A randomized trial of the effect of four forms of psychotherapy on depressive and anxiety disorders

77 Paul Knekt and Olavi Lindfors eds. Kela 2004 Studies in social security and health A randomized trial of the effect of four forms of psychotherapy on depressive and anxiety disorders Design, methods, and results on the effectiveness of short-term psychodynamic psychotherapy and solution-focused therapy during a one-year follow-up

A RANDOMIZED TRIAL OF THE EFFECT OF FOUR FORMS OF PSYCHOTHERAPY ON DEPRESSIVE AND ANXIETY DISORDERS

KELA THE SOCIAL INSURANCE INSTITUTION, FINLAND Studies in social security and health 77 A randomized trial of the effect of four forms of psychotherapy on depressive and anxiety disorders Design, methods, and results on the effectiveness of shortterm psychodynamic psychotherapy and solution-focused therapy during a one-year follow-up Edited by Paul Knekt and Olavi Lindfors YHTEENVETO Satunnaistettu kliininen koe neljän psykoterapiamuodon vaikuttavuudesta masennustiloihin ja ahdistuneisuushäiriöihin Asetelma, menetelmät ja tulokset lyhyen psykodynaamisen psykoterapian ja voimavarasuuntautuneen terapian vaikuttavuudesta yhden vuoden seurannan aikana Helsinki 2004

The publications in this series have undergone a formal referee process. Cover: Petra Niilola and Riitta Nieminen Picture: Harri Heikkilä ISBN 951-669-639-2 ISSN 1238-5050 Printed by Edita Prima Ltd. Helsinki 2004

AUTHORS Paul Knekt, P.H.D., Research Professor The Social Insurance Institution and National Public Health Institute, Helsinki paul.knekt@ktl.fi Olavi Lindfors, Lic. Psych., Psychoanalyst (IFPS) Biomedicum Helsinki, Helsinki olavi.lindfors@hus.fi Camilla Renlund, M.D., Psychoanalyst (IPA) Biomedicum Helsinki, Helsinki camilla.renlund@hus.fi Markku Kaipainen, M.D., Ph.D., Psychiatrist Hospital District of Helsinki and Uusimaa, Department of Psychiatry, Helsinki University Central Hospital, Helsinki markku.kaipainen@hus.fi Päivi Mäkelä, M.Pol.Sc., Researcher Biomedicum Helsinki, Helsinki paivi.makela@hus.fi Aila Järvikoski, Ph.D., Professor Rehabilitation Foundation, Helsinki aila.jarvikoski@kuntoutussaatio.fi Timo Maljanen, M.Pol.Sc., Researcher The Social Insurance Institution, Helsinki timo.maljanen@kela.fi Mauri Marttunen, M.D., Ph.D., Professor Biomedicum Helsinki, Helsinki mauri.marttunen@hus.fi Raimo Raitasalo, Ph.D., Docent The Social Insurance Institution, Helsinki raimo.raitasalo@kela.fi

Tommi Härkänen, Ph.D., Statistician Biomedicum Helsinki, Helsinki tommi.harkanen@ktl.fi Esa Virtala, Data Manager Biomedicum Helsinki, Helsinki esa.virtala@ktl.fi Harri Rissanen, Research Assistant Biomedicum Helsinki, Helsinki harri.rissanen@ktl.fi Hanna Laine, M.A. (Educ.), Researcher Biomedicum Helsinki, Helsinki hanna.laine@hus.fi Jorma Hannula, M.D., Psychiatrist Biomedicum Helsinki, Helsinki jorma.hannula@marenovum.fi Veikko Aalberg, M.D., Ph.D., Docent Hospital District of Helsinki and Uusimaa, Hospital for Children and Adolescents, Helsinki University Central Hospital, Helsinki veikko.aalberg@hus.fi and the Helsinki Psychotherapy Study group (Appendix 1)

Abstract Knekt P, Lindfors O, eds. A randomized trial of the effect of four forms of psychotherapy on depressive and anxiety disorders. Design, methods, and results on the effectiveness of short-term psychodynamic psychotherapy and solution-focused therapy during a one-year follow-up. Helsinki: The Social Insurance Institution, Finland, Studies in social security and health 77, 2004. 112 pp. ISBN 951-669-639-2. The Helsinki Psychotherapy Study (HPS) is a randomized clinical trial comparing the effectiveness of four forms of psychotherapy in the treatment of depressive and anxiety disorders. A total of 367 Finnish psychiatric outpatients from the Helsinki region, 20 46 years of age and suffering from depressive or anxiety disorders, were recruited for the study in 1994 2000. A total of 326 patients were randomly assigned to one of 3 treatment groups: solution-focused therapy, short-term psychodynamic psychotherapy, and long-term psychodynamic psychotherapy. The patients assigned to the longterm psychodynamic psychotherapy group and 41 patients self-selected for psychoanalysis were included in a quasi-experimental design. The primary outcome measures were depressive and anxiety symptoms, while secondary measures included work ability, need for treatment, personality functions, social functioning, and life style. Cost-effectiveness was determined. The data were collected from interviews, questionnaires, psychological tests, and public health registers. The outcome measures were assessed up to 9 times during a 5-year follow-up. Patients on short-term psychodynamic psychotherapy and solution-focused therapy showed considerable decline in depressive and anxiety symptoms during the first year of follow-up, whereas work ability, personality functions, and social functioning were only slightly improved. The result did not differ between the 2 forms of therapy; both types are thus effective in the treatment of depressive and anxiety disorders but for the majority of patients they are not sufficient in producing recovery. Longer follow-ups are needed to evaluate the duration of treatment effects in the 2 groups. The HPS is one of the largest clinical trials on the effect of psychotherapy in the treatment of depressive and anxiety disorders. The results are likely to be incorporated into clinical practice and to impact public health. Key words: anxiety disorder, depressive disorder, psychodynamic psychotherapy, psychoanalysis, randomization, solution-focused therapy, clinical trial

