Abstract
The aim of this study is to evaluate a cognitive-behavioral treatment for children and adolescents with tic disorder including habit reversal training (HRT) in a sample of children and adolescents (n = 27). Multiple outcome measures were used to assess the effects of the treatment on tic symptoms. In addition, impairment/subjective burden ratings and the self-efficacy to control tics were assessed. A within-subject design with two phases (8 weeks diagnostic; 16 sessions treatment) was analyzed using multilevel modeling (MLM). During the treatment phase, significant improvements in tic symptoms, impairment, and self-efficacy to control tics were found on most outcome measures. Treatment effects were found on the clinical rating of tic symptoms (Yale Global Tic Severity Scale [YGTSS]), the self-efficacy to control tics, and the video-observed motor tic frequency by comparing the improvements during treatment with the course of the outcome measures during the preceding diagnostic phase.
Keywords
Introduction
Tic disorders are neuropsychiatric disorders that are distinguished by involuntarily sudden and rapid movements or vocal expressions (American Psychiatric Association, 2013), which are often preceded by a premonitory urge. Regarding the treatment of tic disorders, pharmacological treatment (particularly antipsychotics) is commonly used and is especially indicated in cases of severe symptoms and strong impairment. Although such treatment is effective, it rarely eliminates tics and often has adverse effects (Rössner et al., 2011). Cuenca et al. (2015) found that families frequently described adverse effects of medication (e.g., drowsiness, tiredness, weight gain) and that these were common reasons for stopping medication.
Cognitive-behavioral treatment, especially habit reversal training (HRT) or exposure and response prevention (ERP), is strongly recommended as initial treatment by the European Clinical Guidelines (Verdellen, van de Griendt, Hartmann, & Murphy, 2011) and the Canadian guidelines for evidence-based treatment of tic disorders (Steeves et al., 2012). Similarly, Murphy, Lewin, Storch, and Stock (2013), as well as Döpfner, Rössner, Woitecki, and Rothenberger (2010), recommended behavioral interventions in children and adolescents with chronic tic disorders that cause impairment and are low to moderate in severity, as well as in patients with severe tics but no compliance with or strong adverse effects of medication. Behavioral therapy is especially welcomed by families as it shows few or no adverse effects (Cuenca et al., 2015). Desired outcomes include reducing or stopping tics, gaining a sense of control over tics, and reducing or managing the urge to tic, and reducing negative emotions associated with tics.
HRT (Azrin & Nunn, 1973) aims to enable patients to gain control over their tics and to manage the urge. Its most important components are awareness training for tics and competing response training. The patient learns to initiate a response to the urge, which is incompatible with the tic (e.g., strengthen the antagonist muscles), and, thus, to control the tic.
A recent meta-analysis found a medium effect for HRT on tic symptoms (McGuire et al., 2014). In their review, Cook and Blacher (2007) concluded that HRT is a well-established treatment according to the American Psychological Association’s Division 12 Task Force criteria. However, despite the research in adult therapy (e.g., Wilhelm et al., 2003) or studies examining therapy for both children and adults (e.g., Verdellen, Keijsers, Cath, & Hoogduin, 2004), there are only a small number of studies investigating HRT for children and adolescents with chronic tic disorders (Hwang, Tillberg, & Scahill, 2012). There is only one large randomized controlled trial, the Comprehensive Behavioral Intervention for Tics (CBIT, Piacentini et al., 2010) study for children and adolescents (9 to 17 years), which examined the effects of HRT (8 sessions) compared with education and supportive psychotherapy (8 sessions). On the primary outcomes, independent masked clinical ratings (Clinical Global Impression Improvement Scale [CGI-I], Guy & Bonato, 1970; Yale Global Tic Severity Scale [YGTSS], Leckman et al., 1989), the reduction of tic symptoms was significantly stronger in the HRT group than in the group receiving education and supportive therapy. Medium-sized effects were found. A smaller randomized controlled trial recently found that HRT in group settings was effective in reducing motor tics but not vocal tics (clinical rating, YGTSS; Yates et al., 2016).
