Interpersonal psychotherapy for mood and behavior dysregulation: Pilot randomized trial

Abstract

Background:

Youth with chronic irritability and excessive reactivity, diagnosed as disruptive mood dysregulation disorder (DMDD), have social impairment in multiple settings (i.e., peers, school, and home). This paper presents a pilot randomized trial assessing the feasibility, acceptability, and preliminary efficacy of interpersonal psychotherapy (IPT) for mood and behavior dysregulation (IPT-MBD), an adapted version of IPT for depressed adolescents. IPT-MBD focuses on decreasing outbursts and irritability and improving interpersonal interactions.

Methods:

Nineteen adolescents (aged 12–17) with DMDD or its research precursor, severe mood dysregulation, were randomly assigned to IPT-MBD (n = 10) or treatment-as-usual (TAU, n = 9) in a 24-week psychosocial intervention study. Assessments of mood symptoms and overall functioning were conducted by an independent evaluator, blinded to treatment, every 4 weeks. Parent and self-report irritability measures were collected every 4 weeks.

Results:

Eighty percent of participants randomized to the IPT-MBD arm completed the study. Also, participants enrolled in the IPT-MBD arm attended >80% of therapy sessions. Parents and teens agreed that the frequency and duration of therapy were appropriate and were satisfied with IPT-MBD treatment. Clinical global impression scales for severity and improvement showed statistically greater improvement in the IPT-MBD group compared to TAU.

Conclusions:

In this small pilot randomized trial, IPT-MBD was feasible and acceptable to parents and teens. There was significantly more improvement in the IPT-MBD group compared to TAU. IPT-MBD holds promise as a potentially effective psychosocial intervention for clinically impaired youth with DMDD and warrants further investigation in a larger randomized trial.

1 |. INTRODUCTION

Debate in child and adolescent mental health has centered on the conceptualization of pediatric bipolar disorder (PBD). Central to the debate is whether PBD is characterized by a chronic course (Biederman et al., 2005) or by distinct episodes of mania (Leibenluft, 2011). This issue is clinically important because the conceptualization of PBD drives treatment recommendations. For this reason, researchers at NIMH created a clinical construct labeled severe mood dysregulation (SMD; Leibenluft, Charney, Towbin, Bhangoo, & Pine, 2003) to study chronic irritability and compare them to youth with bipolar disorder (i.e., youth with episodic irritability). Based on numerous studies comparing youth with chronic versus episodic irritability, researchers determined that chronic irritability in youth is not a developmental phenotype of PBD but a distinct construct (Brotman et al., 2006, 2007; Leibenluft, Cohen, Gorrindo, Brook, & Pine, 2006; Rich et al., 2008). The culmination of this research is the Diagnostic and Statistical Manual of Mental Disorders 5th Edition’s (DSM-5) inclusion of a new mood disorder, disruptive mood dysregulation disorder (DMDD; American Psychiatric Association, 2013).

Prevalence rates of DMDD range from 1 to 3% (Copeland, Angold, Costello, & Egger, 2013). Childhood DMDD is associated with increased rates of anxiety and depression in young adulthood (Copeland, Shanahan, Egger, Angold, & Costello, 2014). DMDD in childhood is also associated with high levels of social impairment, including in parent, teacher, sibling (Copeland et al., 2013), and peer relationships (Mulraney et al., 2016). Long-term outcomes of childhood irritability include low educational attainment, financial impoverishment, and increased likelihood of multiple psychiatric disorders (Copeland et al., 2014).

Since outbursts are the most apparent symptom of DMDD, behavioral management strategies are often utilized in treatment. Regarding pharmacologic management, in a research sample of youth with SMD, youth were prescribed an average of two medications (Deveney et al., 2015), including stimulants, selective serotonin reuptake inhibitors, mood stabilizers, and antipsychotic medications. There are limited data supporting the efficacy of interventions for youth with DMDD. Lithium is ineffective (Dickstein et al., 2009). Though stimulant optimization resulted in improvement in externalizing symptoms in youth with attention-deficit hyperactivity disorder and DMDD, youth remained globally impaired (Baweja et al., 2016). In a randomized study, utilizing an 11-week group psychosocial intervention for preadolescents with SMD, there was greater improvement in irritability in the intervention group compared to community care (Waxmonsky et al., 2016). In a randomized trial of preadolescents with DMDD, dialectical behavior therapy demonstrated feasibility and preliminary efficacy compared to treatment-as-usual (TAU) (Perepletchikova et al., 2017).