Tiivistelmä Knekt P, Lindfors O, toim. Satunnaistettu kliininen koe neljän psykoterapiamuodon vaikuttavuudesta masennustiloihin ja ahdistuneisuushäiriöihin. Asetelma, menetelmät ja tulokset lyhyen psykodynaamisen psykoterapian ja voimavarasuuntautuneen terapian vaikuttavuudesta yhden vuoden seurannan aikana. Helsinki: Kela, Sosiaali- ja terveysturvan tutkimuksia 77, 2004. 112 s. ISBN 951-669-639-2. Helsingin Psykoterapiaprojekti (HPTP) on satunnaistettu kliininen koe, jossa verrataan neljän psykoterapiamuodon vaikuttavuutta masennustilojen ja ahdistuneisuushäiriöiden hoidossa. Tutkimukseen valittiin kaikkiaan 367 iältään 20 46-vuotiasta masennustilasta tai ahdistuneisuushäiriöstä kärsivää, Helsingin seudulla asuvaa avohoitopotilasta vuosina 1994 2000. Yhteensä 326 potilasta satunnaistettiin yhteen kolmesta hoitoryhmästä: voimavarasuuntautuneeseen (ratkaisukeskeiseen) terapiaan, lyhytkestoiseen psykodynaamiseen psykoterapiaan ja pitkäkestoiseen psykodynaamiseen psykoterapiaan. Pitkään psykodynaamiseen psykoterapiaan satunnaistetut ja 41 psykoanalyysiin itse ohjautunutta potilasta sisällytettiin kvasikokeelliseen tutkimusasetelmaan. Ensisijaisia vaikuttavuusindikaattoreita olivat masennus- ja ahdistuneisuusoiremittarit. Muita arviointikohteita olivat työkyky, hoidon tarve, persoonallisuuden toimivuus, sosiaalinen toimintakyky ja terveyskäyttäytyminen. Hoitojen kustannusvaikuttavuus arvioitiin. Tiedot kerättiin haastattelu- ja kyselymenetelmillä ja psykologisilla testeillä sekä julkisista terveydentilaa koskevista rekistereistä. Mittauksia tehtiin yhdeksän kertaa viiden vuoden seuranta-aikana. Sekä lyhyessä psykodynaamisessa psykoterapiassa että voimavarasuuntautuneessa terapiassa masennus- ja ahdistuneisuusoireilu vähenivät merkittävästi, kun taas työkyvyssä, persoonallisuuden toimivuudessa ja sosiaalisessa toimintakyvyssä ilmeni vain vähäistä paranemista yhden vuoden seurannan aikana. Tulokset olivat samankaltaiset kummassakin terapiamuodossa. Molemmat terapiamuodot sopivat siten masennustilojen ja ahdistuneisuushäiriöiden hoitoon, mutta suurimmalle osalle potilaista ne eivät riitä tuottamaan täyttä toipumista. Tarvitaan myös pidempiä seurantaaikoja vaikutuksen pysyvyyden arvioimiseksi molemmissa ryhmissä. HPTP on yksi laajimmista kliinisistä kokeista, jossa selvitetään masennustiloihin ja ahdistuneisuushäiriöihin kohdistuvan psykoterapian vaikuttavuutta. Tulokset tulevat mitä ilmeisimmin vaikuttamaan kliinisten käytäntöjen kehittymiseen ja niiden soveltamiseen terveydenhuollossa. Suomenkielinen yhteenveto s. 82 92. Avainsanat: ahdistuneisuushäiriö, kliininen koe, masennustila, psykoanalyysi, psykodynaaminen psykoterapia, satunnaistaminen, voimavarasuuntautunut terapia

Abstrakt Knekt P, Lindfors O, red. En experimentell studie rörande effekten av fyra psykoterapiformer på depression och ångestsyndrom. Design, metoder och resultat rörande effektiviteten av kort psykodynamisk psykoterapi och lösningsinriktad terapi under ett års uppföljningstid. Helsingfors: Fpa, Social trygghet och hälsa: undersökningar 77, 2004. 112 s. ISBN 951-669-639-2. Helsingfors Psykoterapistudie (HPS) är en randomiserad klinisk experimentell studie, som jämför effektiviteten av fyra terapiformer vid vården av depression och ångestsyndrom. Sammanlagt 367 finländska psykiatriska patienter från öppenvården i Helsingforsregionen, 20 46 år gamla och lidande av depression eller ångestsyndrom, rekryterades för denna studie åren 1994 2000. Sammanlagt 326 patienter randomiserades till en av tre vårdgrupper: lösningsinriktad terapi, kort psykodynamisk psykoterapi och lång psykodynamisk psykoterapi. Patienterna som randomiserats till lång psykodynamisk terapi och 41 patienter som själva valt psykoanalys inkluderades i en kvasiexperimentell studie. Primära effektivitetsmått var symptom på depressivitet och ångest. Sekundära mått var arbetsförmåga, behov av vård, personlighetsfunktioner, social funktionsförmåga och livsstil. Kostnadseffektiviteten bestämdes. Informationen insamlades med hjälp av intervjuer, enkäter och psykologiska tester och från allmänna hälsoregister. Mätningarna gjordes upp till nio gånger under en 5-årig uppföljningsperiod. Så väl depressions- som ångestsymptomen minskade betydligt under det första uppföljningsåret både bland patienterna som fick kort psykodynamisk psykoterapi och bland dem som fick lösningsinriktad terapi. Däremot förbättrades arbetsförmågan, personlighetsfunktionerna och den sociala funktionsförmågan endast i mindre mån. Resultaten var likadana för vardera vårdgruppen. Båda formerna av terapi lämpar sig således för vården av depression och ångestsyndrom men den största delen av patienterna blir därigenom inte helt återställda. Längre uppföljningstider är nödvändiga för att man skall kunna bestämma effektens varaktighet i de två terapigrupperna. Resultaten av studien förväntas bli inkorporerade i klinisk praxis och har följdaktligen betydelse för befolkningens hälsa. Nyckelord: depression, klinisk experimentell studie, lösningsinriktad terapi, psykoanalys, psykodynamisk psykoterapi, randomisation, ångestsyndrom