To the best of our knowledge, only one pilot study, published by Woitecki and Döpfner (2011), has also assessed tic symptoms based on self-reports of children and adolescents in addition to parent ratings and clinical ratings. The study, which included 16 children and adolescents aged 8 to 18 years, revealed a decrease in tic symptoms during treatment according to parent and clinical ratings, weekly home observations (self- and parent ratings), and observations of an independent rater. However, no such improvement was found on the self-rated tic questionnaire.
The importance of integrating outcome measures based on multiple informants (e.g., parent ratings, self-reports of the patients) is underscored by the finding of only moderate correlations between parent reports and self-reports of tic symptoms (Döpfner & Görtz-Dorten, 2017). Besides parent ratings, clinical ratings, or self-ratings, teacher ratings may also be important to assess symptoms and treatment outcome at school. To our knowledge, teacher ratings have not yet been used in treatment research with tic disorders.
Moreover, behavioral observations may be another important outcome measure, as they may be free of subjective rater bias. However, behavioral observations have predominantly been used in single-case or small-group studies, while such measures are rarely used in larger studies (Verdellen et al., 2004; Woitecki & Döpfner, 2011). Moreover, to verify whether learned HRT strategies are generalized to the situation at home, it is desirable to conduct observations at home. Himle et al. (2006) found that clinic-based observations corresponded well to home-based observations (rated by clinicians), but Verdellen et al. (2004) found only a weak association between clinic-based observation and home-based observation measured by family members. However, only a few of the larger studies employed direct observations at home (e.g., Verdellen et al., 2004; Woitecki & Döpfner, 2011).
The present study investigates the efficacy of the German treatment program for children and adolescents with chronic tic disorders (Therapieprogramm für Kinder und Jugendliche mit Tic-Störungen [THICS]; Woitecki & Döpfner, 2015) and extends previous research by using multiple informants and different assessment levels. Besides clinical ratings, self-, parent, and teacher ratings were used. Moreover, several direct observations were conducted: clinical observation during therapy session, video recordings rated by independent raters, and home observation of tics by parents and patients. As mentioned above, the patients’ and families’ desired outcome, besides reducing or stopping tics, is to gain a sense of control over tics (Cuenca et al., 2015). Therefore, we assessed the self- and parent-rated sense of controllability over the tics. We expected to find an improvement in tic symptoms as rated by parents (primary outcome), patients, teachers, and clinicians, and based on clinical and home observation. We, in addition, expected an improvement in impairment and subjective burden (parent rating, self-rating, clinical rating) and sense of controllability (parent rating, self-rating) during treatment on all outcome measures. Moreover, a stronger improvement during treatment compared with during a preceding diagnostic phase was expected.
Method
Inclusion Criteria
Children and adolescents were eligible to take part in the study if they were aged 8 to 19 years and had been diagnosed with Tourette’s disorder (F95.2) or a chronic motor/vocal tic disorder (F95.1) according to the 10th revision of the International Classification of Diseases and Related Health Problems (ICD-10) criteria (World Health Organization, 1992). Further inclusion criteria were (a) at least moderate severity of tic symptoms, as measured by the YGTSS Total Tic Score >13 for patients with Tourette’s disorder and >9 for children with motor or vocal tics only (cf. Piacentini et al., 2010); (b) IQ > 80; (c) agreement to attend at least 21 outpatient treatment sessions conducted once per week; (d) children receiving tic medication were allowed to participate if the dosage had been stable for at least 1 month without any planned changes during study participation; (e) tic disorders had to be the primary diagnosis. Exclusion criteria were (a) comorbid diagnosis of autism or psychosis, (b) parallel continuous behavioral therapy.
Study Design
The study was approved by the ethics committee of the University Hospital, Cologne, and was registered on Clinical trials.gov (Identifier: NCT02190383). Informed consent was obtained from patients and their parents prior to inclusion in the study. This analysis is part of a larger clinical trial that aims to assess the effects of THICS in a combined within- and between-subject design. While the between-subject design analyzes the effects of THICS in comparison to an active control group (resource activation), the present analysis uses a within-subject control group design and provides information about the effects of the treatment in comparison to a diagnostic and waiting period, in which no intervention was conducted.
