Journal of Psychiatric Research 50 (2014) 106e112

Contents lists available at ScienceDirect

Journal of Psychiatric Research journal homepage: www.elsevier.com/locate/psychires

A meta-analysis of behavior therapy for Tourette Syndrome Joseph F. McGuire a, b, *, John Piacentini c, Erin A. Brennan b, Adam B. Lewin a, b, d, Tanya K. Murphy b, d, Brent J. Small e, Eric A. Storch a, b, d a

Department of Psychology, University of South Florida, 4202 E. Fowler Avenue, PCD 4118G, Tampa, FL 33620, USA Department of Pediatrics, University of South Florida Tampa, FL, USA c Semel Institute of Neuroscience and Human Behavior, University of California Los Angeles, Los Angeles, CA, USA d Departments of Psychiatry and Behavioral Neurosciences, University of South Florida Tampa, FL, USA e School of Aging Studies, University of South Florida Tampa, FL, USA b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 31 October 2013 Received in revised form 30 November 2013 Accepted 18 December 2013

Individual randomized controlled trials (RCTs) of habit reversal training and a Comprehensive Behavioral Intervention for Tics (collectively referred to as behavior therapy, BT) have demonstrated efficacy in reducing tic severity for individuals with Tourette Syndrome and Chronic Tic Disorders (collectively referred to as TS), with no examination of treatment moderators. The present meta-analysis synthesized the treatment effect sizes (ES) of BT relative to comparison conditions, and examined moderators of treatment. A comprehensive literature search identified eight RCTs that met inclusion criteria, and produced a total sample of 438 participants. A random effects meta-analysis found a medium to large ES for BT relative to comparison conditions. Participant mean age, average number of therapy sessions, and the percentage of participants with co-occurring attention deficit hyperactivity disorder (ADHD) were found to moderate treatment effects. Participants receiving BT were more likely to exhibit a treatment response compared to control interventions, and identified a number needed to treat (NNT) of three. Sensitivity analyses failed to identify publication bias. Overall, BT trials yield medium to large effects for TS that are comparable to treatment effects identified by meta-analyses of antipsychotic medication RCTs. Larger treatment effects may be observed among BT trials with older participants, more therapeutic contact, and less co-occurring ADHD. Ó 2013 Elsevier Ltd. All rights reserved.

Keywords: Tourette disorder Chronic Tic Disorder Habit reversal training Comprehensive Behavioral Intervention for Tics Treatment outcome

1. Introduction Chronic Tic Disorders and Tourette Syndrome (afterward collectively referred to as TS), are characterized by the presence of sudden motor movements and/or vocalizations (referred to as tics) that persist for more than a year. Tourette Syndrome commonly begins in childhood, increases in severity into early adolescence, and may persist into adulthood (Bloch and Leckman, 2009). Tic symptoms show little difference between youth and adults with TS, with common tics including eye blinking, head jerk movements, mouth movements, and simple vocalizations (McGuire et al., 2013). The estimated prevalence rate for TS in children ranges from three to eight per 1000 (Centers for Disease Control and Prevention, 2009). Individuals with TS frequently experience co-occurring obsessive compulsive disorder (OCD), non-OCD anxiety disorders, attention-deficit/hyperactivity disorder (ADHD), and/or disruptive

* Corresponding author. Department of Psychology, University of South Florida, 4202 E. Fowler Avenue, PCD 4118G, Tampa, FL 33620, USA. Tel.: þ1 727 767 8230; fax: þ1 727 767 7786. E-mail address: [email protected] (J.F. McGuire). 0022-3956/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jpsychires.2013.12.009

behaviors (Freeman et al., 2000). Tics and co-occurring conditions have been associated with functional impairment (Conelea et al., 2011, 2013), and a diminished quality of life (Jalenques et al., 2012; Storch et al., 2007). Traditionally, TS has been managed with psychotropic medications such as antipsychotics and alpha-2 agonists (Scahill et al., 2006). A meta-analysis of five randomized controlled trials (RCTs) of antipsychotic medications identified a significant reduction in tic severity relative to placebo [standard mean difference (SMD) ¼ 0.58], with no significant difference between medications (Weisman et al., 2012). Despite their efficacy, antipsychotics may be associated with side effects that may limit tolerability (Scahill et al., 2006). Alpha-2 agonists present another treatment option, with a meta-analysis of six RCTs indentifying a more modest (SMD ¼ 0.31), yet still significant benefit relative to placebo (Weisman et al., 2012). As a result, these medications are commonly recommended as a first-line treatment by professional organizations due to less severe side effects (Murphy et al., 2013). Behavioral interventions have also demonstrated success in reducing tic severity (Verdellen et al., 2011). Among these interventions, only habit reversal training (HRT) has consistently