PREFACE An average of one person out of five suffers from some mental disorder in Finland. Thus mental disorders cause considerable subjective suffering, use of health services, and work disability. About 50% of all mental disorders are either mood or anxiety disorders. Although the effectiveness of psychotherapy as a treatment for mental disorders has been documented, there is insufficient evidence for the basis of treatment choice among the psychotherapy modalities most used in depressive and anxiety disorders. The Helsinki Psychotherapy Study (HPS) was initiated in 1994 to evaluate the effects of 4 psychotherapy modalities (solution-focused therapy, shortterm and long-term psychodynamic psychotherapy, and psychoanalysis) in the treatment of patients with depressive or anxiety disorders. The study is funded mainly by the Social Insurance Institution and carried out in cooperation with the Hospital District of Helsinki and Uusimaa, Biomedicum Helsinki, National Public Health Institute, and Rehabilitation Foundation. The study group is located at the Department of Psychiatry of Helsinki University Central Hospital. A total of 150 persons have functioned in some professional role during the study project, of whom 71 therapists were responsible for actually carrying out the therapies of the patients referred to the study. This report was compiled after the solution-focused therapy and short-term psychodynamic psychotherapy had ended and includes a description of the design, methods, and participants of the study, as well as the effectiveness results during a one-year follow-up after initiation of these short-term therapies. The Helsinki Psychotherapy Study group is looking forward to promoting further research-based decisions in treatment choices for depression and anxiety disorders. Our warmest thanks go to the Board of the Social Insurance Institution for the support without which this research would not have been possible. We also want to thank all the persons involved in the study in its different phases. We are grateful to the patients for their cooperation. Special thanks are due to all the persons working at the Psychiatric Clinic of the Helsinki University Central Hospital, at the Department of Health and Functional Capacity in the National Public Health Institute, and at the Information Systems Department of the Social Insurance Institution whose dedication and long-term commitment have been needed to compile this publication. Helsinki, March 2004 Paul Knekt

ESIPUHE Noin joka viides suomalainen kärsii mielenterveyden häiriöstä. Mielenterveyden häiriöt aiheuttavat paitsi subjektiivista kärsimystä myös huomattavaa terveyspalvelujen käyttöä ja työkyvyttömyyttä. Mielenterveyden häiriöistä noin puolet on joko mieliala- tai ahdistuneisuushäiriöitä. Vaikka on tiedossa, että psykoterapia soveltuu mielenterveyden häiriöiden hoitoon, tiedot Suomessa yleisesti käytettyjen psykoterapiamuotojen soveltuvuudesta masennustilojen ja ahdistuneisuushäiriöiden hoitoon ovat puutteellisia. Helsingin Psykoterapiaprojekti aloitettiin vuonna 1994 neljän terapiamuodon voimavarasuuntautuneen terapian, lyhyen ja pitkän psykodynaamisen psykoterapian ja psykoanalyysin vaikuttavuuden arvioimiseksi masennustilojen ja ahdistuneisuushäiriöiden hoidossa. Projekti on pääosin Kelan rahoittama, Helsingin ja Uudenmaan sairaanhoitopiirin, Biomedicum Helsinki -säätiön, Kansanterveyslaitoksen ja Kuntoutussäätiön kanssa tehtävä yhteistyöhanke, jota toteutetaan Helsingin yliopistollisen keskussairaalan psykiatrian klinikalla. Yhteensä noin 150 henkilöä on toiminut jossain työtehtävässä projektissa. Projektiin rekrytoitujen potilaiden terapioista huolehti 71 terapeuttia. Voimavarasuuntautuneen terapian ja lyhyen psykodynaamisen psykoterapian päätyttyä laadittiin tämä raportti, joka kuvaa projektissa käytettyjä aineistoja, tutkimusasetelmia ja tutkimusmenetelmiä sekä kahden lyhytkestoisen terapian vaikuttavuutta vuoden seurannassa. Tutkijaryhmä toivoo nyt käyttöön tulevien tietojen edistävän päätöksentekoa masennustilojen ja ahdistuneisuushäiriöiden hoitomenettelyjen valinnassa. Osoitamme lämpimät kiitokset Kelan hallitukselle, joka tuellaan on mahdollistanut tämän tutkimuksen toteutumisen. Kiitokset myös kaikille hankkeeseen sen eri vaiheissa osallistuneille henkilöille. Olemme kiitollisia potilaille yhteistyöstä, joka mahdollisti tutkimusaineiston muodostamisen. Erityiskiitokset kuuluvat Psykiatrian klinikalla, Kansanterveyslaitoksen terveyden ja toimintakyvyn osastolla ja Kelan tietojenkäsittelyosastolla työskenteleville henkilöille, jotka sinnikkäällä ja pitkäjännitteisellä työllä ovat huolehtineet siitä, että tämä raportti voidaan nyt julkaista. Helsingissä maaliskuussa 2004 Paul Knekt

CONTENTS 1 INTRODUCTION... 15 1.1 Background... 15 1.2 Aim of the study... 16 2 STUDY POPULATION AND METHODS... 18 2.1 Patients and settings... 18 2.2 Study design... 18 2.3 Treatments... 19 2.4 Therapists... 20 2.5 Compliance... 20 2.6 Assessments... 22 2.7 Follow-up of population registers... 28 2.8 Data monitoring... 28 2.9 Quality control... 28 2.10 Statistical analysis... 29 2.11 Ethics... 31 2.12 Patient enrollment... 31 2.13 Baseline characteristics... 33 2.14 Discussion... 44 3 RESULTS... 46 3.1 Participation in therapy and compliance... 46 3.2 Dropout during follow-up... 48 3.3 Treatment effects... 49 3.4 Discussion... 68 4 SUMMARY... 73 4.1 Introduction... 73 4.1.1 Background... 73 4.1.2 Aim of study... 74 4.2 Population and methods... 74 4.2.1 Study design... 74 4.2.2 Patients... 74 4.2.3 Treatments and therapists... 75 4.2.4 Assessments... 76 4.2.5 Follow-up of population registers... 77 4.2.6 Quality control... 77 4.2.7 Statistical analysis... 77 4.2.8 Ethics... 78