Tic symptoms were measured at four assessment points (T0-T3). T0 took place at the start of an 8-week diagnostic phase, which ended with T1. This was followed by a 16-week treatment phase. T2 occurred after the first half of this treatment phase, that is, after 8 (with a range of ±1) weekly patient sessions (50 min). The second half of the treatment phase, also comprising 8 (with a range of ±1) weekly patient sessions (50 min), ended with the final assessment point T3. In addition, two parent sessions were usually carried out during each treatment phase.
Statistical Analysis
The within-subject design was analyzed using multilevel modeling (MLM; Goldstein, 2003; Hox, 2002; Raudenbush & Bryk, 2002; Snijders & Bosker, 1999), and piecewise linear growth models were computed (Raudenbush & Bryk, 2002; Singer & Willett, 2003). Two different growth rates were calculated for two different time periods: The diagnostic phase was the first time period, and changes during this phase (T0-T1) were covered by the growth rate βdiagnostic. The second time period was the treatment phase (T1-T3) and was covered by the growth rate βtreatment. The first main objective of the analysis was to show that the growth rate βtreatment (change during treatment) was significant. The second main objective was to show that the growth rate βtreatment was significantly larger than the growth rate βdiagnostic (change during diagnostic phase) as a test for treatment effects. To test βtreatment against βdiagnostic, contrasts were defined (βincremental). The intercept of the model was assumed to be random, and the growth rates were fixed for reasons of model identification. The effect sizes (ES) were calculated using the growth rate multiplied by the measurement time intervals divided by the initial standard deviations (T0). In multilevel modeling, incomplete cases remain in the analysis (Maas & Snijders, 2003). This strategy was shown to be appropriate if missing data are missing at random (Rubin, 1976). Regarding descriptive statistics, missing values were not imputed.
Outcome Measures
The primary outcome measure was the symptom score derived from the parent-rated Symptom Checklist for Tic Disorders (SCL-TIC-P), which is an integral part of the diagnostic system for mental disorders according to the ICD-10 and Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV; American Psychiatric Association, 1994) for children and adolescents (DISYPS-II, Döpfner, Görtz-Dorten, & Lehmkuhl, 2008). In the first part of the SCL-TIC-P, the presence of different motor and vocal tic symptoms is assessed by 14 items. For each tic, the frequency of occurrence (not at all to constantly—every few minutes, range 0-4) and intensity of present tics (very mild to severe, irritates others, range 1-4) is assessed for the last week. A tic symptom score (range 0-16) was calculated by multiplying the frequency and intensity ratings for each item and then adding the products and dividing by all given tics (14). The internal consistency of the SCL-TIC-P has been found to be satisfactory (α = .70 to α = .79; Döpfner & Görtz-Dorten, 2017).
In addition, 1 item of the SCL-TIC-P assesses the sense of controllability of tic symptoms on a 5-point scale (1 = very low to 5 = very high). Two further items assess the subjective burden (SB) according to the parent’s perspective: one concerning the SB for the child and one concerning the SB for the parent him or herself (1 = very low, hardly disturbs to 5 = extreme).
The teacher-rated and the self-rated Symptom Checklist for Tic Disorders (SCL-TIC-T, SCL-TIC-S) have the same structure as the SCL-TIC-P. The SCL-TIC-S is designed for children aged 11 and above. For the purpose of this study, it was completed by all children. If needed, they were helped by their therapists, which was especially the case in younger children, children with a lower cognitive level, or children with attention problems. The internal consistency of the SCL-TIC-S is not satisfactory (α = .61 to α = .64), which is unsurprising given the heterogeneous symptoms of tic disorders and the fact that each item assesses a single tic or area of tics. The internal consistency of the SCL-TIC-T is satisfactory (α = .74 to α = .76; Döpfner & Görtz-Dorten, 2017).
To assess tic symptoms observed during diagnostic sessions, a clinical tic observation sheet (SCL-TIC-CO) was designed for the purpose of this study, based on the 14 items of the SCL-TIC-P and recording the frequency of observed tics. Ratings ranged from 0 = not observed to 4 = more than 15 times per therapy session. The intensity rating is similar to that of the SCL-TIC-P.
