Interpersonal Psychotherapy for Adolescents and Young Adults

Aim: To analyze the effects and the usefulness of IPT for adolescents with depression .

Site Norrköping Child and Adolescent Psychiatric Services BUP Norrköping receives approximately 820 new patient every year, and about 285 of these are diagnosed with depression. Seven therapists are working with IPT for adolescents. (Depression data)

Patients Inclusion and exclusion criteria will be evaluated using the Mini-International Neuropsychiatric Interview.

Therapists The requirement of therapists is to have IPT training of at least level B. All therapists will be employed within the Child and Adolescent Psychiatric Services, and work under the regulations of the Health Care Act.

Sample Size It is estimated that recruiting and treating 15-20 patients a year will be feasible. In four years, 60 adolescents with depression will be recruited.

Statistical Power Power was calculated by running Monte Carlo simulation analyses. The difference between treatment and baseline was assumed to be similar to the difference between treatment and waiting lists in between-group randomized trials, which is usually above d = 0.80. Power was evaluated for the same statistical method that will be used in outcome analysis. Assuming an effect size of d = 0.80 and a sample size of 60, statistical power for the difference between the treatment and baseline slopes was 84%.

Measurement Instruments For outcome measures, longer scales are used before starting and at the end of treatment, and shorter measures are filled out prior to each session. The process measures are used each session, and match roughly the theoretically identified IPT change mechanisms (Lipsitz \& Markowitz, 2013).

Outcome Measures: Depression (i) Montgomery Åsberg Depression Rating Scale is a widely used measurement instrument for depressive symptoms. It has been validated for use with adolescents. The observer version of MADRS will be administered before waiting list, at treatment start, and at termination, using the structured interview SIGMA.

(ii) Patient Health Questionnaire - 9M is based on the DSM criteria for major depressive disorder, with nine items focusing on depressive symptoms and ask the individual to rate the duration for which they experienced each symptom. The 9M version is adjusted specifically to the diagnostic criteria for adolescent depression. Items are scored on a Likert scale of 0 (not at all) to 3 (nearly every day), with a total score ranging from 0 to 27. The PHQ-9 has demonstrated high sensitivity (89.5%) and good specificity (78.8%) in detecting depression in adolescents in a care sample. The PHQ-9 will be administered before the start of each session.

Change Process Measures Rated by External Observers (i) Symptom-Specific Reflective Functioning is a measure of the capacity for understanding connections between psychiatric symptoms, mental states, and interpersonal relationships; in other words, mentalizing about psychiatric symptoms. This capacity has been postulated to be especially relevant for IPT, due to its focus on connecting symptoms with interpersonal relationships and emotions. External raters will rate SSRF from the first few minutes of each session, in which IPT therapists ask about symptoms and interpersonal relationships from the previous week.

Change Process Measures Rated by Therapists (i) Working Alliance Inventory - Short form Revised, therapist version is a 12-item measure of the quality of the working alliance, i.e., collaborative relationship, between therapist and patient. This measure is administered to therapists immediately after each session.

Change Process Measures Rated by Patients (i) Session Alliance Inventory is a six-item patient-report measure of the quality of the working alliance, i.e., collaborative relationship, between therapist and patient during the previous session. This measure is administered to both patients and therapists immediately after each session.

(ii) Brief Mentalized Affectivity Scale is a validated patient-report measure for assessing emotion regulation. The short version (12 items) will be filled out immediately before each session.

(iii) Social Support Questionnaire - 6 asks about the person's available network for social support, the availability and quality of support from the network. This will be filled out at treatment start, at mid-treatment, and at treatment termination.

Study 1: Test the Effectiveness of IPT using a Quasi-Experimental Multiple-Baseline Design.

A non-concurrent multiple baseline design will be used, that is, not all patients start treatment at the same time. The sequence will be an A1BA2 design, in which A means no treatment and B means treatment. A1 is the baseline and A2 is the follow-up period. The distinction between A1 and A2 is due to the fact that reversibility of effects in the follow-up phase are not expected. The fact that waiting times vary quasi-randomly according to availability in the therapists' calendars will be utilized, with a minimum waiting of three weeks to establish a slope for the baseline period. Data for Study 1 will be analyzed using a piecewise multilevel growth curve model, in which two separate slopes for change over time in the outcome measure - one for the baseline phase and one for the treatment phase - are compared. If treatment is effective, symptoms should improve at a faster rate during treatment than before treatment. Thus, the test is the average change rate during the treatment phase minus the average change rate during the baseline phase.

Study 2: Mechanisms of change in IPT-A for depression Longitudinal (session-by-session) observation study using data from the treatment phase (B). Specifically, the proposed change mechanisms (mentalization, emotion regulation, social support, and working alliance) are measured each session, and used to predict next-session outcome. In addition, post-treatment qualitative interviews with patients will be used to retrospectively study the participants' experiences of change and what contributed to this. For Study 3, disaggregated time-lagged panel models will be used, in which candidate mechanisms of change are used as predictors of change in the outcome measure from one session to the following session.

Study 3: Predictors of improvement in IPT-A for depression Longitudinal observation design using baseline data on patients to predict differential trajectories during the treatment- and follow-up phases to explore for which patients IPT-A is effective. Session-wise measurements will be used to enable tests of causal direction for the candidate mechanisms. Multilevel growth curve modeling using pre-treatment variables as predictors of the intercepts and slopes of the outcome measure will be used. The cross-level interaction between a pre-treatment variable and the slope of time is the test of interest. Based on previous research, severity of symptoms (depression/eating disorder), comorbid anxiety disorder, the presence and severity of parent-child conflicts, quality of interpersonal relationships, experiences of bullying, and school functioning will be explored as potential predictors of IPT effectiveness.