4.3 Results... 78 4.4 Conclusions... 80 5 YHTEENVETO... 82 5.1 Johdanto... 82 5.1.1 Tausta... 82 5.1.2 Tutkimuksen tarkoitus... 83 5.2 Aineisto ja menetelmät... 84 5.2.1 Tutkimusasetelma... 84 5.2.2 Potilaat... 84 5.2.3 Terapiat ja terapeutit... 84 5.2.4 Mittausmenetelmät... 86 5.2.5 Rekisteriseuranta... 88 5.2.6 Laaduntarkkailu... 88 5.2.7 Tilastolliset menetelmät... 88 5.2.8 Eettiset kysymykset... 89 5.3 Tulokset... 89 5.4 Päätelmät... 92 REFERENCES... 93 APPENDICES... 101

1 INTRODUCTION 1.1 Background Mood and anxiety disorders are among the most prevalent psychiatric disorders, causing considerable subjective suffering, use of health services, and working disability, thus making them a major public health concern (Murray and Lopez 1997). The prevalence of mood disorders varies from 4 11%, and that of anxiety disorders from 4 17% (Aromaa and Koskinen 2002; The WHO International Consortium in Psychiatric Epidemiology 2000). Psychotherapies are widely applied in the treatment of depressive and anxiety disorders. The effectiveness of psychotherapy in general, and particularly shortterm psychotherapies, has been demonstrated in comparative trials (Barlow and Lehman 1996; Clarkin et al. 1996; Robinson et al. 1990). Short-term psychodynamic psychotherapy is widely used in clinical practice (Anderson and Lambert 1995). Developed from psychoanalytic orientation, short-term psychodynamic psychotherapy aims to produce lasting changes through the explorative resolution of focal conflicts in the transference relationship. Patients assessed as appropriate for the treatment are expected to undergo significant personality changes and gain symptomatic relief (Malan 1976; Strupp and Binder 1984). Thus far, about 20 trials have been published on this type of psychotherapy dealing with the treatment of depressive and anxiety disorders or symptoms (Crits-Christoph 1992; Leichsenring 2001). Short-term psychodynamic psychotherapy has been shown to have better effectiveness compared with being on a waiting list (Shefler et al. 1995) or with outpatient treatment in general (Guthrie et al. 1999). When compared with other psychotherapies, the results appear inconsistent. Some studies have reported the effectiveness of short-term psychodynamic psychotherapy to be lower than that of cognitive, behavioral, or cognitive-behavioral therapy (Barkham et al. 1999; Durham et al. 1994; Shapiro and Firth 1987), while others have found no difference in effectiveness (Brockman et al. 1987; Gallagher-Thompson and Steffen 1994; Hersen et al. 1984; Pierloot and Vinck 1978). Theories of change and the explorative therapeutic technique of psychodynamic therapies have recently been challenged by solution-focused therapy with its strength and resource-oriented approach and a social constructionism orientation. Solution-focused therapy was developed from therapies applying a problem solving approach and systemic family therapy (Gingerich and Eisengart 2000). It emphasizes the identification of a problem, collabo- 15

rative efforts to maintain a focus on the problem, and the limiting of therapy to find a solution to the problem while maintaining a positive working alliance (Johnson and Miller 1994; Lambert et al. 1998). Problem solving therapy has been shown to be more effective in the acute treatment of depression compared with placebo or no treatment (Dowrick et al. 2000; Mynors-Wallis et al. 1995), and as effective as antidepressant medication (Mynors-Wallis et al. 2000). Solution-focused therapy has been shown to have significant beneficial effects on outpatients with varying psychiatric disorders. Compared with other therapies, the treatment effects have emerged already after a few sessions (Lambert et al. 1998). Apparently, however, no randomized clinical trials on the effectiveness of this form of therapy in comparison with shortterm psychodynamic psychotherapy in depressive or anxiety disorder have been published thus far. One widely applied form of psychotherapy is long-term psychodynamic psychotherapy, which commonly lasts at least 2 3 years, thus incurring considerable costs. Likewise, psychoanalysis, which typically lasts 5 years, aims to produce long-term beneficial effects that would surmount those elicited by shorter therapies. However, only a few naturalistic and quasi-experimental studies have been published and thus only scarce information is available so far on the effectiveness of these forms of psychotherapy in comparison with short-term psychotherapies (Doidge 1997). Accordingly, the information on cost-effectiveness between short- and long-term therapy is almost non-existing. 1.2 Aim of the study The Helsinki Psychotherapy Study (HPS) was initiated to evaluate the effects of long-term and short-term psychodynamic psychotherapy, solutionfocused therapy, and psychoanalysis in the treatment of depressive or anxiety disorders, excluding severe personality disorders. The primary objective was to compare the effects of the 4 different forms of psychotherapy, all estimated to be suitable for the patient group, on psychiatric symptoms, especially those relevant to depressive and anxiety disorders. The secondary aim of the study was to assess the effect of the intervention on the need for posttherapeutic treatment, social functioning and work ability, psychological functioning, and lifestyle, and to compare the cost-effectiveness of the different forms of psychotherapy. 16

The present report of the HPS describes the rationale, design, methods, and baseline characteristics of the participants and results of the comparison of short-term psychodynamic psychotherapy and solution-focused therapy in the treatment of depressive and anxiety disorders during a one-year followup after initiation of treatment. 17