The YGTSS is a clinician-rated, semistructured interview, which starts with a checklist of all tics present during the past week. Motor and vocal tics are rated separately on 5 domains: number, frequency, intensity, complexity, and interference (each score from 0-5). The domains are summed up to a Total Motor Score and a Total Vocal Score (range of each score: 0-25). These two scores are added up to a Total Tic Score (range 0-50). In addition, an Overall Impairment Rating is assessed (none to severe, range: 0-50). The YGTSS has shown satisfactory validity and reliability in several studies (Leckman et al., 1989; Storch et al., 2005; Walkup, Rosenberg, Brown, & Singer, 1992).
Direct behavioral observations in the clinical setting were conducted via videos of the patient at the end of every treatment session while the child was playing with the therapist (5 min). In addition, at every assessment point (T0-T3), the patient was video-recorded while filling out questionnaires. These videos were randomized regarding sessions and were analyzed by two trained and blinded raters. Before counting the tics of each child, to become sensitized to the child’s tics, raters watched about 15 min of video material of the respective child, which was not used for this study. The partial-interval method was used, as it has been shown to correlate highly with event-frequency coding (Himle et al., 2006). As video counts of at least 5 min duration were found to be reliable and stable (Chappell et al., 1994; Himle et al., 2006), 5-min videos were analyzed. An interval length of 10 s was chosen, as it is the most commonly used interval length in studies (Himle et al., 2006); thus, 30 intervals were analyzed. The Programs “CowLog” (Pastell, 2016) and “Time Sampler” (Franke, 2016) were used to count the tics. Each interval was separately scored for motor (range 0-30) and vocal tics (range 0-30). Analyses showed satisfactory interrater reliability (between rICC = .85 and rICC = .95) and validity (Ruch, 2017; Wälde, 2017). The 8-week retest reliability was not satisfactory, which is unsurprising given the waxing and waning of tics. Other studies that used shorter retest periods found better retest reliability for video observations (Barnea et al., 2016; Chappell et al., 1994; Piacentini et al., 2006).
Direct behavioral observations at home were conducted by the parents and the patients (Woitecki & Döpfner, 2015). One to four individually defined tics were counted once per week (patients alone and patients with parents) using an individually defined observation period (range 5-20 min). The frequency of each individual tic was divided by the highest tic frequency measured for this patient and this specific tic during the whole observation period. A tic frequency score was calculated by summing up this tic frequency ratio for all tics and all situations and dividing the score by the number of tics and the number of situations.
As a global measure of treatment response, the CGI-I was used (Guy, 1976; Guy & Bonato, 1970) at the last assessment point (T3). A modified version was employed, as two ratings were made: tic-specific improvement and improvement in other domains (e.g., self-worth, comorbid problems). The ratings were discussed in a case conference comprising the treating therapist, supervisor, and a minimum of one more therapist, and consensus was reached. Categories ranged from very much improved = 1 to very much worse = 7. The CGI-I shows satisfactory validity (Berk et al., 2008).
Furthermore, a modified tic-specific goal attainment scale was used (Woitecki & Döpfner, 2015) to assess the degree of attainment of a predefined therapeutic goal. The goal attainment was indicated on an individually defined 5-level scale ranging from worsened = −1, through unchanged = 0, to significantly improved, currently no longer a problem = +3.
Treatment satisfaction was measured at the end of the therapy (T3) using the Therapy Evaluation Questionnaire (TEQ, Mattejat & Remschmidt, 1998) for patients (11 years or above) and parents. The questionnaire consists of 20 items rated by the patient and 21 items rated by the parent, with each item rated from 0 = poor to 4 = excellent. For both versions, a total score and a success score were calculated. The internal consistencies for the different questionnaire versions are satisfactory (Cronbach’s α ⩾ .80 for most scales) and the retest reliability for the total score is between r =. 68 and r =. 77.