Due to the social impairment associated with DMDD, improving interpersonal relationships is a crucial treatment target. Furthermore, outbursts typically occur in an interpersonal context, thereby adversely affecting relationships with family, peers, and teachers. Interpersonal psychotherapy (IPT) is an empirically supported psychotherapy, which focuses on improving mood symptoms while improving relationships, through building adaptive interpersonal skills. Thus, IPT for Depressed Adolescents (IPT-A) (Mufson, Dorta, Wickramaratne et al., 2004; Mufson, Dorta, Moreau, & Weissman, 2004; Mufson, Weissman, Moreau, & Garfinkel, 1999) was adapted for youth with DMDD/SMD to create interpersonal psychotherapy for mood and behavior dysregulation (IPT-MBD; Miller, Hlastala, & Mufson, 2015), a 24-week psychosocial intervention. IPT is based on the premise that there is an association between the onset and/or maintenance of mood symptoms and a person’s interpersonal relationships (Mufson, Dorta, Moreau, et al., 2004). The goals of IPT-MBD are to improve interpersonal skills and decrease irritability and outbursts.

This pilot randomized trial assesses the feasibility and acceptability of IPT-MBD in DMDD and explores the efficacy of IPT-MBD compared to TAU. We hypothesized that adolescents in the IPT-MBD group will demonstrate an overall retention rate ≥80% (based on similar pilot feasibility studies) and that teen and parent satisfaction scores would average 6 (higher is more satisfied) on a 7-point Likert scale. We also hypothesized that the IPT-MBD group would have more improvement in irritability and outbursts compared to the TAU group.

2 |. MATERIALS AND METHODS

2.1 |. Participants

Adolescents (age 12–17) were recruited from outpatient psychiatry clinics and through advertisements. Of the 31 adolescents screened, 12 were excluded for subthreshold mood symptoms for DMDD/SMD (n = 5), comorbid diagnoses (n = 3), and acute danger to self (n = 2). One participant was lost to follow-up, and one participant was not randomized due to feasibility (extended trip during study). Nineteen participants meeting criteria for DMDD/SMD were enrolled and randomized (Figure 1) over a 14-month period. The symptoms of DMDD/SMD are chronic irritability, severe and recurrent outbursts, and impairment in two of three domains (peers, school and home), with severe impairment in one. The main difference between SMD and DMDD is the former includes hyperarousal symptoms. Recruitment for this study began before the creation of DMDD; therefore, youth with SMD or DMDD were included.

FIGURE 1.

CONSORT flow diagram

This study was approved by the University Institutional Review Board. Consent was obtained from the guardian and assent from the adolescent.

Diagnoses were determined by master-level clinicians utilizing the schedule for Affective Disorders and Schizophrenia for School-Age Children (KSADS-PL). Other inclusion criteria included are as follows: estimated IQ > 70, Children’s Global Assessment Scale (CGAS) ≤ 60, and Clinical Global Impression Scale-Severity DMDD/SMD (CGI-S) ≥ 4. Exclusion criteria included are as follows: pregnancy, significant risk for dangerousness to self or others, or disorders including current episode of major depressive disorder, autism, schizophrenia, bipolar disorder, eating disorder, and substance dependence/use.

2.2 |. Procedure

Enrolled participants were randomly assigned to IPT-MBD (n = 10) or TAU (n = 9) using a computerized block randomization schedule. The principal investigator (PI) remained blinded to the randomization schedule. Only the research coordinator was unblinded to the randomization scheme and assigned the treatment arm once the participant was enrolled. Participants in the IPT-MBD group received weekly psychotherapy for 24 weeks. Participants in the TAU group continued with their current provider if already established in therapy or were referred to a nonstudy therapist. TAU therapists, either within or outside of the same hospital clinic system, determined type and structure (family, individual, etc.) of therapy, as well as, frequency of visits. TAU therapists were either psychologists, social workers, or licensed counselors.