J.F. McGuire et al. / Journal of Psychiatric Research 50 (2014) 106e112

demonstrated efficacy in RCTs. The core components of HRT are considered to be awareness training and competing response training (Miltenberger et al., 1998), with adjunctive therapeutic components including self-monitoring, relaxation training, contingency management, motivational procedures, and generalization training (Piacentini and Chang, 2006). Awareness training involves the detection of premonitory urges and/or early tic movements that precede a tic. Competing response training involves the identification of behaviors that are physically incompatible with a targeted tic, which are implemented upon early tic detection (e.g., premonitory urges, early tic movements). Habit reversal training serves as the principle therapeutic component in the Comprehensive Behavioral Intervention for Tics (CBIT) (Woods et al., 2008), which integrates HRT with functional assessment and function-based intervention procedures designed to mitigate influences of daily life that worsen tics. A related behavioral intervention called exposure and response prevention (ERP) involves exposure to sensations/urges that precede tics with response prevention of tics (Verdellen et al., 2004). These two treatments share similarities in terms of procedure (tic prevention) and mechanisms of change (habituation to urges), but are distinguished by the use of competing responses. The efficacy of HRT and CBIT (afterward collectively referred to as behavior therapy, BT) has been supported across several RCTs with two noted exceptions (Azrin and Peterson, 1990; Deckersbach et al., 2006; O’Connor et al., 2001; Piacentini et al., 2002, 2010; Wilhelm et al., 2003; Wilhelm et al., 2012). One exception is a comparison trial of HRT to ERP. Participants in the ERP condition experienced greater reductions in tic severity that trended toward significance (Verdellen et al., 2004). Difference in duration of therapeutic contact hours may account for discrepancies as ERP participants received 24 therapeutic contact hours, whereas HRT participants received only 10. The other exception is a comparison trial of HRT to awareness training (a component of HRT), with results indicating only a minimal benefit of HRT above awareness training (Piacentini et al., 2002). While both of these trials compared BT to another active behavioral intervention, the remaining RCTs utilized waitlist or non-directive therapy as control conditions. When making treatment recommendations, it is important to synthesize empirical evidence to guide clinical decisions (Murad and Montori, 2013). Meta-analyses provide a quantitative synthesis of treatment trials, and can examine moderators of treatment effects across studies. To date, only two meta-analyses have examined the efficacy of behavioral interventions for tics, albeit with noted limitations (Bate et al., 2011; Wile and Pringsheim, 2013). Bate et al., 2011 examined the efficacy of HRT across multiple habit disorders (e.g., TS, trichotillomania, nail biting, thumb sucking, stuttering, teeth grinding). While Bate et al. found HRT to have a large effect for TS (Cohen’s d ¼ 0.78), this examination had several limitations. For instance, Bate and colleagues excluded RCTs that used core BT components alongside adjunctive therapeutic components (O’Connor et al., 2001), and did not include a large trial of BT for TS (Wilhelm et al., 2012) or the comparison trial of HRT to awareness training (Piacentini et al., 2002). Furthermore, Bate and colleagues did not identify the measures used to extract treatment effects. As the psychometric properties of tic severity ratings scales differ from one another (McGuire et al., 2012), preference should be given to clinician-administered scales that have demonstrated reliability and validity. Wile and Pringsheim (2013) examined the efficacy of BT for individuals with TS in two meta-analyses that each included two RCTs. On the clinician-rated Yale Global Tic Severity Scale (YGTSS; Leckman et al., 1989), Wile and Pringsheim (2013) found a difference in YGTSS Total Tic Scores between behavior therapy groups