Treatment
During the diagnostic phase, only an intelligence test with the patient (1 session) and an anamnesis with the parents (1 session) took place; no further sessions or interventions were conducted. The treatment phase drew on the therapy program for children and adolescents with chronic tic disorders (THICS; Woitecki & Döpfner, 2015), which is based on the components of HRT according to Azrin and Nunn (1973) plus psychoeducation and further behavioral interventions (e.g., functional interventions). The first treatment phase contained psychoeducation and tic awareness training consisting of (a) Description of the tic reactions, (b) Self-monitoring, (c) Training of tic reaction detection, (d) Training of perception of the premonitory urge and of perception of situational influences. This should sensitize the patient to his or her tics and their interference. In the second treatment phase, the competing response training was conducted. This was first practiced in the therapeutic setting and then transferred to everyday life. In addition, patients practiced at least one relaxation method (breathing techniques or progressive muscle relaxation).
Results
Participant Recruitment
Participants were recruited through a call in a local newspaper and flyers sent to local doctors, psychotherapists, and clinicians. In addition, patients were referred from the outpatient unit of the School of Child and Adolescent Cognitive Behavior Therapy at the University Hospital Cologne. All parents and children gave consent to participate in the study after the procedure had been fully explained. Participants were included between December 2013 and February 2017. The participant flow of the study is shown in Figure 1.
Flow diagram of the study.
A total of 51 patients were eligible to participate in one of two ongoing studies for tic patients. Of these, 27 patients took part in the present study, while the remaining 24 patients took part in another tic study (resource activation). Table 1 shows the demographic and clinical characteristics of the sample. The participants’ age ranged from 8 to 16 years (M = 11;5, SD = 2;3); 17 (62.9%) of them were boys, and 23 (85.2%) met the criteria for Tourette’s disorder. One patient (3.7%) was receiving medication for tics, and two patients (7.4%) were receiving medication for comorbid disorders (attention-deficit/hyperactivity disorder [ADHD].
Baseline Demographic and Clinical Characteristics.
Note: ADHD = Attention-deficit/hyperactivity disorder.
Treatment Effects
Table 2 summarizes the means and standard deviations, and Table 3 shows the results for the growth rates of the diagnostic and treatment phase as well as the incremental effect.
Means and Standard Deviations for All Outcome Measures Assessed at the Assessment Points T0 to T3.
Note. SCL-TIC-P = parent-rated Symptom Checklist for Tic Disorders; SCL-TIC-S = self-rated Symptom Checklist for Tic Disorders; SCL-TIC-T = teacher-rated Symptom Checklist for Tic Disorders; SCL-TIC-CO = clinical tic observation sheet; YGTSS = Yale Global Tic Severity Scale.
Results of the Multilevel Analyses and Effect Sizes for All Outcome Measures (Clinical, Parent, Teacher, Patient).
Note. n = 27. β = estimated growth rate; df = degrees of freedom; ES = effect size; SCL-TIC-P = parent-rated Symptom Checklist for Tic Disorders; SCL-TIC-S = self-rated Symptom Checklist for Tic Disorders; SCL-TIC-T = teacher-rated Symptom Checklist for Tic Disorders; SCL-TIC-CO = clinical tic observation sheet; YGTSS = Yale Global Tic Severity Scale.
During the diagnostic phase (βdiagnostic), the growth rates for the questionnaire data did not differ significantly from zero, except for the SCL-TIC-CO, which showed a significant decrease of symptoms during this phase, with a small effect size (ES = 0.26). During the treatment phase, growth rates (βtreatment) of all tic-specific questionnaires showed significant improvements, with mostly moderate to high effects (0.39-1.21), indicating a reduction of tic symptoms as rated by parents, teachers, clinicians, and patients, a reduction in impairment and SB as rated by parents, patients, and clinicians, and an increase in controllability as rated by parents and patients.
Regarding the incremental effect of the growth rates (βincremental) between the treatment and diagnostic phase, significant effects only emerged for the clinician-rated tic symptoms (YGTSS Total Tic Score: ES = 0.89) and the self-rated sense of controllability of tics (ES = 1.67), indicating greater improvement during the treatment phase compared with the diagnostic phase. No significant effects emerged for the other assessments, although small to medium effect sizes were found in self- and teacher ratings as well as in the YGTSS impairment rating (ES = 0.31 to ES = 0.57).