Both groups were assessed every 4 weeks over the 24-week protocol by an independent evaluator (IE), blinded to treatment assignment. One of the two IE’s completed the CGI, CGAS, Mood Symptoms Questionnaire (MSQ), and the Children’s Depression Rating Scale-Revised (CDRS-R). Both IEs observed several administrations of these measures as part of an ongoing NIMH study for youth with DMDD/SMD. Though interrater reliability was not formally tested, the IEs had ongoing communication with research staff at NIMH to facilitate consensus with NIMH ratings. Parents and teens completed irritability and anxiety measures (see below). The KSADS-PL was repeated at week 24 with the reporting time frame being the last 4 weeks of the study. Parents and children in the IPT-MBD group completed a treatment satisfaction questionnaire. Parents in the TAU group were asked about frequency of treatment visits. Medication changes were permitted and recorded. At baseline, female participants had a urine pregnancy test, and all participants had a urine toxicology screening test.

2.2.1 |. Interpersonal psychotherapy for mood and behavior dysregulation

IPT-MBD (Miller et al., 2015) was adapted from IPT-A to treat the chronic irritability and verbal and physical outbursts of DMDD/SMD. IPT-MBD was developed in collaboration with the developers of Interpersonal and Social Rhythm Therapy for Adolescents with Bipolar Disorder (IPSRT-A; Hlastala, Kotler, McClellan, & McCauley, 2010) and IPT-A (Mufson et al., 1999; Mufson, Dorta, Moreau, et al., 2004; Mufson, Dorta, Wickramaratne, et al., 2004). Modifications include longer duration of treatment, changes to the psychoeducation component, inclusion of anger/outburst ratings and outburst data gathering, and discussion of family accommodations (decreased expectations to avoid outbursts). Other modifications include creation of outburst safety plan (individualized plan to help teen de-escalate to decrease physically aggressive outbursts), modified social rhythm metric (to stabilize circadian rhythms; adapted from IPSRT-A), and increase in parental involvement (for a detailed description of IPT-MBD and evidence-based case study, see Miller, Hlastala, Mufson, Leibenluft, & Riddle, 2016).

There are three phases of IPT-MBD (initial, middle, and termination phases) and four problems areas of focus: interpersonal role transitions, role disputes, deficits, and grief. During the initial phase, the therapist meets individually with the adolescent to gather information on outbursts and important relationships, to identify patterns of outbursts in the context of interpersonal interactions, and to develop a formulation focused on a specific problem area most linked to the mood symptoms. In the middle and termination phases, all sessions begin with the therapist meeting individually with the adolescent. The middle phase focuses on increasing awareness of one’s emotions and the emotions of others through encouragement of affect and perspective taking, as well as, improving communication and problem-solving skills. The termination phase focuses on solidifying and generalizing skills for relapse prevention. At minimum, the therapist meets with the parent during the initial phase for psychoeducation and rationale for IPT-MBD, in the middle phase for work as needed on the parent–child relationship, and during termination to address clinical change and assess need for further treatment. Additional sessions with the parent may be scheduled as needed by the therapist.

Prior to this study, a proof of concept 20-week study was conducted where the PI delivered IPT-MBD to three adolescent participants with SMD (Miller, 2015). Most improvement in symptoms occurred from week 16 to 20; therefore, treatment was extended to 24 weekly sessions to enhance solidification of skills. Sessions ranged from 45 to 60 min.

One of the three clinicians (the PI, a child and adolescent psychiatrist, and a licensed clinical professional counselor) delivered IPT-MBD. Didactic training was provided prior to the study. The PI treated six participants, while the other clinicians each treated one participant. The PI completed two standard IPT-A and seven IPT-MBD cases under direct supervision prior to providing supervision to the other clinicians. For this study, all sessions of the PIs first two cases were reviewed by SH. For the subsequent two cases, half the sessions were reviewed by SH; for the remainder, sessions were randomly chosen by SH for review. For the duration of the study, weekly supervision calls between the PI and SH were conducted. The PI reviewed all audiotaped sessions and provided weekly supervision for the other treating clinicians. The PI also reviewed these cases with SH.

