Study design
The current study used an open pre-post intervention design that can be used to determine the effectiveness of a single intervention. The strength of this design is the suggestion that the outcome is determined by the intervention, even though pre-post studies do not have control over other elements affecting treatment outcome (16). Patients were recruited from January 2012 through to December 2017 at the NPI, a specialised service for PD treatment that is part of the Arkin mental health institute in Amsterdam. The study was granted an exemption from the provisions of the Medical Research Involving Human Subjects Act (Wmo) by the Medical Ethics Review Committee of VU University Medical Center in Amsterdam and approved by the ethics board of the mental health institute ARKIN in Amsterdam. All patients in the study gave informed consent.
Participants
The NPI treatment programme consists of three treatment pathways: a pathway focusing exclusively on symptoms and coping ( < 1 year treatment), a short-term pathway oriented towards personality change (< 1 year treatment) and a pathway oriented towards structural personality change (> 1 year treatment).
After intake, patients are referred to one of these pathways. Subsequently, after the refinement of the diagnosis and a shared decion-making process, one of the modalities of the treatment pathway is selected.
In addition to the group therapy studied here (SCBT-g), the short-term pathway oriented towards personality change consisted of two individual modalities: Affect Phobia Therapy and Short-term Psychodynamic Supportive Psychotherapy (SPSP).
The inclusion criteria for SCBT-g were: age 18 to 65 years and fulfilment of the DSM-IV criteria for at least one PD. The diagnosis was made in clinical interviews. The exclusion criteria were: severe suicidality, antisocial personality disorder, severe somatic problems/illness, acute and disruptive psychosocial problems such as homelessness, no income or high debts and inability to participate in a group due to communication problems (stuttering, deafness or language barrier).
During the study period (January 2012 to December 2017), approximately 1100 patients were referred to the short-term treatment pathway oriented towards personality change. Of these patients, 225 (20.5%) were selected for the SCBT-g modality on the basis of the inclusion criteria listed above and after a shared decision-making process that could also involve practical considerations such as the availability of groups or the times at which the patient was available to attend therapy etc.
Intervention
The SCBT-g is a highly structured group therapy format based on the protocol by Broersen & van Vreeswijk (15). It consists of twenty weekly sessions of group therapy with 8 or 9 patients. Every session lasts two hours, including a short break. The programme comprised two phases: the conceptualisation phase and the schema-change phase. In the conceptualisation phase, the patients identified their three main schemas by discussing the results from the Young Schema Questionnaire, through psycho-education about the schema model and by discussing the origins of the patients’ schemas. The schema-change phase consists of interventions focused on challenging and changing the maladaptive schemas and schema behaviour into more adaptive schema behaviour patterns with cognitive modification techniques, behaviour experiments and experiential interventions. Before the start of the group therapy, the patients were invited to attend two individual introduction sessions at which the SCBT-g was explained and a final eligibility check took place. Evaluation sessions were individual and took place at mid-treatment, treatment termination and three months after the end of therapy.
During the study period, 26 therapists were assigned to pairs who worked with a total of 31 parallel groups.
Each group had one pair of therapists, with at least one therapist being a general mental health psychologist. Thirteen therapists were general mental health psychologists, one was a clinical psychologist, two were psychiatrists, two were psychotherapists, two were resident psychiatrists, one was a resident clinical psychologist and five were social psychiatric nurses.
All therapists completed a 56-hour course in schema therapy and at least 50 hours of group supervision for schema therapy chaired by a schema therapist registered as a supervisor with the Dutch Association of Schema Therapy. In addition, all therapists attended a weekly peer supervision session lasting one hour.
Measurements
Baseline assessments
Personality disorder
Before the intervention in question, patients were assessed in a standard intake procedure (i.e. clinical interview) conducted by government registered psychologists or psychiatrists. The intake procedure included two interviews, the first to make a general evaluation of the patient’s psychopathology, the second to establish a biography for the patient. Insurance requirements meant that only patients with a confirmed DSM personality disorder diagnosis could be treated in the NPI.
Comorbid depressive symptoms
The severity of comorbid depressive symptoms was measured with the Symptom Check List depression scale (SCL-90-R). On the basis of the Dutch norms for an outpatient psychiatric population (17) for the Symptom Checklist-90-R, a cut-off score of 48 points was used, above which patients were considered to have depressive symptoms.
Measurement instruments
All measurement instruments for outcome were completed by patients at baseline, after 10 weeks, at treatment termination (20 weeks) and at three-month follow-up. The data were collected and ordered by trained research assistants (master-level graduate students in clinical psychology).
The following measurement instruments were used:
The Symptom Checklist 90-Revised (SCL-90-R)(18) Dutch translation(17), a self-report instrument consist of 90 items covering different symptom scales rated from ‘1, not at all’ to ‘5, could not be worse’. The scales are: anxiety, phobic anxiety, depression, somatisation, insufficiency, interpersonal sensitivity, hostility, sleep problems, and a Global Severity Index (GSI) scale. This last scale is the mean for all items. The instrument is well validated and internal consistency is high (Cronbach α=.82-.97). Test-retest reliability is good (19).
The Young Schema Questionnaire (YSQ) (20) Dutch version(21) is a 205-item self-report questionnaire that is scored on a six-point Likert scale. It is used to measure 16 maladaptive schemas (core beliefs) as defined by Young et al, (22). These sixteen schemas are grouped in five schema domains. Schema domain 1= disconnection and rejection (schemas:, abandonment/instability, mistrust/abuse, emotional deprivation, social isolation and social undesirability), Schema domain 2 = impaired autonomy (dependency/incompetence, undeveloped self/enmeshment, defectiveness/shame, and failure to achieve), schema domain 3 = impaired limits (entitlement and insufficient self-control/discipline), schema domain 4 = other directedness (subjugation and self-sacrifice), schema domain 5 = over-vigilance and inhibition (emotional inhibition, unrelenting standards and vulnerability to harm/illness). Research has shown that, in the Dutch version of the YSQ, internal consistency is adequate to high in all schema scales (Cronbach α = 0.73-0.93) (23) and that reliability is good (squared multiple correlation, R² = 0.75)(23).
The ‘Utrecht’ Coping List (UCL)(24) is a self-report questionnaire that aims to measure cognitive and behavioural coping patterns to determine which characteristic coping style is used when confronting problems or complex situations. The UCL covers 47 items. The following seven scales were extracted by factor analysis from 44 scaled items: active coping (7 items), palliative reaction pattern (8 items), avoidance (8 items), seeking social support (6 items), passive reaction pattern (7 items), expression of emotions (3 items) and reassuring thought (5 items). Each of the items is rated on a four-point scale from ‘Doesn’t apply not to me’; ‘Applies seldom to me’, ‘Applies often to me’ to ‘Applies to me’. The UCL has good psychometric properties. Internal consistency for the seven scales (Cronbach α = 0.43-0.88)(24) and reliability (r = 0.45-0.88) (24) is good.
We measured outcomes at three levels: general symptom severity (General Severity Index scale (GSI) of the SCL-90-R); severity of maladaptive schemas (Young Schema Questionnaire) and coping styles ( ‘Utrecht’ Coping List for measuring coping mechanisms).
Secondly, we determined treatment success with the two-step approach of Jacobson&Truax (25) based on pre- to post- and follow-up treatment changes on the SCL-90 Global Severity Index (GSI). The Reliable Clinical Index (25) was calculated first, followed by the symptom remission rate.