2 STUDY POPULATION AND METHODS 2.1 Patients and settings Outpatients were recruited from psychiatric services in the Helsinki region from June 1994 to June 2000. Local psychiatrists were informed about the project and they carried out the first phase screening by referring to the project patients who were evaluated to fulfill the inclusion criteria and suitable for the treatments. The psychiatrists were working in private practice (29%), the community mental health care (20%), the primary health care (16%), the student health care (20%), and occupational health care (10%). The patients represented individuals normally treated by psychotherapy in southern Finland. Eligible patients were 20 45 years of age and had a longstanding (> 1 year) disorder causing social dysfunction in work ability. They had to meet the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) criteria (American Psychiatric Association 1994) for an anxiety or depressive disorder and be estimated on a psychodynamic scale of suffering from neurosis to high-level borderline disorder. Patients were excluded from the study for the following reasons: psychotic disorder or severe personality disorder, bipolar I disorder, adjustment disorder, substance abuse, organic brain disease or other severe organic disease, and mental retardation. Individuals treated with psychotherapy within the previous 2 years, psychiatric health employees and persons known to the research team members were also excluded. 2.2 Study design Two study designs were defined: randomized and quasi-experimental. In the randomized design, those patients who remained eligible at baseline were randomly assigned according to a central computerized randomization schedule in a 1:1:1.3 ratio to solution-focused therapy and to short-term and long-term psychodynamic psychotherapy. A research associate assigned consecutively numbered envelopes containing concealed assignment codes sequentially to eligible patients. The quasi-experimental design consisted of patients randomly assigned to long-term psychotherapy and patients self-selected and assessed to be suitable for psychoanalysis. 18

2.3 Treatments Solution-focused therapy is a brief therapy approach which emphasizes the identification of a problem and collaborative efforts to maintain a focus on finding a solution to the problem (Johnson and Miller 1994; Lambert et al. 1998). No single, accepted theory of solution-focused therapy exists but the approach is closely related to postmodernism, narrative theory, and language theory (Miller et al. 1995). In this study, the main components included were the search for pre-session change, (i.e. questions about change that occurred before treatment began) goal-setting, use of miracle and scaling questions, exploration of exceptions, use of one-way mirror and consulting break, and use of positive feedback and home assignments. The frequency of sessions was flexible, usually one every second or third week, up to a maximum of 12 sessions, over no more than 8 months. Short-term psychodynamic psychotherapy is characterized by the exploration of a focus, which can be identified by both the therapist and the patient. This consists of material from current and past interpersonal and intrapsychic conflicts. Therapist's role in this approach is active in creating the alliance and ensuring the time-limited focus (Malan 1976). Major therapeutic interventions used were confrontation, clarification and interpretation. The therapy was scheduled for 20 treatment sessions, one session a week, over 5 6 months. Long-term psychodynamic psychotherapy is characterized by a framework in which the central elements are exploration of unconscious conflicts, deficits and distortions of intrapsychic structures. Confrontation, clarification and interpretation were major elements, as well as therapist's actions in ensuring the alliance and working through in the therapeutic relationship (Gabbard 1994). The therapy process was oriented towards conflict resolution and greater self-awareness. The frequency of sessions was 2 3 sessions a week and the duration of therapy up to 3 years. In psychoanalysis, current and past interpersonal and intrapsychic conflicts, transference phenomena, and developmental arrests are explored. The central aim in this study therapy was the enhancement of self-awareness of unconscious motives, impulses, fears and conflicts and thereby a thorough restructuring of personality (Greenson 1985) as well as a new developmental growth process within the therapeutic relationship (Tähkä 1993). The frequency of sessions was 4 sessions a week, and the duration up to 5 years. 19

2.4 Therapists Altogether 71 therapists participated in the study. A total of 6 therapists gave solution-focused therapy, 12 short-term psychodynamic psychotherapy, 41 long-term psychodynamic psychotherapy, and 30 psychoanalysis. The therapists giving short-term and long-term psychodynamic psychotherapy and psychoanalysis were mainly psychologists (78%) whereas those giving solution-focused therapy had a more heterogeneous educational background (psychologists, physicians or social workers). The mean age of the therapists was 49 (SD = 6.6) years and 69% of them were women. The therapists had practiced for at least 2 years after training in the specific form of psychotherapy. The mean years of experience were 9 (SD = 4.8) for both short-term psychodynamic psychotherapy and the solution-focused groups of therapists and 18 (SD = 5.6) and 15 (SD = 5.4) years for the therapists giving longterm psychotherapy and psychoanalysis, respectively. The therapists giving short-term psychodynamic therapy had given long-term psychodynamic psychotherapy for an average of 16 years. The mean number of patients treated before this investigation in the specific form of therapy were 14 and 155, for the therapists giving short-term psychodynamic therapy and solution-focused therapy, respectively. The corresponding values for therapists giving long-term psychodynamic psychotherapy and psychoanalysis were 66 and 16 patients, respectively. The therapies were conducted in conformance with clinical practice. No therapy manuals were used and no video or audio taping was carried out during the sessions. 2.5 Compliance The external and internal quality of the study treatment and the use of auxiliary treatment during and after the study treatment were assessed. We also suggested all patients to undergo a washout by stopping the use of psychotropic medication one month before the baseline examination. External quality of study treatment The external quality of study treatment describes how well the treatment satisfied criteria based on the characteristics of the treatment intended. The characteristics considered were waiting time to beginning of the therapy, fre- 20

