Lack of treatment compliance and premature treatment termination in youth forensic psychiatry
August 2013
R.E.S.R. van Veldhuijsen 10409270
Forensische Orthopedagogiek, Graduate School of Child Development and Education University of Amsterdam
Supervisor UvA: Drs. E. Kornelis
GGzE Centre for Child and Adolescent Psychiatry Supervisor GGzE: Dr. I.L. Bongers
2 Contents Acknowledgements ... 3 Abstract ... 4 Introduction ... 5 Method ... 9 Participants ... 9 Procedure ... 10 Measures ... 11
Strategy for analyses ... 12
Results ... 12
Descriptives... 12
Correlations ... 14
Logistic regression analyses ... 15
Discussion ... 16
References ... 20
3
Acknowledgements
I would like to give thanks to Drs. E. Kornelis and Professor G. J. Stams, my thesis supervisors from the University of Amsterdam, for their patient guidance and useful critiques on this research work and presentation. I would like to express my deep gratitude to Dr. I.L. Bongers and Drs. E.A.W. de Ruijter, my thesis supervisors from GGzE Centre for Child and Adolescent Psychiatry, for their patient guidance and constructive suggestions during the process of writing this research work. I am grateful for the opportunity of enabling my research and writing my Master’s thesis within the Research Group Forensic Mental Health Care of Professor Ch. van Nieuwenhuizen. I would also like to give thanks to all staff who helped and guided me while writing this research work. I offer my regards to all who supported me in any way writing my thesis. Finally, I wish to thank my parents for supporting me throughout my studies.
4
Abstract
This study investigated whether the lack of treatment compliance within youth forensic psychiatric patients is associated with the premature termination or completion of their treatment. Psychopathology was expected to influence this association. Participants were adolescent forensic psychiatric patients (N = 97) admitted at Youth Forensic Psychiatric Hospital ‘de Catamaran’, Institute for Mental Health Care Eindhoven, the Netherlands. Using multiple logistic regression analyses relations between treatment compliance, psychopathology and treatment termination were investigated. In contrast to expectations treatment compliance is not associated with psychopathology, and psychopathology is not related to treatment termination. Treatment compliance is a predictor for treatment termination, though this relation is marginally significant only for treatment compliance assessed closer to the end of treatment. Future studies should investigate treatment compliance in a more dynamic way, such as treatment engagement as defined by Drieschner, Lammers & Van der Staak (2004). Since engagement in treatment appears to be a dynamic process, monitoring of engagement will contribute to clarify engagement profiles in order to predict treatment termination more accurate. An implication for clinical practice arising from this study is to measure patients’ improvement in functioning on more specific areas of individual functioning that are related to motivation and engagement, instead of a global assessment of their functioning.
5
Introduction
Multiple theories give explanations for complex processes and factors involved in motivation and engagement in treatment among youth in forensic psychiatric setting (Drieschner, et al., 2004; Prochaska & DiClemente, 1982) These youngsters generally have a lack of motivation in treatment which is associated with premature treatment termination (Drieschner & Verschuur, 2010). For that reason, it is important to improve motivation and engagement within youth forensic psychiatric patients. This study investigated whether treatment engagement1 in youth forensic psychiatry is associated with the premature termination or completion of their treatment.
Treatment success in youth forensic psychiatry depends on many factors and is difficult to predict: motivation and engagement in treatment are found to be central factors (Abram, Teplin, McClelland, & Dulcan, 2003; Drieschner, et al., 2004; Smith, Duffee, Steinke, Huang, & Larkin, 2008). Youth forensic psychiatric patients have complex psychiatric and behavioural problems which make participation in society difficult (Harder, Knorth, & Kalverboer, 2012). These youngsters often have been in contact with child care and previous treatments might have been insufficient. Admission to forensic psychiatric setting is necessary because of involvement with the criminal justice system or the risk they pose to themselves or others. Treatment in forensic psychiatry aims to decrease recidivism and to improve youth’s behaviour, competency and mental health so the patient can participate in society as self-reliant as possible (Harder, et al., 2012). Treatment success contains many aspects and is related to various domains in one's life. Literature defines treatment success as follows: measurements of treatment success used in recent studies are retention in treatment, achievement of goals and aspects of treatment outcome, such as decreased recidivism, improved behaviour, mental health and competency (Harder, et al., 2012; Roedelof, Bongers, & van Nieuwenhuizen, 2013). Risk assessment is a predictor of treatment success: the dynamic individual items of the Structural Assessment of Violence Risk in Youth (SAVRY) are found to be predictors of violent recidivism (Lodewijks, Doreleijer, & De Ruiter, 2008; Hilterman, Nicholls, & Van Nieuwenhuizen, 2013). Recent research shows that environmental factors such as community environment and treatment setting are associated with treatment progress and retention in treatment (Bartak, et al., 2011; Luborsky, et al., 2002). Moreover, perceived positive treatment environment is related to increased social skills and decreased criminal cognitions (Van der Helm, Stams, & van der
1 This study used a more narrow definition for treatment engagement and operationalized treatment engagement as lack
of treatment compliance, using the dynamic item of the SAVRY, an instrument for risk assessment (Borum, 2006). Therefore, the concept used in this study is defined as ‘lack of treatment compliance’, concepts from literature are defined as ‘treatment engagement’.
