The Effects of Schema Therapy in Outpatient Forensic Mental Health Care: a Single Case Multiple-baseline Study.

Personality disorders (PDs) are relatively common. Having a PD has a major impact on a patient, their environment and society. It is often accompanied by having other mental disorders such as an addiction or mood or anxiety disorders. It lowers quality of life, lowers life expectancy, and involves more costs due to frequent use of health care.

In forensic psychiatry approximately between 42 and 84% of the patients have one or more PD. Individuals with a PD have an increased risk of violence. Moreover, offenders with a PD have a higher recidivism risk than offenders without a PD. Forensic psychiatry aims to reduce recidivism risk by treating mental or psychiatric problems. Consequently, in outpatient forensic mental health settings, PDs are both assessed and treated. A frequently offered treatment for PDs is schematherapy (ST).

Treatment of PDs with Sschematherapy (ST) was demonstrated to be effective in regular mental health care. The presence of PD symptoms decreased, depression and anxiety decreased and quality of life improved. ST is a form of psychotherapy developed by Young that focuses on early maladaptive schemas (EMS) and schema modes (SM). For forensic patients, ST was adjusted by adding specific SM. This adjustment showed promising results in forensic hospitals in terms of decreases in PD symptoms in a study of Bernstein et al. However, the study was limited to closed forensic psychiatric hospitals where patients were admitted mandatorily.

In recent years, there has been a development in the field of personality and PD with more attention for personality functioning (PF) as the core of personality pathology. This is described in Criterion A of the Alternative Model for Personality Disorders (AMPD) in the DSM-5 section III. This dimensional view fits better with clinical practice in which PF is considered a continuum from maladaptive to adaptive PF and where treatment of patients attempts to improve their PF. Several instruments have been developed to measure this PF, including self-report questionnaires such as the Severity Indices of Personality Problems (SIPP-118), the Level of Personality functioning Scale-Brief Form 2.0 (LPFS-BF 2.0), and the General Assessment of Personality Disorder (GAPD).

In forensic outpatient mental health, as far as we know, no specific instrument has been identified to be useful for measuring routine outcome monitoring (ROM) during PD treatment. Questionnaires assessing PF, such as the SIPP-118 or the LPFS-BF 2.0 might be suitable for this, as they can capture changes in PF during therapy. In forensic psychiatry effect of therapy is usually assessed by evaluating the recidivism risk, or likelihood of delinquency, by using a risk assessment instrument (RAI).

When considering the dynamic factors of the Forensic Ambulatory Risk Evaluation (FARE) and the facets of the SIPP-118, there appears to be overlap between several components (besides unique variance being captured). The LPFS-BF 2.0 also shows overlap with concepts used in the FARE. It is therefore possible that these PF instruments could be of significance in monitoring treatment in forensic outpatient mental health, and a measure of PF might be even more sensitive to changes by treatment in case of PDs as this is the 'core' of personality pathology.

The current study will examine the outcome of ST for PDs in forensic outpatient mental health. To our knowledge this has not been studied before. We will examine three primary outcomes. A first outcome whether ST is an effective treatment is measured in terms of changes towards more adaptive EMS and SM (the classic outcome variables for ST). A second outcome is defined in terms of reducing recidivism risk (the classic outcome variable for forensic psychiatry). Thirdly, we will investigate whether the concept of severity of PF as described in Criterion A of the AMPD in the DSM-5 is useful to monitor the effect of ST treatment for these patients.

Because having a PD is known to correlate with experiencing a lesser quality of life and having other psychological problems, both concepts (quality of life and the presence of psychological complaints) are secondary outcome variables for the effect of treatment.

Hypotheses:

1. Treatment of PD with ST in forensic outpatient mental health shows a positive change in EMS and SM as measured by a self-evaluation of the SM vulnerable child (VC), the main coping mode (MCM) and healthy adult (HA), and by the YSQ-3 and the SMI questionnaires. (Primary outcome variable.)
2. Treatment of PD with ST in forensic outpatient mental health shows a positive change in recidivism risk as measured with the RAI FARE. (Primary outcome variable.)
3. Treatment of PD with ST in outpatient forensic mental health shows a positive change in PD as described in the AMPD and as measured with the self-report questionnaires SIPP-118 and LPFS-BF 2.0. (Primary outcome variables.)
4. Treatment of PD with ST in outpatient forensic mental health shows a positive change for quality of life as measured with the World Health Organization Quality of Life Brief Form. (Secondary outcome variable.)
5. Treatment of PD with ST in outpatient forensic mental health shows a positive change for psychological complaints as measured with the Brief Symptom Inventory. (Secondary outcome variable).

In this study a non-concurrent multiple baseline Single Case Experimental Design (SCED) will be done. A SCED is specific for small sample sizes. No control group is needed with this design. Each participant starts at its own date, in this way we create a within-subject design.

For PDs, ST is standard care at The Rooyse Wissel outpatient treatment centers. Waiting time between intake and start of the therapy was (unfortunately) common practice due to capacity problems in certified therapists. This waiting time will be used as the baseline phase in our study. There will be no extra waiting time for patients who participate in our study. During the waiting period no other form of psychotherapy will be administered which is also standard procedure. Other forms of treatment (for instance pharmacotherapy) will be registered. In the waiting period we will start with gathering data. After start of the ST, with a maximum of 2 years, during the treatment phase, data will be gathered. We aim at starting the study in 2022.

A minimum of 8 patients is targeted, as was used in a similar study with this design that also examined schematherapy. Since this is a SCED no sample size calculation is needed.

Information is extracted from the electronic patient file (EPD) and participants are also asked to fill in questionnaires.

Missing values will be noted by reviewing the completed questionnaires and will be checked with the participants as much as possible. Due to the limited number of intended participants, this is a realistic thing to do. If any values are missing, this will be mentioned when describing the results.

Besides the visual inspection that is typical for a SCED, a mixed-effects model with time as a continuous variable will be used to assess differences within each participant and among treatment phases on the primary and secondary outcomes of forensic care. The outcomes will be assessed longitudinally using fixed effects of time, the therapy x time interaction, and other baseline covariates. The random model part will be selected using the Bayesian information criterion (BIC) from either an 'unstructured', an Autoregressive (AR1) or an Autoregressive Moving Average (ARMA11) within-subject covariance structure. The model results will be presented as P values and 95% confidence intervals. Normal quantile plots of residuals, standardized residuals, and random effects will be used for model diagnostics. If the underlying assumptions underlying the mixed-effects model analysis are violated, we will conduct the analysis using a generalized estimation equation.

The effect of time on therapy will be assessed using Cohen's d. These effect sizes will be calculated as the change between the last intervention value and the baseline value to measure the strength of the treatment effect. Analyses will be done with SPSS.

The research data that are traceable (patient number of tRWot and date of birth) and the assigned subject number, are stored at the VUB in a secured Word file that is only accessible by B. van Reijswoud.

For the remaining data, this subject number is always used for identification. These data are put directly into SPSS under the subject number. The questionnaires that are completed on paper by the participants are also stored at the VUB (only) with the subject number.