2.3 |. Measures

2.3.1 |. Feasibility and acceptability

Feasibility was assessed via number of participants who completed the study and number of therapy sessions attended. Acceptability was measured with a satisfaction questionnaire completed by teens and parents. This measure consisted of 10 Likert-scored items (7-point scale, higher scores indicate more satisfied) and included questions about therapeutic alliance, satisfaction with length/frequency of sessions, and satisfaction with IPT-MBD. This assessment was modified from the satisfaction questionnaires from the Treatment for Adolescents with Depression Study (TADS, 2003) and IPSRT-A study (Hlastala et al., 2010).

Diagnostic evaluation

The KSADS-PL (Kaufman et al., 1997) is a semistructured diagnostic interview used to diagnose pediatric psychiatric disorders. The KSADS-PL was administered by trained master’s level clinicians. Training included observation of KSADS-PL interviews completed by experts and observation administering the KSADS-PL. Diagnoses were determined by best-estimate consensus. Parents completed a Social Communication Questionnaire Lifetime form (Rutter, Bailey, & Lord, 2012) at baseline to identify those participants with a high likelihood of an autism spectrum disorder (ASD), in addition to the completion of the KSADS-PL ASD screen.

2.3.2 |. Mood and anxiety ratings

Irritability and outbursts

MSQ is a semistructured clinician administered interview to assess severity and frequency of DMDD/SMD symptoms. The MSQ was used to obtain DMDD/SMD-specific CGI ratings. From the MSQ, a CGI score is generated for outbursts and mood between outbursts, which is used to determine an overall CGI-DMDD/SMD score. The CGI (Guy, 1976) is a clinician-administered measure consisting of two subscales, severity and improvement. The severity subscale (CGI-S) is based on severity of current symptoms (1–7, lower score less severe), and the improvement subscale (CGI-I) is based on the current week’s symptoms relative to baseline (1–8, lower score more improvement).

Irritability

The Affective Reactivity Index (ARI; Stringaris et al., 2012), a parent and self-report dimensional measure of irritability has demonstrated excellent internal consistency and differentiation between clinical and nonclinical populations (Stringaris et al., 2012).

Depression

CDRS-R (Poznanski & Mokros, 1996) is a clinician-administered instrument that assesses for depressive symptoms and has been utilized in adolescent depression treatment studies (Brent et al., 2009; TADS, 2003). T scores ≤ 54 are unlikely to indicate a depressive disorder.

Anxiety

The Screen for Child Anxiety Related Disorders (SCARED; Birmaher et al., 1999) is a parent and self-report measure of anxiety, and a score of ≥25 indicates the possibility of an anxiety disorder. The SCARED has demonstrated good validity in pediatric outpatient psychiatric populations (Monga et al., 2000).

2.3.3 |. Functional measures

CGAS (Shaffer et al., 1983), a reliable and valid scale used by clinicians to assess a child’s overall functioning, has been used in treatment studies of youth with SMD (Dickstein et al., 2009). A lower score indicates more impairment.

Cognitive screen

IQ was estimated via the Kaufman Brief Intelligence Test-2 (Kaufman & Kaufman, 2004), which includes verbal, nonverbal, and composite IQ scores.

2.3.4 |. Data analytic strategy

Attendance is summarized as the average number of sessions attended, as well as proportion of completers. As this study was not powered to demonstrate a difference between conditions, preliminary efficacy is reported. Preliminary efficacy was assessed using primary (CGI, CGAS) and secondary (ARI, CDRS-R, SCARED) outcome variables.

Baseline measures of functioning and symptoms were compared by treatment arm to assess potential imbalance in distributions between treatment arms. In the absence of any clinically meaningful differences in these measures by treatment arm, efficacy of the IPT intervention at 24 weeks post-baseline was assessed using Wilcoxon rank-sum non-parametric test to compare the distributions of the primary and secondary outcome variables between arms. In addition, the within-group change in distributions of these measures at 24 weeks from baseline using Wilcoxon matched-pairs signed-ranks test was assessed.

Linear mixed effects models with random intercept for participant were used for each outcome to model the linear trajectory (over 24 weeks) by intervention arm, taking into account within-person correlation of the study outcomes over time. There were four dropouts from the study at week 24 (two in each arm). Based on our data for the reasons for leaving the study, we assumed missing at random and used linear mixed models to appropriately account for this missing data mechanism by borrowing information within and across study participants (Schafer & Graham, 2002). All six measurements (every 4 weeks) were used to look at the longitudinal change. The model included follow-up time (in weeks), treatment arm, and time by treatment arm interaction as the primary predictors. Statistically significant interaction indicates that the patient-specific trajectories for these outcome measures differ by treatment condition. The above-described model was modified to include (1) random slope for time and (2) autoregressive residuals. These additional models were compared using Bayesian Information Criterion (BIC). The most parsimonious model (smallest BIC) is presented for each outcome.