Based on the inclusion criteria of YGTSS Total Tic Score >13 for patients with Tourette’s disorder and >9 for patients with chronic tic disorders, a tic score below these cut-offs at the final assessment can be interpreted as a clinical success. A total of 12 of the 25 patients (48%) achieved a lower score than the inclusion criteria at the last assessment point (T3).
The home observation of tics could not be evaluated for all patients, as some patients and their families did not complete it due to several reasons (e.g., high tic variability and, therefore, many changes; forgot to fill out the observation sheet; observation led to higher stress). For all patients (n = 25; βtreatment = −0.03, SE = 0.01, t(242) = −9.12, p = .000; ES = 1.24) and their parents (n = 25; βtreatment = −0.03, SE = 0.01, t(240) = −5.48, p = .000; ES = 1.50) who completed at least one observation, a significant reduction of tics during the treatment phase was demonstrated.
The results of the video ratings of the two independent and blinded raters are shown in Table 4 (means and standard deviations) and Table 5 (MLM).
Means and Standard Deviations for Video Ratings (Playing and Questionnaire Situation) Assessed at the Assessment Points T0 to T3.
Note. Total included cases n = 25 (one patient did not give consent to video observations, and videos of another patient were not usable due to technical problems); due to some technical problems, several data points for a few patients are missing (the range of usable videos regarding the single data points was n = 14 to n = 25).
Multilevel Analyses and Effect Sizes: Frequency Video Ratings (Playing Situation: Every Therapy Session; Questionnaire Situation: Every Assessment Point.
Note. Tic frequency: time sampling method, 30 intervals. β = estimated growth rate; df = degrees of freedom; ES = effect size.
For the motor tic frequency, a statistically significant decrease of tic symptoms in the playing situation and in the questionnaire situation was found during the treatment phase. In addition, an incremental effect was found for both situations, indicating a greater decrease during the treatment phase than during the diagnostic phase (ES = 0.91 and ES = 1.04). The frequency of the vocal tics was not significantly reduced during the treatment phase (playing and questionnaire situation), and no incremental effect was found.
Regarding the modified tic-specific goal attainment scale, which was assessed during treatment only, a significant increase during the treatment phase emerged, βtreatment = 0.55, SE = 0.04, t(90) = 13.07, p = .000, indicating that children and adolescents attained their individualized tic-specific goal.
Regarding the clinician-rated tic-specific improvement assessed at the end of the treatment (CGI-I), 10 patients (37.0%) were rated as very much improved, 9 (33.3%) as much improved, and 6 (22.2%) as minimally improved; 2 (7.4%) ratings were missing, as these patients discontinued the intervention. The clinician-rated improvement in other domains was as follows: 3 patients (11.1%) very much improved; 4 (14.8%) much improved; 6 (22.2%) minimally improved; 8 (29.6%) no change; 2 (7.4%) minimally worse; 2 (7.4%) much worse; and 2 (7.4%) missing as described above.
The average overall treatment satisfaction was 3.49 (SD = 0.47) as measured by the TEQ for parents (n = 25) and 3.24 (SD = 0.54) as measured by the patients (11 years or older, n = 16). On average, the parents estimated the success of the treatment as 2.81 (SD = 0.90) and the patients as 3.01 (SD = 0.75). A range between 3.5 and 2.5 can be interpreted as “good.”