quency of sessions, duration of therapy, number of sessions, unusual breaks in treatment, change of therapist, and discontinuation of therapy. The waiting time to beginning of the therapy was defined as the number of days from the baseline measurement to the first session of therapy. The reason or reasons for the delay is given for subjects for whom the waiting time exceeded 90 days. The frequency of sessions of the study treatment was defined as the mean of the frequency reported at follow-up measurements during the treatment (taking differences in time intervals into account). For short-term psychodynamic therapy, the session frequency criterion was met if the mean session frequency was one session per week, whereas no restrictions were set for solution-focused therapy in which an individually tailored frequency and very brief therapy were acceptable. The duration of therapy was the number of days from the first to the last session. For both therapy groups the duration of therapy criterion was met if the therapy lasted for 245 days at most. For short-term psychodynamic therapy, the number of sessions criterion was met if the number of sessions was between 15 and 25, and for solution-focused therapy if the number of sessions was 15 at most, with no minimum set. Unusual breaks in treatment was rated positively, if individually for each patient, there were unplanned and disturbing interruptions of therapy for 2 weeks or more, as rated by the interviewer. Discontinuing study treatment was rated positively if the interviewer judged that the treatment was interrupted prematurely; in most cases, this was evident and in concordance with the patient s view. Internal quality of study treatment The internal quality of the treatment was monitored as the mean values of the patient s and therapist s evaluations of the Working Alliance Inventory (WAI) (Horvath and Greenberg 1989). The interviewer, who collected the follow-up data from the patients, evaluated the quality of patient-therapist cooperation, the quality of patient-therapist fit, and the degree of patient s experience of being heard and understood on the basis of the Psychotherapy Process Assessment (PPA). The assessment was performed at the first follow-up interview after therapy was terminated. The assessment methods are described in more detail in Appendix 2. 21

Auxiliary treatment during study treatment or follow-up Use of auxiliary treatment during and after the study treatment was assessed by collecting data on the use of psychotropic medication, psychotherapy, and psychiatric hospitalization. The use of psychotropic medication and, as a subgroup, of antidepressant medication, was rated positively if the patient in any follow-up measurement during study treatment reported regular use of psychotropic medication. The use of psychotherapy over and above the study treatment was rated positively if the interviewer at any time in follow-up during study treatment rated the patient as having initiated psychotherapy. Psychiatric hospitalization was rated positively if the patient in follow-up during study treatment reported periods of hospital treatment for psychiatric problems, hospital treatments due to suicide attempts, or any nonspecified treatment in a psychiatric hospital. 2.6 Assessments Approved methods were used for assessment of symptoms and psychiatric diagnosis, need for post-therapeutic treatment, work ability, personality functions, social functioning, lifestyle, and cost-effectiveness (Table 1). The measurements were carried out as ratings based on interviews or self-report questionnaires and as psychological tests. Laboratory measurements and follow-up of nationwide health registers were also carried out. Experienced clinical raters, who were not involved in patients assignments or treatment, conducted the interviews. The quality of the interview data (i.e. the agreement between raters and the long-term stability of ratings) was continuously controlled. The interviews, although not blinded (except at baseline), and the treatment sessions were carried out at separate physical locations. The assessments were completed at baseline examination and during the follow-up after 3, 7, and 9 months and 1, 1.5, 2, 3, 4, and 5 years. Questionnaires were administered at each of these occasions. The questionnaires administered after 3 and 9 months, and 1.5, 2, and 4 years were brief. Interviews were repeated 4 times, i.e., after 7 months, and 1, 3, and 5 years. Psychological tests and laboratory determinations were repeated after 3 and 5 years of follow-up. Furthermore, the end point data of the patients were measured when the study treatment was finished. The assessment methods are briefly mentioned as follows and described in more detail in Appendix 2. 22

Time of measurement (months) Assessment method Method < 0 0 3 7 9 12 18 24 36 48 60 Reference Psychiatric diagnosis and symptoms Psychiatric diagnosis (DSM-IV) Interview x x x x x (American Psychiatric Association 1994) Global Assessment of Functioning scale (GAF) Interview x x x x x (American Psychiatric Association 1994) Beck Depression Inventory (BDI) Questionnaire x x x x x x x x x x (Beck et al. 1961; Beck 1970) Symptom Check List, Questionnaire x x x x x x x x x x x (Derogatis et al. 1973) Global Severity Index (SCL-90-GSI) Symptom Check List, depression scale (SCL-90-DEP) Questionnaire x x x x x x x x x x x (Derogatis et al. 1973) Symptom Check List, anxiety scale (SCL-90-ANX) Questionnaire x x x x x x x x x x x (Derogatis et al. 1973) Hamilton Depression Rating Scale (HDRS) Interview x x x x x (Hamilton 1960; Williams 1988) Hamilton Anxiety Rating Scale (HARS) Interview x x x x x (Hamilton 1959; Bruss et al. 1994) Suicidal ideation (one item from HDRS) Interview x x x x x (Hamilton 1960; Williams 1988) Target Complaints (TC) Questionnaire x x x x x x x (Battle et al. 1966) Psychiatric Symptoms Questionnaire (PSQ) Interview x x x x x x x x x (Lindfors et al. 2004) Table 1. Assessment methods used in the Helsinki Psychotherapy Study. Post-therapeutic treatment Perceived need for post-therapeutic treatment Questionnaire x x x x x x x x x - Post-therapeutic treatment Register, Quest. x x x x x x x x - Work ability Work Ability Index Questionnaire x x x x x x x (Tuomi et al. 1997, 1998; Ilmarinen et al. 1997) SAS Work (work subscale of SAS-SR) Questionnaire x x x x x x x x x x (Weissman and Bothwell 1976) Perceived psychological functioning Questionnaire x x x x x x x (Lehtinen et al. 1991) Sick leave Questionnaire x x x x x - Table 1 continues. 23