Continuous measures are summarized using mean ± SD, while frequencies and percentages are presented for categorical variables. Pilot data were also used to evaluate the effect size at week 24 by calculating Cohen’s d and 95% confidence intervals (CI).

3 |. RESULTS

3.1 |. Sample description and covariate balance

Of the 19 participants enrolled, the mean age was 13.7 years (SD = 1.19), and 58% were male. The sample was 57.9% Caucasian, 31.6% African–American, and 10.5% Hispanic/Latino. Most subjects met criteria for comorbid diagnoses (Table 1). Sixty-three percent of randomized participants were prescribed medication at baseline, most commonly a stimulant (83%) and/or selective serotonin reuptake inhibitor (42%). Of the seven participants not prescribed medication at baseline, five were medication naïve. See Table 1 for demographics and medical history characteristics and Table 2 for baseline functional measure summaries by group. There were more females in the IPT group (70 vs. 11%), and no appreciable differences in the distributions of functional measures at baseline by treatment arm (Table 2). Due to the small sample size, we did not adjust for gender differences between the treatment arms, nor did we explore potential effect modification by gender.

TABLE 1.

Demographic and medical history characteristics by treatment arm

Participant characteristics	IPT-MBD (n = 10)	TAU (n = 9)
	Mean (SD)	Mean (SD)
Age	14 (1.2)	13.7 (1.2)
	N (%)	N (%)
Gender
Female	7 (70)	1 (11)
Male	3 (30)	8 (89)
Race/ethnicity
Caucasian	6 (60)	5 (56)
African-American	2 (20)	4 (44)
Hispanic/Latino	2 (20)	0
Household status
Single-parent household	6 (60)	4 (44)
Annual income ≤60,000	5 (50)	5 (56)
Education history
Current or past IEP or 504 plan	6 (60)	7 (78)
Suspension or expulsion	5 (50)	7 (78)
Primary diagnosis
SMD and DMDD	4 (40)	7 (78)
SMD alone	6 (60)	1 (11)
DMDD alone	0	1 (11)
Comorbid current diagnoses
ODD	10 (100)	9 (100)
ADHD	10 (100)	8 (89)
Anxiety	6 (60)	4 (44)
Past depressive episode	2 (20)	1 (11)
Medication at baseline	6 (60)	6 (67)
Stimulant	6 (60)	4 (44)
SSRI	1 (10)	4 (44)
Nonstimulant ADHD medication	1 (10)	1 (11)
Atypical antipsychotic	1 (10)	0
Mood stabilizer	0	1 (11)
No medication at baseline	4 (40)	3 (33)
No past medication trial	3 (30)	2 (22)

IPT-MBD, interpersonal psychotherapy for mood and behavior dysregulation; TAU, treatment-as-usual; IEP, Individualized Education Program; DMDD, disruptive mood dysregulation disorder; SMD, severe mood dysregulation; ODD, oppositional defiant disorder; ADHD, attention deficit hyperactivity disorder; SSRI, selective serotonin reuptake inhibitor.

TABLE 2.

Baseline functional measures by treatment arm (N = 19)

	IPT-MBD (n = 10)	TAU (n = 9)
	Mean (SD)	Mean (SD)
KBIT	96.4 (15.5)	101.0 (12.2)
SCQ	5.9 (5.1)	9.7 (3.7)
CGI-S	4.6 (0.5)	4.9 (0.3)
CDRS-R	55.1 (4.8)	54.4 (3.3)
CGAS	43.8 (1.8)	43.2 (2.1)
SCARED-Parenta	23.6 (9.8)	21.2 (11.4)
SCARED-Self	19.4 (12.7)	15.0 (12.2)
ARI-Parenta	8.7 (2.9)	9.4 (2.2)
ARI-Self	7.2 (2.8)	4.7 (3.5)

IPT-MBD, interpersonal psychotherapy for mood and behavior dysregulation; TAU, treatment-as-usual; KBIT, Kaufman Brief Intelligence Test; SCQ, Social Communication Questionnaire; CGI-S, Clinical Global Impressions Scale-Severity; CDRS-R, Children’s Depression Rating Scale-Revised; CGAS, Children’s Global Assessment Scale; SCARED, The Screen for Child Anxiety Related Disorders; ARI, Affective Reactivity Index.