Discussion
This study investigated the course of tic symptoms, impairment, and SB during a cognitive-behavioral intervention for children and adolescents with Tourette’s disorder or chronic motor/vocal tic disorder. Moreover, we analyzed the effects of the treatment by comparing improvements during treatment with the course of outcome measures during a preceding diagnostic phase. Multiple outcome measures including parent ratings, teacher ratings, patient ratings, clinical ratings, and behavioral observations in different settings were applied. A significant improvement during the treatment phase was found on the primary outcome (parent-rated Symptom Checklist, SCL-TIC-P) and on all other tic-related questionnaires including tic symptoms (except for vocal tic frequency in video ratings) as well as on impairment, SB, and controllability of tics. Pre-post effect sizes were mostly moderate to large (ES = 0.39 to ES = 1.21). The expected stronger effects of the treatment phase compared with the diagnostic phase were found for the clinician-rated tic symptom severity (YGTSS) and the self-rated controllability of tics, as well as for blinded video ratings for motor tics, with large effect sizes (ES = 0.89 to ES = 1.67). However, despite the moderate and significant reduction of tic symptoms during treatment on the primary outcome (parent-rated SCL-TIC-P), which on a descriptive level was somewhat stronger than the reduction of symptoms during the diagnostic phase, statistically significant treatment effects did not emerge (ES = 0.25). Nevertheless, treatment effects were found on the clinician-rated YGTSS, which is comparable to the results found in the CBIT study (Piacentini et al., 2010).
The significant decrease of symptoms in the SCL-TIC-CO during the diagnostic phase may be explained by an exaggeration of tics during the first diagnostic session due to excitement or stress about the new situation (e.g., meeting their therapist for the first time and talking about their tics to a stranger), which faded throughout the second diagnostic session. Multiple studies have found that tics increase in emotional and stressful situations (see review in Conelea & Woods, 2008). Another possible explanation is that many patients begin their therapy when symptoms are severe. In terms of the typical waxing and waning of tics, patients might have shown a relatively high level at the beginning, while after 8 weeks, their tics had decreased. This would also explain the results of other ratings (especially parent ratings of tics, SB, and sense of controllability), where symptoms improved with up to medium-sized effects during the diagnostic phase. Furthermore, some unspecific effects, arising, for instance, from positive expectations and feeling understood by a specialized therapist (e.g., Frank & Frank, 1991; Grawe & Grawe-Gerber, 1999), might explain the decrease during the diagnostic phase. Similar unspecific improvements during the diagnostic or waiting phase were found in several other studies (McGuire et al., 2015; Woitecki & Döpfner, 2011). These aforementioned unspecific improvements during the diagnostic phase minimize the change in significant incremental effects. However, the effect sizes of the treatment phase are superior to those of the diagnostic phase. Thus, although not significant, small to medium incremental effects were found for the self- and teacher-rated tic symptoms (SCL-TIC-S/T) as well as the YGTSS Impairment rating (ES = 0.31 to ES = 0.57).
At the end of the treatment, almost half of the patients had a tic score below the cut-off for inclusion, indicating a clinical success. The total tic improvement of about 7 points on the YGTSS Total Tic Score during the treatment phase was slightly below the improvement found in the CBIT study (Piacentini et al., 2010). In terms of percentage improvement, the present study found a reduction of 32% on the YGTSS during treatment, which is comparable with the reduction of 31% found in the CBIT study. This improvement is slightly below the improvement rates found for pharmacotherapy with ziprasidone (35%, Sallee et al., 2000) and risperidone (36%, Scahill, Leckman, Schultz, Katsovich, & Peterson, 2003) and substantially lower than in recent studies for risperidone and aripiprazole, which lay at 50% or more (e.g., Ghanizadeh & Haghighi, 2014; Sallee et al., 2017).
The present study also found an improvement of impairment measured by the YGTSS. This improvement lay at 9.5 points (46%) during treatment, which is slightly below that found in the CBIT study. However, the treatment in the present study was longer than that in the CBIT study (more than 13 hr in the present study vs. 9 hr of treatment in the CBIT study). We do not know whether a somewhat shorter treatment duration would have had similar effects.
The significant reductions with small to medium effects in teacher-rated (ES = 0.39) and patient-rated (ES = 0.52) tic symptoms during treatment indicate that an improvement in symptom severity was also confirmed by the patients themselves and observed by their teachers in the school setting. Despite the stronger improvement during the treatment compared with the diagnostic phase on a descriptive level, this effect was not statistically significant. These results extend the results found in a smaller study by our own research group (Woitecki & Döpfner, 2011). While the improvement in parent-rated tics is similar in both studies, the present study showed larger improvement rates in self-rated tics compared with our previous study.
