The main aim of the project is to develop couple therapy for depression with special interest in dialogical and narrative processes in therapy. Parenthood, and different psychosocial phenomena, such as domestic violence, suicidal behaviour and alcohol abuse in relation to the client's specific depressive state will be analysed. The second main aim of the study is to analyse the effectiveness couple therapy for depression as a real world study. The good and poor outcome of different therapy processes will be compared and thus information about precise change processes will be reached. Owing to the real world setting, the inclusion criterion is a broad one: a primary diagnosis of depression (F32 and F33 in ICD-10).

The project will be conducted as a multicentre study so that in each centre the included patients will be randomized into experiment and control groups. The study centres are (1) Pohjois-Savon sairaanhoitopiiri in Kuopio, (2) Länsi-Pohjan sairaanhoitopiiri in Tornio and Kemi and (3) HUS/ Jorvin sairaala regional psychiatric polyclinics in Espoo. In each province the entire adult population has free access to the psychiatric outpatient clinics in their specific catchment area and thus no selection for the patients is done.

In each of these research centres qualified family therapists will conduct the couple therapy in case of depressed patients in a team of two therapists. To be a responsible therapist in this study the minimum criterion is three years specialist- level family therapy training and thus a qualification as psychotherapist according to Finnish law. The therapist shall have special experience and interest in dialogical and narrative processes in therapy. All the family therapists will also have taken a training program that focuses on systemic family therapy, narrative therapies and open dialogue in a family setting. No special therapy manual has been written, but the treatment processes will be analyzed after the therapy in relation to the outcome. The treatment processes are expected to last for as long as required according to the specific needs of each couple. This will allow for the focus on external validity: Clients will be recruited in the course of daily clinical practice without any selection outside the diagnosis and the therapies are taking place for as long as needed. The analyses of the treatment processes in relation to outcome, has termed as progress research by Pinsof & Wynne (2000)

In couple therapy some guidelines are followed. Since every family therapist has training in systemic therapy, the therapy can be termed as systemic couple therapy with a special focus on dialogues and narratives. The two therapists and the couple form a system in which the first aim is to generate dialogue to increase understanding of couple's situation and the role of depression in their lives. To this end, spouses are encouraged to speak of subjects that they find most important. While one person is speaking, the other three take a reflective position in their inner dialogues and afterwards describe their thoughts about what was said. The shift between the speaking and listening positions in the topics that the couple choose is the basic element in each therapy process. In these processes a new joint language is created for experiences without words or experiences manifested in symptoms, marital problems or in problematic behaviour.

Three types of analysis will be conducted:

1. The effectiveness of systemic couple therapy will be analyzed by (a) the assessment of depression by the Hamilton Depression Rating Scale inventory and the BDI and SCL self-report ratings scales; (b) the assessment of couple satisfaction by the Dyadic Adjustment Scale (DAS) self-report rating scale; and (c) the assessment of the psychological status of the patient by GAF rating. In addition to these, (e) domestic violence, use of alcohol and other psychosocial factors will be analyzed in interviews. Assessments will be conducted before the outset of the treatment, after 6 months, after 12 months and after 18 months from the beginning of the treatment. Changes in the scales between experiment and control groups will be compared. Because of the small size of the experimental group and the naturalistic design, no noteworthy differences in the results are expected as such. To test hypothetical explanatory power a prove was conducted. In the power calculation, a hypothetical sample of 35 received the value .82 in cases where the mean difference between two measurements divided by the corresponding standard deviation was 0.6. Similarly, if the difference in the means between two measurements was 0.5, the explanatory power was .66. Differences that are strong enough cannot be expected if only the group differences in the mean values between two measurements are analyzed without relating the results to the therapy processes (point 2) and to the processes of change (point 3). The explanatory power may be higher, however, after the Swedish study site has been added to the sample. The entire sample size would then consist of about 55 patients in each group.
2. Treatment processes will be studied by analysing (a) video or audio recordings of the dialogues and (b) the entire treatment processes in a specific follow-up interview three months after termination of the couple therapy. This interview will be organized as a learning forum, that is, the follow-up interviews will be conducted in a way that enables the therapists to evaluate their work together with the couple and an outsider interviewer. The aim is at developing the family therapists' therapeutic skills. Reflective processes and learning on the part of the families are essential in this co-research practice (Andersen 2006). The family and all staff members included in the specific treatment process will be invited to this interview. First the therapy team will be interviewed about their experiences of the treatment, and then the family members will be interviewed about their experiences of the therapy and about their thoughts regarding the therapists' experiences. The aim of this interview is to gain information about the specific processes through which clients can perform their own psychological work with their problems.
3. Processes of changes will be studied by conducting (a) a Outcome Ratings Scale among clients before each session and a Session Rating Scale among both clients and therapists after each therapy session and; and by (b) analyzing the qualities of the couple therapies by Psychotherapy Q-Sort (PQS) categories (Jones & Jones, 1985). In the PQS 100 characteristics in 9 categories will be analyzed for one therapy session. The characteristics describe 1) attitude, behaviour and explicative experience of the patient; 2) actions and attitudes of the therapists; and 3) the nature of the therapeutic relationship, environment and atmosphere of the session (Sirigatti, 2004). Although the PQS has mostly been used in analysing manualized psychotherapies (Ablon & Jones, 2002), implications has recently been developed for non-prescriptive therapies (Ablon et al., 2005) and systemic therapy (Sirigatti, 2004). Both analyses will be related to the outcome of the couple therapies and thus systematic information will be obtained as to which type of dialogues are connected with good outcomes.

Control group is formed in each of the three centres. These clients will not have family therapy sessions during the first 12 months; instead, according to the manner usual in each centre they will visit a psychiatrist or other therapists. Clients will be randomly assigned to the experimental and control groups after the doctor admitting them has assessed their inclusion by diagnosis and the Hamilton Rating Scale and after the first research interview has been conducted.

The inclusion period started in February 2006 and lasted until summer 2007. The first follow-up interviews will continue until the end of 2009. In this specific project the analysis of the data will begin in 2007 and continue to the end of 2010 concerning this specific project. Thereafter the data will be available for post graduate and doctoral students under the supervision of the principal investigators.

Department of Psychology at the University of Jyväskylä is the study site, in which all the data will be gathered. In each centre the research data will be stored outside of patient records. Information of the ratings scales can be added to specific patient records.

Inclusion Criteria:

• A primary diagnosis of depression (F32 and F33 in ICD-10);
• Under 65 years of age, living in a heterosexual couple relationship;
• Contacting the study centres either by oneself or by a referral because of depression (F32 or F33 in ICD-10) rating at least 14 on the Hamilton Rating Scale.

Exclusion Criteria:

• Violence;
• Drug abuse;
• Clear psychotic symptoms, organic brain disorder or severe suicidal behaviour that prevents participation in therapy discussions;
• Earlier family or couple therapy due to depression during the two years before the index treatment.