6 Laan, 2011). Further, specific factors, such as the patient therapist relationship, are associated with treatment outcome and retention in treatment (Karver, Handelsman, Fields, & Bickman, 2006). In addition, experiencing positive relationships with youth care workers leads to positive outcomes for youth in residential settings (Harder, et al., 2012). Finally, patient factors, such as outcome expectancy, have an influence on treatment progress (Harder et al., 2012). External, treatment and patient factors can influence the progress of treatment. However, in order to achieve treatment success the patient needs to make an effort to change (Derisley & Reynolds, 2000; Drieschner, et al., 2004; Joe, Simpson, Greener, & Rowan-Szal, 1999; Roedelof, et al., 2013)
The Trans Theoretical Model and the Integral Model are two theories that give explanations for the complex processes associated with motivation for treatment and patients’ engagement in treatment. According to the Trans Theoretical Model presented by Prochaska and DiClemente (1982) change of behaviour occurs through four stages. Motivation for change unfolds in the contemplation and determination stages, in which cognitive and emotional processes prepare the patient to act, after this the patient commits to change and acts upon this in the behavioural processes in the action and maintenance stages. Engagement to treatment is seen as an expression of commitment to change, and is what patients show in their actual behaviour during treatment (Prochaska & DiClemente, 1982). However, Drieschner et al. (2004) distinguished treatment motivation and treatment engagement, rather than stages in the process of change, by presenting an Integral Model of treatment motivation and related concepts. In this model patients’ treatment motivation is influenced by internal determinants, such as distress, problem recognition and outcome expectancy. The internal determinants in turn are influenced by external factors and patients’ perception of these internal determinants. The treatment motivation is assumed to influence the treatment engagement. In contrast to the Trans Theoretical Model, where motivation unfolds in stages and engagement is a behavioural expression of motivation, the Integral Model clearly distinguished both concepts: treatment motivation is defined as the willingness to change and treatment engagement is the actual behaviour the patient is showing in order to change (Drieschner, et al., 2004).
The use of this latter theory might improve the comprehension of predictors of treatment outcome within youth in forensic psychiatry by examining treatment engagement as predictor of treatment outcome. Patients with high engagement profiles are found to benefit most from treatment, and in addition, treatment outcome can be predicted based on treatment engagement at the start of treatment (Roedelof, et al., 2013). Previous research found patients who were engaged in treatment to have positive treatment outcomes (Karver, et al., 2006; Raftery, Steinke, & Nickerson, 2010; Roedelof, et al., 2013; Smith, et al., 2008). Ferrin et al. (2012) found attitudes towards treatment
7 engagement to be related to perceived benefits of treatment and to be predictors of actual treatment engagement within youth diagnosed with attention deficit hyperactivity disorder. More autonomous motivational attitude is found to predict treatment engagement (Klag, Creed, & O'Callaghan, 2010), suggesting intrinsic motivation to be a stronger predictor of treatment engagement than extrinsic motivation. Drieschner and Verschuur (2010) found treatment engagement to predict reduction of recidivism and even stronger to predict premature treatment termination (e.g. drop-out, step-out), in which treatment engagement closer to ending of treatment predicted premature termination of treatment stronger than engagement in the beginning.
This study defined treatment engagement in the perspective of risk assessment as lack of treatment compliance, causing this definition to be more narrow than treatment engagement as defined in literature (Drieschner, et al., 2004, Roedelof et al., 2012). Operationalized as measurement for risk assessment, the lack of compliance with treatment is also seen as a risk factor for violent behaviour and recidivism (Borum, 2006). Generally youth in forensic psychiatry are more often placed under coercion compared to youth in other treatment settings. Experiencing coercion for placement and treatment might increase perception of legal pressure and costs of treatment, and decreases perception of suitability or necessity of treatment. Consequently, youth in forensic settings are often extrinsically motivated, and therefore more reluctant to treatment and fail to comply with treatment (Abram, et al., 2003; Roedelof, et al., 2013). For that reason, a better understanding of treatment engagement and related factors could be of benefit to clinical practice and scientific knowledge.
Extrinsic motivation might cause failing to comply with treatment for youth in forensic psychiatric setting, although their psychopathology might also be a factor influencing their engagement in treatment (Drieschner et al., 2010). Psychopathology is found to influence the internal determinants related to treatment engagement: psychopathology impedes problem recognition, increases distress and reduces outcome expectancy (Drieschner, et al., 2004), which makes showing engagement in treatment more difficult for patients with severe psychopathology. Youth in forensic psychiatry are more likely than other adolescents to have one or more psychiatric disorders, as well as they are more likely to have multiple problems and risk-factors on psychological, behavioural and social areas affecting functioning and development (Abram, et al., 2003; Colins, et al., 2011; Hussey, Drinkard, Falletta, & Flannery, 2008; Jochems, et al., 2012). Suffering from (severe) psychopathology might influence the capabilities to show the necessary behaviour in treatment for these youngsters, in other words their treatment compliance. Youth with severe psychopathology are found to be more reluctant to treatment and to have more complex needs (Abram, et al., 2003). Psychiatric disorders which are commonly found in youth psychiatric patients are pervasive
8 developmental disorders (PDD), attention deficit (hyperactivity) disorders (ADHD), conduct disorders (CD), psychotic disorders and substance use disorder (SUD) (De Ruijter, Vermunt, & Van Nieuwenhuizen, 2012). Having a mental disorder in combination with substance abuse is common among youth in forensic psychiatry which is associated with less compliance with treatment (Abram, et al., 2003; Hussey, et al., 2008). The craving for substances caused by the substance dependency might be more important than their motivation in treatment, which will influence the treatment compliance. (Abram et al., 2003; Klag et al., 2010). In addition, patients with conduct disorder are found to be less complied with treatment (Abram et al., 2003; Colins et al., 2011). Behavioural aspects of the conduct disorder such as reluctance, disobedience and rule-breaking behaviour might make these patients less able to be in compliance with treatment as necessary.