^aIf two parents completed forms, the average was calculated.

3.2 |. Retention and satisfaction

Fifteen participants completed the study: 8/10 (80%) in the IPT and 7/9 (78%) in the TAU arm (Figure 1). In the IPT-MBD group, one participant dropped out due to a move out of state, and another dropped out due to maternal stressors. In the TAU group, one participant dropped out due to dissatisfaction with TAU assignment, and the other did not complete the last two assessments.

Of the eight (80%) participants in the IPT-MBD arm who completed treatment, the average number of therapy sessions attended was 20 ± 3 out of 24 sessions (83%). The average satisfaction scores for the IPT-MBD intervention were high, 6.2 ± 0.2 for parents and 6.6 ± 0.2 for adolescents (1–7, 7 indicates higher score). Mean parent and teen scores on therapeutic alliance were 6.5 ± 0.2 and 6.2 ± 0.0, respectively (1–7). Teens and parents felt the length and frequency of sessions were appropriate.

In the TAU group, though all participants were given referrals for providers, one family chose not to pursue psychotherapy and received only medication management. One family delayed pursuing therapy and did not want medication management. Six participants received both psychotherapy and medication management. Of those who regularly engaged in therapy, sessions were typically weekly or biweekly. The number of therapy sessions varied from 0 to 16 during the 24-week study.

3.3 |. Mood and anxiety measures and diagnoses

Table 3 shows comparison of response rates at week 24 by treatment condition. The IPT-MBD group had significantly lower scores on the CGI-S (2.8 ± 1) compared to TAU (4.1 ± 0.7; P = 0.01). On average, the IPT group were considered “mildly ill” at week 24 compared to the TAU group who were on average “moderately ill,” according to the CGI-S. IPT-MBD also demonstrated more improvement on the CGI-I (2.6 ± 1.1) compared to TAU (3.9 ± 0.9; P = 0.04). On average, the CGI-I score for IPT-MBD was “improved,” whereas for the TAU group, it was “minimally improved.”

TABLE 3.

Posttreatment week 24 outcomes by treatment arm (N = 15)

	Mean (SD)
	IPT-MBD (n = 8)	TAU (n = 7)	P-valuea
CGI-S	2.8 (1.0)	4.1 (0.7)	0.01
CGI-I	2.6 (1.1)	3.9 (0.9)	0.04
CGAS	59.8 (13.1)	48.4 (4.4)	0.04
CDRS-R	44.2 (6.9)	47.4 (5.7)	0.45
SCARED-Parentb	13.4 (7.8)	8.7 (7.7)	0.22
SCARED-Self	9.0 (7.7)	2.7 (3.6)	0.14
ARI-Parentb	5.2 (4.2)	8.1 (2.8)	0.15
ARI-Self	3.1 (3.2)	3.6 (2.5)	0.52

CGI-S, Clinical Global Impressions Scale-Severity; CGI-I, Clinical Global Impressions Scale-Improvement; CGAS, Children’s Global Assessment Scale; CDRS-R, Children’s Depression Rating Scale-Revised; SCARED, Screen For Child Anxiety Related Disorders; ARI, Affective Reactivity Index.

^aWilcoxon rank-sum test.

^bIf two parents completed forms, the average score was calculated.

Data were also analyzed to examine linear trajectories of outcomes over time (Table 4, Figure 2). There was a significant difference in rate of change in CGI-S scores from baseline to week 24 for IPT-MBD compared to TAU (P = 0.008). The model estimates that a participant in the IPT-MBD group had CGI-S improvement of 1.7 from baseline to week 24 (95% CI: 1.2–2.2, P < 0.001) compared to 0.7 (95% CI: 0.2–1.3, P = 0.007) for a TAU participant.

TABLE 4.