Table 1 continued. Time of measurement (months) Assessment method Method < 0 0 3 7 9 12 18 24 36 48 60 Reference Personality functions Quality of Object Relations Scale (QORS) Interview x x x x x (Azim et al. 1991) Level of Personality Organization (LPO) Interview x x x x x (Kernberg 1984 and 1996) Defense Style Questionnaire (DSQ) Questionnaire x x x x x x x (Andrews et al. 1989) Structural Aspects of Social Behaviour (SASB) Questionnaire x x x x x x x (Benjamin 1996) Inventory of Interpersonal Problems (IIP) Questionnaire x x x x x (Horowitz et al. 1988) Rorschach Inkblot Method Test x x x (Exner 1993; Urist 1977; Cooper et al. 1988) Social functioning Perceived competence Questionnaire x x x x x x x x x x (Smith et al. 1991; Härkäpää 1995; Wallston 1990) Life Orientation Test (LOT) Questionnaire x x x x x (Scheier and Carver 1985) Social Adjustment Scale, self-report (SAS-SR) Questionnaire x x x x x x x x x x (Weissman and Bothwell 1976) Sense of Coherence scale (SOC) Questionnaire x x x x x (Antonovsky 1993) Life Situation Survey (LSS) Questionnaire x x x x x (Chubon 1987) Life style and somatic health Smoking, alcohol consumption, leisure time exercise, Questionnaire x x x x x x x (Aromaa et al. 1989) body mass index Serum determination Laboratory x x x - Diseases, hospitalization, use of medication Register, Quest. x x x x x x x x x x x - Health economics data Register, Quest. x x x x x x x x x x x - 24

Psychiatric diagnosis and symptoms The primary outcomes measured, specified a priori, were depressive and anxiety symptoms. The symptoms of depression were assessed with the 21- item Beck Depression Inventory (BDI) (Beck et al. 1961) and the observerrated Hamilton Depression Rating Scale (HDRS) (Hamilton 1960). The symptoms of anxiety were assessed with the self-reported Symptom Check List Anxiety scale (SCL-90-ANX) (Derogatis et al. 1973) and the observerrated 14-item Hamilton Anxiety Rating Scale (HARS) (Hamilton 1959). Self-reported remission among patients with depressive disorder was defined as a total score < 9 in the BDI. Observer-rated remission among patients with depressive disorder was defined a priori as a total score < 7 in the HDRS, and among those with anxiety disorder a total score of < 7 in the HARS. Dichotomous proxy measures covered the presence of symptoms for all patients treated using the higher categories of the BDI, HDRS, and HARS for participants and information on the presence of symptoms and need for treatment from the Psychiatric Symptoms Questionnaire (PSQ) (Lindfors et al. 2004) for dropouts. The secondary measures described general psychiatric symptoms and psychiatric diagnoses. The Symptom Check List Global Severity Index (SCL- 90-GSI) (Derogatis et al. 1973), the Global Assessment of Functioning scale (GAF) (American Psychiatric Association 1994) and the severity of the main complaint on the Target Complaints (TC) scale (Battle et al. 1966) were used as outcome measures of general psychiatric symptoms. Suicidal ideation was assessed with a single item from the observer-rated HDRS (Hamilton 1960; Williams 1988). Psychiatric diagnoses (axes I and II) were assessed according to the DSM-IV diagnostic criteria (American Psychiatric Association 1994) using a semistructured interview. When a subject had more than one diagnosis, all were listed. Other outcome measurements included self-report and interviewer-rated instruments. Need for post-therapeutic treatment The patient s perceived need for post-therapeutic treatment was gauged with a single self-report questionnaire item after the end of treatment. Information on the use of post-therapeutic treatment, i.e., prescription medication, psychotherapy or hospitalization was collected by self-report questionnaires, interview-rated questionnaires and from public registers. 25

Work ability Self-reported current ability to work was measured by a modification of the Work Ability Index (Ilmarinen et al. 1997; Tuomi et al. 1997 and 1998). Performance in work was also measured using the work subscale of the Social Adjustment Scale (SAS-SR) (Weissman and Bothwell 1976). Furthermore, perceived psychological functioning (Lehtinen et al. 1991) and sick leave from work were also measured. The incidence of more than 20 sick leave days during the preceding 3 months was used as an outcome measure. Personality functions Developmental level of object relations was assessed with the Quality of Object Relations Scale (QORS) (Azim et al. 1991). The personality organization was evaluated with the Level of Personality Organization (LPO) interview, a modification of Kernberg s (1981) structural interview. Psychological defense styles were assessed with the Defense Style Questionnaire (DSQ) (Andrews et al. 1989). Self-concept was measured with the Structural Aspects of Social Behavior (SASB) introject questionnaire (Benjamin 1996). Amount of interpersonal problems was measured with the self-report Inventory of Interpersonal Problems (IIP) (Horowitz et al. 1988). The Rorschach Inkblot Method, administrated according to the Comprehensive System, was used as a measure of personality organization and maladaptation (Exner 1993). The protocol was also coded for the Mutuality of Autonomy scale (MOA) (Urist 1977) and for defense categories (Cooper et al. 1988). Social functioning Perceived competence was measured with a modification of the Self-Performance Survey (Härkäpää 1995; Smith et al. 1991; Wallston 1990). Dispositional optimism was measured with the Life Orientation Test (LOT) (Scheier and Carver 1985) and social functioning with the Social Adjustment Scale (SAS-SR) (Weissman and Bothwell 1976). Coping was measured with the Sense of Coherence Scale (SOC) (Antonovsky 1993). Quality of life was measured with the Life Situation Survey (LSS) (Chubon 1987). 26

Lifestyle and somatic health Data on smoking, alcohol consumption, body mass index, leisure time exercise, and pregnancies were collected with a questionnaire, and serum determinations (e.g. cholesterol, glucose, thyroid hormone, and minerals) were performed. Somatic health was also determined with self-report questionnaires (perceived general health, disease symptoms, medication, and hospitalization) and with data from population registers (medication and hospitalization). Cost-effectiveness In the main analysis all relevant direct and indirect costs due to mental health problems were estimated and related to the effectiveness of the treatments. The perspective was societal, i.e. all costs were included regardless of one who bore these costs. In the 3-year and longer follow-up studies, the costs were discounted. Data on the use of different health care resources were collected from national health registers and, if necessary, completed with data from questionnaires. In the secondary analysis, costs due to somatic diseases were also included. Potential confounding and effect-modifying factors Sociodemographic (sex, age, marital status, education, and socioeconomic status), and psychiatric history data (age at onset of first psychiatric problems and duration of present disorder, recurrences of major depressive disorder, separation experiences, mental disorders in first degree relatives and lifetime occurrence of attempted suicides), psychiatric diagnoses (axes I and II), and suitability for psychotherapy (e.g. motivation and self-reflection ability) were assessed at baseline using questionnaire and interviews. Previous psychiatric treatment (psychotherapy, psychopharmacological treatment, and hospitalization) and somatic health (perceived general health, disease symptoms, medication, and hospitalization) were obtained by linking the study population to nationwide health registers. 27