The aim of this study is to investigate whether lack of treatment compliance predicts premature treatment termination and how treatment compliance is influenced by psychopathology. Further, the differences between treatment compliance and severity of psychopathology at the beginning and ending of treatment are investigated. Considering this, the following research questions are formulated:
1. Does treatment compliance predict premature treatment termination?
It was hypothesised that treatment compliance predicts treatment termination: lower treatment compliance is expected to lead to premature treatment termination, whereas higher treatment compliance would lead to treatment completion (Drieschner & Verschuur, 2010).
2. How is treatment compliance influenced by severity of psychopathology?
Severity of psychopathology is expected to negatively influence treatment compliance (Drieschner, et al., 2004; Roedelof, et al., 2013).
3. Are there differences between severity of psychopathology and treatment compliance at the beginning and ending of treatment?
Expected is that there are differences between beginning and ending of treatment for severity of psychopathology and treatment compliance (Drieschner & Verschuur, 2010; Roedelof, et al., 2013). Recent studies found different engagement profiles during the course of treatment (Roedelof, et al., 2013), these might be relevant to treatment compliance as well. No expectations regarding the direction of change of treatment compliance are made.
4. Are severity of psychopathology and treatment compliance at the ending of treatment stronger predictors for the premature termination of treatment compared to severity of psychopathology and treatment compliance at the beginning of treatment?
Severity of psychopathology and treatment compliance levels at ending of treatment are expected to be stronger predictors of treatment outcome (premature termination or treatment
9 completion) than at beginning of treatment (Drieschner & Verschuur, 2010; Roedelof, et al., 2013).
5. Does psychopathology influence the relation between treatment compliance and treatment outcome?
Expectations are that the relation between treatment compliance and treatment outcome differ between psychiatric disorders; patients with SUD (or SUD in combination with another psychiatric disorder) are expected to be less complied with and less likely to complete treatment compared to patients with other psychiatric disorders (Abram, et al., 2003; Stein, et al., 2006). In addition, patients with conduct disorders are expected to be less complied with treatment and less likely to complete treatment compared to patients with other psychiatric disorders (Abram et al., 2003).
Method Participants
Data were gathered of patients admitted at Youth Forensic Psychiatric Hospital ‘de Catamaran’, Institute for Mental Health Care Eindhoven, the Netherlands. Participants were male adolescents who were admitted in the period from January 2009 to December 2012 under juvenile civil law, juvenile criminal law or voluntary. Lack of treatment compliance was measured after the first six months of treatment; the ‘start period’, as well as after the last six months of treatment; the ‘end period’ of treatment. Therefore, participants who were in treatment for less than three months were excluded from the sample, since for them no measurement of treatment compliance had been taken place. Participants who were in treatment for less than a year but more than three months (N = 30) had only one measurement of treatment compliance. Consequently, for analyses of the third and fourth research questions two measurements of treatment compliance were required, these participants were excluded from the sample used for analyses of the third and fourth research questions.
The final sample consisted of 97 participants who were included for analyses of the first and second research questions. The participants’ age at admission ranged from 14 to 23 years (M = 16.8 years, SD = 1.9). The length of stay ranged from 3 to 38 months (M = 15.2 months, SD = 9.0). Almost half of the participants (49.5%) were admitted under criminal law, the other participants were admitted under civil law (43.3%) or placed voluntary (7.2%). The majority of this sample (75.3%) had a registered criminal background (criminal background of 10 participants was missing) and the mean number of convictions was 4.41 (SD = 3.75), ranging from 1 thru 15. Most patients were diagnosed with comorbid disorders on Axis 1 of DSM-IV (78.4%); most prevalent disorders on Axis
10 1 of DSM-IV were PDD (44.3%), CD (49.5%), ADHD (26.8%) and SUD (21.6%) on Axis 1 of DSM-IV. Other disorders prevalent in the sample were Reactive Attachment Disorder (15.5%), Mood disorders (7.2%), Anxiety disorders (6.2%), Schizophrenia and other psychotic disorders (6.2%), Sexual and Gender Identity disorders (3.1%), Sleep disorders (1.0%), Impulse-control disorders (7.2%), Adjustment disorders (1.0%), Personality disorders (5.2%), Mental retardation (6.2%) and other conditions that may be a focus of clinical attention (17.5%).
The sample used for the third and fourth research questions consisted of 67 participants. The participants’ age at admission in this sample ranged from 14 to 23 years (M = 16.8 years, SD = 2.13). The length of stay ranged from 3 to 38 months (M = 18.1 months, SD = 9.46). Almost half of the participants (47.8%) were admitted under civil law, the other participants were admitted under criminal law (43.3%) or placed voluntary (9.0%). The majority of this sample (68.7%) had a registered criminal background (criminal background of 8 participants was missing), of which mean number of convictions was 3.63 (SD = 3.69), ranging from 1 thru 15. Most patients were diagnosed with comorbid disorders on Axis 1 of DSM-IV (79.1%); most prevalent disorders on Axis 1 of DSM-IV were PDD (44.8%), CD (50.7%), ADHD (26.9%) and SUD (19.4%) on Axis 1 of DSM-IV. Other disorders prevalent in the sample were Reactive Attachment Disorder (16.4%), Mood disorders (7.5%), Anxiety disorders (7.5%), Schizophrenia and other psychotic disorders (4.5%), Sexual and Gender Identity disorders (4.5%), Impulse-control disorders (9.0%), Adjustment disorders (1.5%), Personality disorders (4.5%), Mental retardation (6.0%) and other conditions that may be a focus of clinical attention (19.4%).