Longitudinal analyses of outcome measures by intervention arm: Results of linear mixed effects models showing estimated change to week 24 from baseline (ß), its 95% confidence interval, and P-value for test of no change in outcome measures

	IPT-MBD			TAU			P-value for interactionb
	(n = 10)			(n = 9)
	βa	95% CI	P-value	β a	95% CI	P-value
CGI-Sc	−1.7	−2.2, −1.2	<0.001	−0.7	−1.3, −0.2	0.007	0.008
CGI-Id	−2.1	−2.9, −1.3.	<0.001	−1.1	.−2.0, −0.2	0.015	0.090
CGASc	13.1	8.9, 17.2	<0.001	4.4	1.2, 7.6	0.006	0.001
CDRS-Rc	−9.9	−13.7, −6.1	<0.001	−9.6	−15.7, −3.6	0.002	0.939
ARI-Pe,f	−3.5	−5.6, −1.5	0.001	−1.5	−3.6, 0.7	0.179	0.175
ARI-Cc	−3.3	−5.7, −1.0	0.005	−0.7	−2.3, 0.8	0.338	0.068
SCARED-Pe,c	−10.3	−13.8, −6.8	<0.001	−9.3	−19.1, 0.4	0.061	0.855
SCARED-Cc	−5.4	−9.2, −1.6	0.005	−7.9	−13.1, −2.7	0.003	0.447

^aEstimated change to week 24 from baseline calculated as 24 × (estimated slope for time) for each treatment arm; positive value indicates an increase.

^bTests for difference in slopes between treatment arms, that isWald test for arm × time interaction term in the model.

^cFrom linearmixed effects model with random intercept for participant, and random slope for time and autoregressive residuals.

^dFrom linearmixed effects model with random intercept for participant.

^eIf two parents completed forms, the average score was calculated.

^fFrom linearmixed effects model with random intercept for participant and random slope for time.

CGI-S, Clinical Global Impressions Scale-Severity; CGI-I, Clinical Global Impressions Scale-Improvement; CGAS, Children’s Global Assessment Scale; CDRSR, Children’s Depression Rating Scale-Revised; SCARED, Screen For Child Anxiety Related Disorders; ARI, Affective Reactivity Index.

FIGURE 2.

Outcome measure trajectories by intervention arm

At week 24, there was no difference between groups on CDRS (P = 0.45), SCARED parent (P = 0.22), SCARED child (P = 0.14), ARI parent (P = 0.15), or ARI child (P = 0.52) (Table 3). The longitudinal analysis also did not reveal any differences in rate of change between treatment arms on these measures (Table 4).

The KSADS-PL was administered at week 24 with the reporting time frame being the last 4 weeks of the study. Of the eight participants in the IPT-MBD group, only one met criteria for having DMDD/SMD, while four of the seven TAU participants continued to meet criteria.

3.4 |. Functional measures and effect size

At week 24, the CGAS (59.8 ± 13.1) was also significantly higher for IPT-MBD than TAU (48.4 ± 4.4; P = 0.04). The linear trajectory demonstrated a significant difference in rate of change in CGAS scores from baseline to week 24 for IPT-MBD compared to TAU (P = 0.001). The estimated improvement from baseline to week 24 in CGAS was 13.1 (95% CI: 8.9–17.2, P < 0.001) for IPT-MBD and 4.4 (95% CI: 1.2–7.6, P = 0.006) for TAU.

The estimated effect size using Cohen’s d and based on mean comparison of CGI-S is 1.56 (95% CI: 0.36, 2.7) and is 1.25 (95% CI: 0.10, 2.4) for CGI-I.

3.5 |. Serious adverse events and medication changes

One participant in the TAU arm had a psychiatric hospitalization for agitation and aggressive behavior during the study. There were no hospitalizations in the IPT-MBD arm.

Medication changes throughout the study were as follows: four participants in each condition who completed treatment had either a change of medication or change in dose of existing medication; one participant in each condition continued baseline medications without changes; one participant in each condition stopped stimulant medication for the summer; and two IPT and one TAU participants were not prescribed medication. Therefore, there was no difference in medication usage between treatment conditions.

4 |. DISCUSSION

The goal of this pilot randomized trial of IPT-MBD compared to TAU was to examine the feasibility and acceptability of an adapted version of IPT-A for youth with DMDD/SMD, a highly impaired population. The feasibility of delivering a time-limited psychotherapy for these youth is supported by the 80% completion rate for those randomized to IPT-MBD. In addition, IPT-MBD participants attended ≥80% of therapy sessions demonstrating the feasibility and acceptability of this intervention in a mental health outpatient setting. Participants and parents also reported high satisfaction with IPT-MBD.