2.7 Follow-up of population registers Individual information on use of psychiatric medication, use of rehabilitation services, days of sickness absence, hospitalization and mortality was obtained by linking the study population to nationwide registers using a unique personal identification number. Information on use of psychotropic medication, on use of rehabilitation services and on sick leave periods came from the medicine, rehabilitation, and sick leave reimbursement registers of the Social Insurance Institution. Information on hospitalization for mental reasons was obtained from the hospital discharge register (Heliövaara et al. 1984). Mortality data were identified from death certificates obtained from Statistics Finland for all the deceased. The follow-up of the population registers was started 2 years before the baseline examination and continues to the end of the study 5 years after baseline. 2.8 Data monitoring General adherence to the study protocol was continuously evaluated by monitoring recruitment success, dropout rates, timeliness and completeness of form handling, and accuracy of the database. Treatment group balance for confounding factors, including disorder factors and information on the therapy process, was continuously evaluated. Other comparisons included dropout rates and missing data. A telephone interview, including questions on symptom status and reason for dropout (Psychiatric Symptoms Questionnaire, PSQ), was completed whenever possible for each dropout patient for whom no questionnaire data were available. Several factors potentially affecting the compliance of the treatment were continuously followed. Use of other treatments (psychotherapy or psychopharmacological) during the 5-year follow-up period from enrollment was evaluated by questionnaires and based on information from population registers. Variables measuring the external quality of therapy were also created and continuously followed to gauge how well the treatments given matched the planned characteristics of the treatments. 2.9 Quality control The quality of the integral data collected was controlled with several separate designs. First, the consistency of the interviewer s ratings was evaluated 28

by repeated control ratings of 39 selected interviews recorded on videotape during the entire follow-up. Based on these ratings, both the agreement between interviewers and long-term stability of the ratings were evaluated. Second, the validity of the diagnosis, based on the semi-structured interview method used, was evaluated by comparison with diagnoses based on Structured Clinical Interview for DSM-IV axis I and axis II disorders (SCID) (First et al. 1995 and 1997) in a sample of 27 selected patients. Third, the patients had to wait a median time of 52 days (range 0 526 days) for start of treatment after admission to the study. Possible changes in symptoms during that period were determined with SCL-90 (Derogatis et al. 1973), which was assessed at the time of admission and repeated at baseline. Fourth, the reliability and validity of the Psychiatric Symptoms Questionnaire were assessed, by determining the agreement between PSQ and HDRS and HARS scales. Finally, the internal consistency was estimated for all scales used. Reliability of the Rorschach test was estimated based on 20 protocols using Comprehensive System guidelines (Exner 1993). All laboratory determinations were made under standard quality control. 2.10 Statistical analysis We estimated that 100 patients in the short-term psychodynamic and the solution-focused therapy groups and 130 in the long-term psychodynamic psychotherapy group was required to have a 95% probability of detecting as significant 20% difference during a 3-year follow-up between the 3 groups in the BDI and SCL-90-ANX. The main analyses were based on the intention-to-treat sample. Since the outcome data available for patients who withdrew from the study were scarce, at-treatment analyses were also performed. The data contained repeated measurements of the response variables. In the case of informative dropouts the values of the outcome variables were completed using multiple imputation (Rubin 1987). Information on atypical therapy, caused by incomplete therapy or use of auxiliary therapies during follow-up, was included as covariates in the model. Alternative Bayes models were also used for assessing the ignorability of the compliance (Hogan and Daniels 2002) and for utilizing surrogate information observed on possibly informative missing data values (Pepe 1992). In the case of continuous response variables the main statistical analyses were based on linear mixed models (Verbeke and Molenberghs 1997), and in 29

the case of binary responses logistic regression models and generalized estimating equations estimation were used (Liang and Zeger 1986). Several model-adjusted statistics were calculated for different design points (Lee 1981). For continuous responses, absolute means and differences and for binary responses prevalences and relative risks were estimated. The delta method was used for calculation of confidence intervals (Migon and Gamerman 1999, p. 138). Statistical significance was tested with the Wald test. Two primary models were used in the main analyses of the one-year followup of the short-term psychodynamic psychotherapy and the solution-focused therapy. The basic model included the main effects of time, treatment group, and diagnosis, difference between theoretical and realized starting times of treatment, waiting time from randomization to initiation of treatment, and of first-order interactions of time and treatment group. A complete model further included the potential confounding factors of age, sex, marital status, education, age at onset of first psychiatric disorder, separation experiences, and axes I and II diagnoses. Tests for significance of the effect modification of baseline diagnosis on the treatment effect was carried out by including an interaction term between diagnosis, time, and treatment group in the model. Since no notable differences were found between the 2 models, the results presented are based on the basic model. The need for first auxiliary treatment after finishing the therapy given in the project was analyzed using the Cox model (Cox 1972). The cost-effectiveness analysis of the study was performed using incremental cost-effectiveness ratios (ICER) based on the average cost and effectiveness figures of the study treatments compared (McGuire 2001). The ICER calculated as the ratio of differences of average costs and average effects defines the incremental cost per unit of additional outcome between the 2 treatments. Multiple imputation was used for augmenting missing data values (Shao and Sitter 1996). The confidence intervals for the ICERs were estimated using bootstrap methods. Multi-dimensional sensitivity analyses were also performed to handle the uncertainty associated with parameter estimates. The agreement between, and repeatability of, measurements in the quality control data were estimated as intraclass correlation coefficients (Fleiss 1981; Winer 1971). 30