Procedure
The Youth Forensic Psychiatric Hospital ‘de Catamaran’, Institute for Mental Health Care aims to offer psychological and psychiatric treatment to youth who have severe behavioural problems causing a risk for themselves or others, or who have been involved with the criminal justice system. Data were gathered by retrospective coding from clinical files of the patients. Background variables such as age, criminal background, and court order, as well as the DSM-IV diagnosis (American Psychiatric Association, 2000) were scored using the medical files and when applicable the criminal files. Treatment engagement was defined in a more narrow way as ‘lack of treatment compliance’, and was operationalized by using the dynamic item that measures lack of treatment compliance of the Structural Assessment of Violence Risk in Youth (SAVRY), a risk assessment instrument (Borum, 2006). The SAVRY is scored by researchers and research interns after the first six months of treatment; the ‘start period’, as well as after the last six months of treatment; the ‘end period’ of
11 treatment for those patients who are in treatment for longer than a year. The SAVRY was scored only once for those who are in treatment for longer than three months but less than one year. Treatment termination was scored by means of consensus rating by researchers and research interns using the discharge letter or, if more information was required, using the last treatment plan as well.
Measures
Lack of treatment compliance
Treatment compliance was measured with the dynamic item ‘Lack of treatment compliance’ of the SAVRY, a risk assessment instrument (Borum, 2006). The SAVRY assesses the risk for violent behaviour and recidivism based on the behaviour shown in the six months prior to assessment. Treatment compliance was measured using the SAVRY scores for the first six months of treatment (start period) and the last six months of treatment (end period). The possible scores were 0 (low lack of compliance), 1 (moderate lack of compliance) and 2 (high lack of compliance). Interrater reliability of 80% was achieved between researchers and research interns. Additionally, a qualified psychologist checked all risk assessment scores of the SAVRY.
Psychopathology
Psychiatric comorbidities and severity were measured by the DSM-IV diagnosis of the patients (American Psychiatric Association, 2000). Based on the literature, severity of psychopathology includes several aspects: a diagnose of a severe psychiatric disorder, treatment or illness duration of at least two years and several disabilities (Jochems, et al., 2012). This study used the Global Assessment of Functioning Scale (GAF), Axis V of the DSM-IV diagnosis at admission as measurement for severity of psychopathology at start of treatment, as well the GAF score in the discharge letter as measurement for severity of psychopathology at the end of treatment, diagnosed by the patients’ psychiatrist or psychologist. Based on most common prevalent psychiatric disorders in this sample psychopathology is divided in the categories 'Pervasive Developmental Disorders' (PDD), 'Attention Deficit (Hyperactivity) Disorders' (ADHD), 'Conduct Disorder' (CD), 'Substance Use Disorders' (SUD), and 'Comorbities of psychopathology' (which includes comorbities on these four most common disorders).
Treatment termination
Treatment termination was measured according to the discharge information from the participants or, if more information was necessary, the last treatment plan was used as well. Treatment termination
12 was categorized as ‘completer’ (clinical treatment completed), ‘drop-out’ (premature termination of treatment by patient), ‘step-out’ (premature termination of treatment in agreement with therapist/clinical staff, e.g. due to transfer to another setting), and ‘push-out’ (premature termination of treatment by therapist/clinical staff). Treatment termination was scored by consensus rating between researcher and research interns.
Strategy for analyses
Multiple analyses are conducted to investigate the relation between treatment compliance and treatment termination, and the influence of psychopathology. First, the descriptives for age at admission, court order, criminal background, treatment compliance at beginning of treatment, and treatment termination were presented. Second, logistic regression analyses were performed to examine whether treatment compliance predicted treatment outcome and whether this relation was mediated by severity of psychopathology. Two models were tested using logistic regression analyses: First, to examine whether treatment compliance at the beginning of treatment influenced treatment outcome and whether this relation was mediated by severity of psychopathology at beginning of treatment. Second, to examine these relations for treatment compliance and severity of psychopathology at the end of treatment. In order to test these models, treatment termination was dichotomised, with the categories ‘0’ (‘premature termination of treatment’, previous drop-out, step-out and push-step-out) and ‘1’ (‘completion of treatment’, previous completer). Additionally, treatment compliance was dichotomised, with categories '1' (previous high and moderate lack of compliance) and '0' (previous low lack of compliance) lack of treatment compliance. The Global Assessment of Functioning Scale (GAF), Axis V of the DSM-IV diagnosis, was transformed into a categorical variable with categories ‘0’ (‘30-39’), ‘1’ (‘40-49’), ‘2’ (‘50-59’), and ‘3’ (‘60-70’). Finally, logistic regression analyses were performed to investigate whether there are differences for type of psychopathology regarding the relation between treatment compliance and treatment outcome; these analyses were performed for measurements both in start period and end period of treatment.
