This study is a multi-centre randomized controlled trial in which treatment drop-out rates of Moroccan and Turkish immigrants with depressive and anxiety disorders, who are treated by therapists trained in cultural competences, are compared to those who receive regular care. Figure shows the design.
Recruitment and measurement protocol.
Recruitment/Setting and locations
Participants are recruited within two outpatient mood disorder clinics in Amsterdam. All new Moroccan and Turkish registries, mostly referred by their general practitioner, are screened on in- and exclusion criteria. If eligible, the new registries are randomized by a research assistant before the intake procedure starts. Then the patients are given an appointment for an intake interview by a therapist from the matching condition. After this intake interview the patients receive study information and the informed consent form from the therapists. Within two weeks after the intake an interviewer contacts a patient by telephone and asks oral consent to participate in the study. If the patient gives permission, an appointment is made for the basic interview in the clinic. If the patient is willing and eligible to participate, a written informed consent is signed.
The patients are randomized by an independent randomization team that uses a computerized random number generator. The patients are randomly assigned for either intake or treatment to therapists who are trained in cultural competence (the intervention group) or to therapists who are not trained in cultural competence (regular care, control group). The research assistant reveals the treatment condition by telephone to the registration staff who plans the intake interview.
Patients (ages 18 to 65) are eligible to participate if:
1. their main problem is a depressive and/or an anxiety disorder for which they are referred to the participating outpatient mood disorder clinics.
2. they are first or second generation Moroccan or Turkish immigrants. The definition for a first generation immigrant is that the patient him- or herself was born in Morocco or Turkey. The definition for the second generation migrant is that at least one of the patient's parents was born in Morocco or Turkey [17
Patients are excluded from the study if their main problem is one of the following disorders: a psychotic disorder, bipolar disorder, organic brain syndrome, substance dependence, or a severe borderline, schizotypical, or antisocial personality disorder.
Intervention versus control-group therapists
In both clinics six therapists are selected for the intervention group and six for the control group of this study. In both groups disciplines of the therapists, the average amount of years of general treatment experience and ethnic background are evenly distributed.
Therapists of the intervention condition have been trained in cultural competence. The training program is based on existing modules that are widely used in the Netherlands and are based on international and national literature [18
]. The aim of the module is to train the intervention therapists' knowledge, awareness and skills in diagnosing and treating Moroccan and Turkish patients with depression and anxiety disorders.
To improve the cultural knowledge the therapists are familiarized with the Moroccan and Turkish patients' background and learn about:
• the specific cultural aspect and elements,
• the impact of migration in first and second generation immigrants,
• the migration history,
• the religious background, (health)habits, beliefs, explanations and expectations,
• traditional health care and health workers.
To improve the cultural awareness therapists are taught to:
• be aware of their own cultural background, attitudes and values,
• recognise prejudices and generalizations regarding the cultural background of the patients.
To improve the intercultural skills the therapists are trained:
• how to give psychoeducation
• how to make use of an interpreter during treatment,
• how to make use of the Cultural Formulation [24
] during intake and treatment,
• in intercultural diagnostics [25
• how to use the intercultural addenda to the treatment guidelines for depression and anxiety disorders [26
The module has an interactive character. It contains exercises in becoming aware of the therapists' own cultural background; a discussion about how to deal with traditional health workers; a role play with an actor in using an interpreter and in using the Cultural Formulation during treatment.
The training takes two days. After the training the intervention therapists join a monthly intercultural peer group to keep knowledge, awareness and skills vivid.
Data are collected at three points in time: within two weeks after the intake (T0), four weeks after the start of the treatment (T1), 16 weeks after the start of the treatment (T2). Table summarizes the measures that are used at each point. The assessments are partly performed face to face by a trained interviewer and partly self report. Most of the interviewers are bilingual. Besides sufficient command of the Dutch language the Turkish, Moroccan-Arabic and Berber languages are requested respectively. The interviewers participate in a three day training. The interview and self report questionnaires consist of pen and paper versions only.
Primary outcome measure
The primary outcome measure is drop-out of treatment. Treatment drop-out is defined as: the patient is in need of more therapy in the therapist's opinion but ignores at least two invitations of the therapist and does not continue the sessions. Data will be gathered by interviewing the therapists and analysing the medical records.
