we have recently developed and are testing a fully automated online version of the intervention, Rhythms and You, for adults with bipolar disorder (
39). The opportune timing in generating these remote interventions has been highlighted by the COVID-19 pandemic, which has necessitated a rapid transformation in health care delivery globally.
Conclusions
This study provides initial evidence that SRT delivered largely by telehealth is feasible and acceptable for adolescents and young adults with bipolar disorder and can reduce mood symptoms and suicide risk.