Results Descriptives
The descriptives for age at admission, court order and criminal background are presented in Table 1. Most prevalent type of crimes in the criminal background of these youngsters were violent crime (light and medium) and crime against property. The percentages of the model variables treatment
13 compliance, severity of psychopathology and treatment termination are presented in Table 1 as well. For conducting the regression analyses treatment compliance and treatment termination were dichotomised, as described above. The majority of the sample showed high lack of treatment compliance (80.4%), which are moderate and high compliance according to the SAVRY together. Treatment termination was transformed into a dichotomous variable as well, with approximately half of the sample completing the treatment (44.3%).
Table 1
Descriptive Statistics of the Background and Model Variables for the Total Sample (N = 97)
Variable N %
Age at admission – Mean (SD) 97 16.8 (1.9)
Court order Civil law Criminal law Voluntary 42 48 7 43.3 49.5 7.2 Criminal backgrounda None Offense Drug offense
Violent crimes (light) Crime against property Violent crimes (medium)
Crime against property with violence Violent crimes (severe)
Sexual crimes (Attempted) murder/manslaughter Arson Homicide 14 14 5 37 39 40 17 4 15 2 7 2 14.4 14.4 5.2 38.1 40.2 41.2 17.5 4.1 15.5 2.1 7.2 2.1 Psychopathologyb PDD ADHD CD SUD 43 26 48 21 44.3 26.8 49.5 21.6 Lack of treatment compliancec
Low Moderate High 19 48 30 19.6 49.5 30.9 Treatment termination Completer Drop-out Step-out Push-out 43 19 12 23 44.3 19.6 12.4 23.7
Note: a Criminal background was missing for 10 of the 97 participants. Participants may have committed multiple crimes,
b Psychopathology at start period of treatment; percentages of Pervasive Developmental Disorder, Attention Deficit
(Hyperactivity) Disorder, Conduct Disorder, Substance Use Disorder as DSM-IV Axis 1 diagnoses, cLack of treatment compliance at start period of treatment.
14 For the model variable psychopathology the prevalence of comorbidities are presented in Table 2. Most patients were diagnosed with two or more disorders on Axis 1 of the DSM-IV regarding the most common disorders (PDD, ADHD, CD and SUD) in this sample. Among this sample various comorbidities of disorders are prevalent as shown in Table 2.
Table 2
Prevalence of Comorbidity among Juvenile Forensic Psychiatric Patients with Pervasive Developmental Disorder, Attention Deficit (Hyperactivity) Disorder, Conduct Disorder or Substance Use Disorder
Disorder Prevalence of Comorbidity, %
PDD (N = 43) ADHD
Conduct Disorder Substance Use Disorder
23.3 32.6 16.3 ADHD (N = 26) PDD Conduct Disorder Substance Use Disorder
38.5 50.0 26.9 Conduct Disorder (N = 48) PDD ADHD
Substance Use Disorder
29.2 27.1 27.1 Substance Use Disorder (N = 21)
PDD ADHD Conduct Disorder 33.3 33.3 61.9
Note. Most common disorders within this sample are analysed (PDD, ADHD, CD and SUD). Percentages reflected for
each type of psychopathology: participants may have more than two disorders, so may be included in more cells, and therefore percentages do not sum to 100. Total sample (N = 97).
Correlations
Correlations between the model variables were examined (Table 4, Appendix). The correlations were significant for measurements at end period of treatment between all model variables: lack of treatment compliance correlated negatively with severity of psychopathology (r = -0.31, p < 0.05) and also negatively with treatment termination (r = -0.25, p < 0.05), and severity of psychopathology correlated positively with treatment termination (r = 0.24, p < 0.05). For the measurements at start period of treatment no significant correlations were found for the model variables (Table 4, Appendix).
15
Logistic regression analyses
Logistic regression analyses were performed testing two models2. First, the relations between treatment compliance and treatment termination with severity of psychopathology as mediator, at the start of treatment were examined. Second, these relations were examined for treatment compliance and severity of psychopathology both at the end of treatment. The results of the regression analyses are presented in Table 3.
Table 3
Logistic Regression Analyses Predicting Treatment Termination from Treatment Compliance and Severity of Psychopathology among Juvenile Forensic Psychiatric Patients
B SE eB
Start of Step 1. Treatment compliance Severity of psychopathology
treatment GAF 30 – 39 .000 28420 .688 1.000
N = 97 GAF 40 – 49 -19.904 16408 .653 .000
GAF 50 – 59 -19.930 16408 .653 .000
GAF 60 – 70a - - -
Step 2. Treatment compliance Treatment termination .778 .520 2.200 Step 3. Severity of psychopathology Treatment termination -.218 .323 .804 Treatment compliance Treatment termination .777 .521 2.176 End of Step 1. Treatment compliance Severity of psychopathology
treatment GAF 40 – 49 20.392 8380 .814 7.180 E8
N = 67 GAF 50 – 59 .245 .728 1.278
GAF 60 – 70a - - -
Step 2. Treatment compliance Treatment termination 1.332 * .675 3.789 Step 3. Severity of psychopathology Treatment termination .451 .381 1.570 Treatment compliance Treatment termination 1.090 .704 2.975
Note. a Reference category, * p < .05.
Results in Table 3 show there are no significant relations in the first model. Furthermore, the results in Table 3 show there is a marginally significant relation in the second model between treatment compliance at the end of treatment and treatment termination (B = 1.33, p < 0.05). The relation between treatment compliance and treatment termination was no longer significant when severity of psychopathology was considered. This gives the suggestion of mediation by severity of psychopathology, however, no significant indirect effect of severity of psychopathology was found.