Significant treatment effects on behavioral observations of tic symptoms emerged for blinded video ratings of motor tics observed in the clinical setting. Therapy effects are, therefore, unlikely to be merely biased by effort justification of parents, patients, and clinicians. An improvement of about 30% was shown for motor tics and between 4% and 43% for vocal tics. Verdellen et al. (2004) found a similar improvement of 32% in children and adults when counting all tics with a mechanical counter.
To our knowledge, apart from the pilot study by our own research group (Woitecki & Döpfner, 2011), this is the only study to find effects of a behavioral treatment on patients’ sense of controllability, which may be a very important treatment goal (Cuenca et al., 2015). Patients play an active role in the HRT and learn that they may gain control over their tics and, therefore, their bodies. Even if some patients are unable to control their tics in every situation, most of them may be able to do so in situations that are important to them. The clinical impression is that many of the patients felt empowered due to the intervention.
The fear of some patients and parents that the tics might worsen due to HRT was not confirmed in the present study, as all children were rated as improved according to the tic-specific CGI-I scale. As the data of the noncompleters of the CGI-I are missing, one could argue that the tics of these patients could have worsened. However, on all SCL-TIC measurements as well as the YGTSS, the two patients who discontinued the therapy showed improved or similar tic scores at T2 compared with T0 and T1, with the exception of the self-rated SCL-TIC, where one of the patients reported a tic increase. This increase might be explained by a higher self-awareness of the tics due to the awareness training rather than by a real increase (cf. Woitecki & Döpfner, 2011). Therefore, the assumption that the data are missing completely at random is supported.
Limitations
One of the main limitations of this study is that it is not a randomized controlled trial. Within-subject analyses have some advantages, for example, error variance is reduced, and fewer participants are needed because patients serve as their own control group (Gliner, Morgan, & Harmon, 2002). Nevertheless, they also entail some disadvantages. When comparing the treatment phase with the diagnostic phase, it was expected that the course found during the diagnostic phase would continue in a linear fashion. However, especially during the first weeks after beginning a treatment, many studies find unspecific treatment effects. It is, therefore, not realistic to assume that this unspecific treatment effect would continue for another 16 weeks. Thus, computing the incremental effect is a very conservative approach, which probably underestimates the treatment effects. Moreover, the small sample size (n = 27), with its limited power, may make it difficult to reliably detect treatment effects or differences between the treatment and diagnostic phases.
A further limitation is that some of the principal investigators were also authors of the evaluated treatment program. Replication by other investigators would rule out the possibility of researcher allegiance. Furthermore, as follow-up data are lacking, the stability and possible long-term effects are unknown. Nevertheless, the long-term effect of this treatment program was assessed for a smaller sample and was found to be stable after 3 months (Woitecki & Döpfner, 2011) and 3 years (Viefhaus, 2012). In addition, this study was conducted in a specialized tic outpatient clinic by well-trained and supervised therapists using a detailed study plan. Further studies are required to determine whether the results can be generalized to routine care conditions.
As one child underwent a temporary change in tic medication (less risperidone for 2 months) and another child received a reduced dosage of methylphenidate during the treatment phase, it could be argued that these changes may have influenced the treatment outcome. Therefore, the data were reanalyzed without all medicated children (n = 24); no changes in the results emerged.
One principal limitation is the lack of a blinded clinical rating. However, due to the positive effects found in blinded video ratings (motor tics), it is unlikely that the effect on the YGTSS can only be explained by rater bias.
Despite some limitations of this study, the findings provide further evidence that a cognitive-behavioral intervention including HRT is effective in the treatment of patients with tic and Tourette’s disorder, in terms of reducing tics, decreasing impairment and SB, and increasing controllability of tics.
Footnotes
Declaration of Conflicting Interests
The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Manfred Döpfner, Anja Görtz-Dorten and Katrin Woitecki are authors of the treatment manual evaluated and of books about tic disorders or questionnaires used in this study, for which they receive royalties from Hogrefe. Other authors have no potential conflicts of interest.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.