Secondary outcome measures
Secondary outcome measures focus on no-show and the patients' perspective of treatment. No-show rates will be collected from the medical records of the participants. The patients' perspective will be measured using the Helping Alliance Questionnaire [28
] and the trust in mental health care questionnaire (a part of the NIVEL consumer panel questionnaire will be used [29
Several possible determinants of drop-out will be explored:
Demographic factors (age, gender, marital status, education); level of functioning (WHO-Disability Schedule II (WHO-DAS II) [30
]); loss of productivity at work and health care utilization (Trimbos/iMTA questionnaire for costs associated with psychiatric illness (TIC-P) [31
]); perceived need for care (Perceived Need for Care questionnaire (PNCQ) [32
]); the diagnosis of depression and anxiety disorders (Composite International Diagnostic Interview (CIDI) depression and anxiety life time version (WHO version 2.1) ([33
]); severity of depressive symptoms (Inventory of Depressive Symptoms self report version (IDS) [36
]); severity of generalised anxiety and panic symptoms (Beck Anxiety Inventory self-report version (BAI)[37
]); somatic disease (perceived somatic problems on an analogue scale from 1 to 10); medication use (Medication use questionnaire); pain (Chronic graded pain scale [38
]); acculturation (the Lowlands Acculturation Scale (LAS) [39
]); social support (Close Person Inventory [41
]); discrimination (a part of the national survey of Midlife development in the US); locus of control (Pearlin and Schooler mastery scale [42
Translation of the instruments
For this project insufficient command of the Dutch language is not an exclusion criterion. Therefore, we use translated questionnaires if necessary. For respondents who only understand the Turkish language, we use translated and validated Turkish instruments (IDS http://www.ids-qids.org
, BAI [43
], CIDI (WHO)) and translated instruments used in the Amsterdam Health Monitor (AHM) study [44
] (Who-Das II, PNCQ, LAS, discrimination, locus of control, NIVEL trust in care). The other questionnaires were translated at our institute. Bilingual (Dutch and Turkish speaking and writing) mental health professionals translated the Dutch version into Turkish and this Turkish version was translated back into Dutch by other bilingual mental health professionals.
Because the Moroccan language is a collection of dialects no validated and translated versions of the instruments are available. For the respondents who only understand a Moroccan language we use several instruments used in the Amsterdam Health Monitor (AHM) (CIDI, LAS, discrimination, locus of control, PNCQ). These instruments were not completely translated into Moroccan Arabic. The Dutch version is used and only a keyword in each question was translated into Moroccan Arabic. For the other interview instruments (demographic, TIC-P, medication use, somatic disease, pain scale) keywords were translated in our institute by a focus group of bilingual students (Dutch-Moroccan-Arabic). For the self-report a complete translated version into Moroccan Arabic is accomplished. We asked the Dutch public translation centre (Tolk Vertaal centrum Nederland or TVcN) to translate the Dutch instruments into modern standard Moroccan Arabic. A focus group of bilingual students (Dutch-Moroccan-Arabic) translated it back into Dutch and adapted the Moroccan translation where necessary. We knew from the AHM study that most of the Moroccan respondents who did not speak Dutch well enough, speak a mix of Berber, Moroccan Arabic and Dutch. The interviewers are trained to cope with these language problems as much as possible. In all interviews the interviewers will specify in which language the interview has been conducted and how many problems they have encountered (language or verbal expression problems, associative behaviour or reluctancy to answer). Interviewers assess the reliability of the answers given.
No comparable intervention studies, aiming to reduce drop-out rates in immigrants, have been found in literature. In an intervention study by Blom et al [11
], post hoc analyses showed that the drop-out rate was 46% in immigrant patients compared to 24% in native patients. Based on these data we assume that the drop-out can be reduced from 45% to 30%, an improvement of 33%. To demonstrate this difference 75 patients in each condition (beta 0.01 and alpha 0,05) should be included [45
Basic characteristics will be compared between treatment conditions. Dropout will be analysed by (non) parametric longitudinal analysis techniques (eg GLM model) using multivariate statistics. Determinants are modelled along. Analysis will be performed according to the intention-to-treat principle.
Participation in this study is voluntary. Participants are informed that they can cancel their participation at any time without disclosing reasons for their cancellation and without negative consequences for their future care. Participants sign a written informed consent.
Vote of the ethics committee
The design and conduct of the study was approved by the Medical Ethics Committee of the VU University Medical Center, Amsterdam.