2 Additional analyses were conducted to test whether there was a relation between severity of psychopathology and
treatment termination and if this relation was mediated by treatment compliance. In order to test this model stepwise multinominal logistic regression analyses were performed. Results of these analyses showed no significant relations in this model (Table 5, Appendix).
16 Finally, nominal logistic regression analyses were performed to test whether the relation between treatment compliance and treatment termination differed by type of psychopathology. These analyses were performed for treatment compliance and psychopathology at the beginning of treatment as well as at the end of treatment. There were no significant relations found performing these analyses.
Discussion
The aim of this study was to investigate whether treatment compliance predicted treatment termination and how treatment compliance is influenced by psychopathology. It was hypothesised that treatment compliance predicted treatment termination and there was a mediation effect from severity of psychopathology on this relation. Furthermore, expectations were that treatment compliance and severity of psychopathology at the ending of treatment were stronger predictors for treatment termination than treatment compliance and severity of psychopathology at the beginning of treatment. Additionally, it was expected that the relation between treatment compliance and treatment termination differed by type of psychopathology: it was expected that patients with substance disorder as well as patients with conduct disorder were more likely to be less in compliance with treatment, and therefore more likely to premature terminate treatment. The findings showed only a marginal trend in the relation between treatment compliance at the ending of treatment and treatment termination, although, this relation was no longer significant when severity of psychopathology was considered. Bear in mind that this study operationalized treatment engagement as lack of treatment compliance, according to risk assessment.
In contrast to previous findings, this study did not find treatment engagement to be a good predictor of treatment termination (Ball, Carroll, Canning-Ball, & Rounsaville, 2006; Drieschner & Verschuur, 2010; Roedelof, et al., 2013), since only a marginally trend was found in the relation between treatment engagement at ending of treatment and treatment termination. According to theories on treatment motivation and engagement the patient has to make an effort in order to achieve successful changes. Consequently, in order to complete treatment in forensic psychiatric setting the patient needs to be engaged in treatment. Roedelof et al. (2013) recently found that among psychiatric patients profiles can be distinguished based on treatment motivation. These profiles fluctuated during treatment revealing the dynamic characteristic of treatment engagement. Furthermore, patients with high engagement profiles showed highest treatment engagement overall during treatment and benefit most from treatment (Roedelof et al., 2013). This study measured
17 treatment engagement at start period and end period of treatment, and consequently engagement profiles can not be distinguished. Future studies should measure treatment engagement at multiple moments during treatment, since engagement is found to take place in a dynamic process. This dynamic process might influence the course of treatment (Roedelof et al., 2013). Using a more dynamic assessment of treatment engagement will contribute to the prediction of treatment termination. Implications arising for clinical practice are to monitor treatment engagement of patients during treatment, so engagement profiles can be differentiated in order to improve treatment engagement and motivation.
Severity of psychopathology was found not to mediate the relation between treatment engagement and treatment termination. Results also showed there is no relation between severity of psychopathology and treatment termination. These results are in contrast to recent research which found youth with severe psychopathology to be more recalcitrant to treatment, and, as a result, to be less engaged in treatment (Abram et al., 2003; Drieschner et al., 2004). Patients with high treatment engagement are found to improve most during treatment on the Global Assessment Functioning scale of the DSM-IV Axis V (Roedelof et al., 2013). For a better understanding of improvement during treatment a more specific evaluation of functioning is suggested. Internal determinants that influence patients’ motivation and engagement in treatment can be important factors to determine the course of treatment and influence the reason for treatment termination (Drieschner et al., 2010). For example, increase in problem recognition or perception of suitability of treatment increases patients’ motivation and engagement during treatment (Harder et al., 2012). When these internal determinants improve, patients’ motivation might shift from extrinsic motivation to more intrinsic motivation. This is particularly relevant for forensic psychiatric setting, since patients are often placed under coercion, and therefore show extrinsic motivation, whereas patients who are intrinsically motivated are more likely to complete treatment. The Global Assessment Functioning scale of the DSM-IV Axis V measures global levels of functioning, but suggested is to measure functioning at more specific areas of individual functioning that are associated with motivation and engagement.
Contrary to previous research, no differences were found regarding psychopathology in the relation between treatment engagement and treatment termination. Patients with SUD or comorbid disorders with SUD are found to be less able to show engagement in treatment than patients with other psychiatric disorders (Abram et al., 2003). In addition, patients with conduct disorders are, due to their often disobedient behaviour, less engaged in treatment than other psychiatric patients (Abram et al., 2003). However, comorbidities were highly prevalent among patients in this study which might decrease the effect of type of disorder on treatment termination, and therefore is not found to be a good predictor of treatment termination in this study. Suggested is that the understanding of
18 treatment engagement will improve by examining more specific areas of individual functioning. More insight will be gained for influences of psychopathology and specific areas of individual functioning on course of treatment, and insight will grow regarding the particular limitations associated with the disorders.