This study also explored the preliminary efficacy of IPT-MBD compared to TAU. Those youth who received IPT-MBD demonstrated greater improvement over the course of treatment compared to those in TAU, evident in the CGI scores. Since the CGI scores for severity and improvement are specific for DMDD/SMD symptoms, improvement in these scores denotes change in the symptoms of irritability and outbursts. Participants in the IPT-MBD group were rated by an IE as mildly ill on average, whereas those who received TAU were rated as moderately ill. IPT-MBD participants were also rated as having more improvement in symptoms and overall functioning compared to TAU. Effect sizes were robust but included large CI, likely due to the small sample size. There were no statistically significant differences in depression or anxiety ratings between the two groups. There was no difference in the rate of change in irritability scores in the longitudinal model between the two groups, likely due to small sample size since irritability ratings were significantly improved in the IPT-MBD group compared to TAU.

As mentioned, there are limited data on interventions for youth with DMDD. There are no other psychosocial treatment studies for adolescents with DMDD. This is the first study focusing on both an adolescent population with DMDD and focusing on improving interpersonal interactions and relationships, which is fundamental to addressing the pervasive social impairment associated with this disorder.

There are limitations to the study. Despite randomization, there were differences in the baseline characteristics between conditions with IPT-MBD being predominantly female (70%) and TAU predominantly male (89%), which could affect generalizability of results to males and the preliminary efficacy estimates, since they were not adjusted for gender differences. Though these results need to be confirmed in a larger trial, we feel the gender imbalance did not critically bias the results. When examining the distribution of functional measures at baseline by gender, scores were on average the same or lower in males. Gender would have been a strong positive confounder if males had higher scores at baseline, which did not occur.

Another limitation is the small sample size, which may limit the ability to find significant differences between treatment conditions; however, the primary goal of the study was to assess feasibility and acceptability and these results should be interpreted as preliminary. Also, one therapist delivered the majority of the therapy, so symptom improvement could be an effect of the particular therapist’s skill rather than an effect of the treatment intervention itself. Therefore, future studies should have more therapists treating participants to confirm the effect of the treatment itself independent of therapist effect. Despite these limitations, the outcome efficacy data demonstrated potential impact of the intervention on functional improvement. These preliminary results should be interpreted in light of the discrepancy in number of therapy sessions between the groups. While the IPT group had an average of 20 sessions in 24 weeks, the TAU group varied in number of sessions ranging from 0 to 16, with some receiving no or few sessions of therapy, despite referrals given to all participants. Therefore, the improvement in symptoms in the IPT-MBD group could be due to higher dose effect of receiving more treatment, rather than the intervention itself.

Also, while there was evaluator training for the KSADS-PL and clinician measures, there was no interrater reliability conducted. Additionally, while there was IE assessment of change in irritability and outbursts over time, as well as, a global assessment of functioning measure, a specific measure to assess social functioning in youth with DMDD/SMD would have added useful information. Moreover, treatment satisfaction was not assessed in the TAU condition. Though the goal of this pilot study was to assess the acceptability of a new therapy, inclusion of a satisfaction instrument for both treatment conditions would have been informative. Lastly, the protocol allowed for medication changes throughout the course of treatment. Given the study’s 24-week duration, not permitting medication changes could have presented feasibility and ethical challenges. However, there were a similar number and type of medication changes in each group, so the improvement in symptoms in the IPT-MBD group is unlikely due to medication changes.

5 |. CONCLUSIONS

In summary, IPT-MBD appears to be a feasible and acceptable psychosocial intervention for youth with DMDD/SMD delivered alone or in combination with medication. The preliminary outcome data suggest the intervention holds promise for improving the psychosocial outcomes of these youth and warrants further investigation. Due to the social impairment associated with having DMDD/SMD, a psychotherapy focusing on improving interpersonal relationships may be particularly beneficial for improving short- and long-term outcomes for these youth. Next steps should include a larger study utilizing an active control condition with similar dose of treatment to further assess efficacy of IPT-MBD on mood and functioning outcomes for adolescents diagnosed with DMDD.