Some limitations should be noted considering this study. First, the major limitation of this study was that treatment engagement was operationalized as lack of treatment compliance by using an instrument for risk assessment. The purpose of scoring compliance with treatment in the perception of risk assessment differs from the perception of reflecting on a patients’ engagement in treatment which is considered to arise from a patients’ motivation. Lack of treatment compliance is comparable to the lack of the behaviour the patient is needed to show during treatment, whereas treatment engagement seems to be the commitment to treatment the patient is making. Therefore, future studies should assess treatment engagement in the perspective that reflects on the patient’s motivation and engagement in treatment, such as the definitions used by Drieschner et al. (2004) and Roedelof et al. (2013). Second, the study used cross-sectional analyses, and therefore it is not possible to investigate associations between treatment engagement and psychopathology over time. When panel analyses are used, relations over time between treatment engagement and psychopathology can be investigated. This would be even more interesting to examine when multiple measurements are made of both treatment engagement and severity of psychopathology or assessment of more specific individual functioning. Third, this study used a relatively small sample which might weaken the power for the analyses. Larger sample size will improve statistical power, and make small effects from treatment engagement and psychopathology on treatment termination more evident.
In conclusion, this study found a trend in the relation between treatment compliance at the ending of treatment and treatment termination. In contrast to expectations, treatment compliance is not found to be a good predictor of treatment termination. Suggested is that treatment engagement should be monitored during treatment, so engagement profiles can be distinguished which will emphasise the dynamic process of engagement in treatment. Based on engagement profiles motivation can be improved during treatment, increasing treatment completion among forensic psychiatric patients. Treatment compliance, in the perspective of risk assessment, is not sufficient for predicting treatment outcome. A broader assessment of treatment engagement is necessary in order to predict treatment outcome. Suggested is to investigate treatment engagement according to the theory of Drieschner et al. (2004). Severity of psychopathology is found to have no influence on treatment engagement or treatment termination. Suggested is to measure more specific areas of individual functioning instead of the global functioning measured by the Global Assessment
19 Functioning scale of the DSM-IV Axis V diagnosis. Investigating internal determinants associated with motivation and engagement (Drieschner et al., 2010) will expand knowledge and give clinical implications regarding individual functioning, engagement in treatment and course of treatment of patients in forensic psychiatric settings. Overall, results of this study implicate more research is needed regarding the treatment engagement and treatment outcome of youth forensic psychiatric patients: how engagement fluctuates during treatment, and which specific areas of individual functioning related to motivation and engagement should be assessed in order to improve treatment outcome.
20
References
Abram, K. M., Teplin, L. A., McClelland, G. M., & Dulcan, M. K. (2003). Comorbid Psychiatric Disorders in Youth in Juvenile Detention. Archives of general psychiatry, 60(11), 1097-1108. American Psychiatric Association. (2000). Diagnostic and Statistical Manual of Mental Disorders:
DSM-IV-TR. Washington, DC: American Psychiatric Association.
Ball, S. A., Carroll, K. M., Canning-Ball, M., & Rounsaville, B. J. (2006). Reasons for dropout from drug abuse treatment: Symptoms, personality, and motivation. Addictive Behaviors, 31, 320 - 330.
Bartak, A., Andrea, H., Spreeuwenberg, M. D., Ziegler, U. M., Dekker, J., Rossum, B. V., et al. (2011). Effectiveness of outpatient, day hospital, and inpatient psychotherapeutic treatment for patients with cluster B personality disorders. Psychotherapy and psychosomatics, 80(1), 28-38.
Borum, R. (2006). Manual for the structured assessment of violence risk in youth (SAVRY).
Colins, O., Vermeiren, R., Vahl, P., Markus, M., Broekaert, E., & Doreleijers, T. (2011). Psychiatric Disorder in Detained Male Adolescents as Risk Factor for Serious Recidivism. Canadian Journal of Psychiatry, 55(4), 255-263.
De Ruijter, E.A.W., Vermunt, J.K., & Van Nieuwenhuizen, Ch. (2012). Can we distinguish different profiles in youngsters admitted in a youth forensic psychiatric hospital? EFCAP 3rd International Congress of the European Association for Forensic Child and Adolescent Psychiatry, Psychology and other involved Professions. Berlin, 7-9 maart 2012.
Derisley, J., & Reynolds, S. (2000). The transitional stages of change as a predictor of premature termination, attendance and alliance in psychotherapy. British Journal of Clinical Psychology Review, 39, 371-382.
Drieschner, K. H., Lammers, S. M., & van der Staak, C. P. (2004). Treatment motivation: An attempt for clarification of an ambiguous concept. Clinical Psychology Review, 23(8), 1115-1137. Drieschner, K. H., & Verschuur, J. (2010). Treatment engagement as a predictor of premature
treatment termination and treatment outcome in a correctional outpatient sample. . Criminal Behaviour and Mental Health, 20, 86-99.
Ferrin, M., Ruiz-Veguilla, M., Blanc-Betes, M., El Abd, S., Lax-Pericall, T., Sinclair, M., et al. (2012). Evaluation of attitudes towards treatment in adolescents with attention deficit hyperactivity disorder (ADHD). European Child and Adolescent Psychiatry, 21(7), 387-401.
21 Harder, A. T., Knorth, E. J., & Kalverboer, M. E. (2012). Securing the Downside Up: Client and Care Factors Associated with Outcomes of Secure Residential Youth Care. Child & Youth Care Forum, 41 (3), 259-276.
Hilterman, E. L., Nicholls, T. L., & van Nieuwenhuizen, Ch. (2013). Predictive Validity of Risk Assessments in Juvenile Offenders: Comparing the SAVRY, PCL: YV, and YLS/CMI With Unstructured Clinical Assessments. Assessment.
Hussey, D. L., Drinkard, A. M., Falletta, L., & Flannery, D. J. (2008). Understanding clinical complexity in delinquent youth: comorbidities, service utilization, cost, and outcomes. Journal of Pshycoactive Drugs, 40(1), 85-95.
Jochems, E. C., Mulder, C. L., van Dam, A., Duivenvoorden, H. J., Scheffer, S. C. M., van der Spek, W., et al. (2012). Motivation and treatment engagement intervention trial (MotivaTe-IT): the effects of motivation feedback to clinicians on treatment engagement in patients with severe mental illness. BMC Psychiatry, 12(1), 209.
Joe, G. W., Simpson, D. D., Greener, J. M., & Rowan-Szal, G. A. (1999). Integrative modelling of client engagement and outcomes during the first 6 months of methadone treatment. Addictive Behaviors, 24(5), 649-659.
Karver, M. S., Handelsman, J. B., Fields, S., & Bickman, L. (2006). Meta-analysis of therapeutic relationship variables in youth and family therapy: The evidence for different relationship variables in the child and adolescent treatment outcome literature. Clinical Psychology Review, 26(1), 50-65.
Klag, S. M. L., Creed, P., & O'Callaghan, F. (2010). Early Motivation, Well-being, and Treatment Engagement of Chronic Substance Users Undergoing Treatment in a Therapeutic Community Setting. . Substance Use & Misuse, 45(7), 1112-1130.
Lodewijks, H. P., Doreleijers, T. A., & De Ruiter, C. (2008). Savry Risk Assessment in Violent Dutch Adolescents Relation to Sentencing and Recidivism. Criminal justice and behavior,
35(6), 696-709.
Luborsky, L., Rosenthal, R., Diguer, L., Andrusyna, T. P., Berman, J. S., Levitt, J. T., et al. (2002). The dodo bird verdict is alive and well-mostly. Clinical Psychology: Science and Practice, 9(1), 2-12.
Prochaska, J. O., & DiClemente, C. C. (1982). Transtheoretical therapy: toward a more integrative model of change. Psychotherapy: Theory, Research and Practice, 19 (3).
Raftery, J. N., Steinke, C. M., & Nickerson, A. B. (2010). Engagement, Residential Treatment Staff Cognitive and Behavioral Disputations, and Youths' Problem-Solving. Child Youth Care Forum, 39(3), 167-185.
22 Roedelof, A. J. M., Bongers, I. L., & Van Nieuwenhuizen, Ch. (2013). Treatment engagement in adolescents with severe psychiatric problems: a latent class analysis. European Child & Adolescent Psychiatry, 1-10.
Smith, B. D., Duffee, D. E., Steinke, D. M., Huang, Y., & Larkin, H. (2008). Outcomes in residential treatment for youth: The role of early engagement. Children and Youth Services Review 30(12), 1425-1436.
Stein, L. A. R., Colby, S. M., Barnett, N. P., Monti, P. M., Golembeske, C., R., L.-C., et al. (2006). Enhancing Substance Abuse Treatment Engagement in Incarcerated Adolescents. Psychological Services, 3(1), 25-34.
Van der Helm, P., Stams, G. J., & Van der Laan, P. (2011). Measuring group climate in prison. The Prison Journal, 91(2), 158-176.
23
Appendix
Table 4
Correlations between the Model Variables for start period and end period of treatment
Variable 1 2 3
1 Treatment compliance - .041 .156
2 Severity of psychopathology -.308 * - -.074
3 Treatment termination -.251 * .241 * -
Note. Correlations above the diagonal are within start period of treatment (N=97), correlations below the diagonal are
24
Table 5
Logistic Regression Analyses Predicting Treatment Termination from Treatment Compliance and Type of Psychopathology among Juvenile Forensic Psychiatric Patients
B SE eB
Start of treatment Step 1. Psychopathologya Treatment compliance
N = 97 PDD -.182 .949 .833
ADHD .511 1.366 1.667
Conduct disorder 1.153 1.104 3.167
SUD 19.665 .000 3.471 E8
Comorbid disorders .154 .910 1.167
Step 2. Psychopathologya Treatment termination
PDD .095 .831 1.100
ADHD -.693 1.118 .500
Conduct disorder -.486 .838 .615
SUD 1.386 1.323 4.000
Comorbid disorders -.693 .791 .500
Step 3. Treatment compliance Treatment termination .886 .539 2.426
Psychopathologya Treatment termination
PDD .067 .847 1.069
ADHD -.639 1.136 .528
Conduct disorder -.358 .854 .699
SUD 1.603 1.336 4.966
Comorbid disorders -.692 .805 .500
End of treatment Step 1. Psychopathologya Treatment compliance
N = 67 PDD -1.099 1.211 .333
ADHD 18.048 .000 68906466.849
Conduct disorder 18.048 5639.414 68906466.849
SUD 18.048 .000 68906466.849
Comorbid disorders -.639 1.177 .528
Step 2. Psychopathologya Treatment termination
PDD .693 .928 2.000
ADHD .288 1.258 1.333
Conduct disorder -.916 1.008 .400
SUD -.405 1.443 .667
Comorbid disorders -.288 .870 .750
Step 3. Treatment compliance Treatment termination 1.187 .725 3.278
Psychopathologya Treatment termination
PDD .508 .956 1.661
ADHD .458 1.272 1.580
Conduct disorder -.746 1.025 ..474
SUD -.235 1.455 .790
Comorbid disorders -.428 .897 .652