107

and comparison groups of 3.66 and 10.52 for blinded and unblinded trials, respectively. Their examination of treatment response in two RCTs on the Clinical Global Impression of Improvement (CGI-Improvement; Guy and Bonato, 1970) identified an odds ratio in favor of BT. Despite its noteworthy contributions, this report had limitations that included a small sample size, utilization of a fixed effect model, and did not explore treatment moderators. Given the collective limitations of existent metaanalyses, further investigation is needed to comprehensively quantify treatment effects of BT, and explore moderators of treatment. A systematic approach to quantifying BT treatment effect is important because it facilitates comparison with existing pharmacotherapy meta-analyses for TS, and provides an informative comparison in the absence of a head-to-head RCT. Similarly, an examination of BT treatment moderators may inform clinical recommendations. The present meta-analysis examined RCTs of BT to determine its efficacy in reducing tic severity and identify the odds ratio of treatment response. Additionally, this meta-analysis examined moderators of BT treatment effects that included: co-occurring OCD and ADHD; number of 1-h therapy sessions; participant age; methodological quality; and tic medication status. 2. Method 2.1. Search strategy PubMED, PsycInfo, and ProQuest Dissertations and Theses Online were searched using key search terms (i.e., “behavior therapy” or “habit reversal” AND either “Tourette” or “chronic tic”). Identified abstracts were reviewed independently by two raters (JM, EB) for appropriateness. The references of eligible treatment trials, and review articles were also searched. Identified abstracts/citations were evaluated for the following inclusion criteria: (1) a RCT; (2) examined the efficacy of BT in treating TS; (3) available in English; and (4) provided sufficient data to allow calculation of treatment effects. Trials were considered randomized when study investigators explicitly represented them as such. Treatments were classified as BT when they included awareness training and competing response training. When treatment effect data was not sufficiently reported, study investigators were contacted to obtain values. If treatment effect data were unavailable for separate conditions and/or the authors did not respond to requests, the trial was excluded from analyses (Azrin et al., 1980). 2.2. Procedures 2.2.1. Outcome measures used to quantify treatment effects Consistent with previous TS meta-analyses, a hierarchy of preferred tic rating scales for the primary outcome measure was established a priori to limit possible investigator reporting bias (Weisman et al., 2012). In order of preference, these tic rating scales included the YGTSS (Leckman et al., 1989), Tourette Syndrome Global Scale (Harcherik et al., 1984), Shapiro Tourette Syndrome Severity Scale (Shapiro and Shapiro, 1984), and Hopkins Motor/ Vocal Tic Severity Scale (Walkup et al., 1992). In the absence of a standardized clinician-administered rating, direct observation of tics (e.g., tic frequency counts, ratings of tic severity) was used. Preference was given to in-clinic observations over home observations due to standardization of recording procedures. 2.2.2. Study coding Trials were coded for the following characteristics: (1) sample size; (2) mean age; (3) percent of sample with ADHD and OCD; (4) percent of sample on tic medication; (5) average number of 1-

108

J.F. McGuire et al. / Journal of Psychiatric Research 50 (2014) 106e112

h therapy sessions; (7) standard mean difference; (8) treatment response on the CGI-Improvement; (9) comparison condition; and (10) study methodology. Comorbid ADHD and OCD were selected due to their common co-occurrence among individuals with TS (Freeman et al., 2000). Tic medication was classified as any antipsychotic medication or alpha-2 agonist. Comparison interventions were classified as either wait-list comparisons (WL), non-directive treatment comparisons (i.e., supportive therapy, psychoeducationsupportive therapy; ND), or behavioral therapy component comparisons (i.e., awareness training, exposure and response prevention; BTC). Study methodology was assessed using a 23-item scale (Moncrieff et al., 2001). Possible scores range from 0 to 46, with higher values corresponding with greater methodological rigor. Study investigators were contacted to request the above information if it was not available in published form. Trials were coded by two raters to ascertain reliability. Rater disagreement was resolved through discussion and consensus.

effect of BT trials (Borenstein et al., 2009). Finally, Rosenthal’s Failsafe N (Rosenthal, 1991), Orwin’s Fail-safe N (Orwin, 1983), and comparison of treatment effects across control conditions were calculated for sensitivity analyses. Rosenthal’s Fail-safe N determines the number of un-retrieved studies (k) with a mean effect of zero that would be needed to make current findings nonsignificant. Orwin’s Fail-safe N determines the number of unretrieved studies (k) with a mean effect of zero that would be needed to reduce the summary ES to a trivial effect. Based on placebo response ES observed pharmacological treatment trials (Cubo et al., 2013), the trivial ES was set at 0.16. An analog to the analysis of variance (ANOVA) examined the heterogeneity of ES across control conditions (WL, ND, BTC) using the Q statistic, with followup pair-wise comparisons. 3. Results 3.1. Included studies

2.2.3. Effect size (ES) calculation The SMD was chosen as the treatment ES statistic for the metaanalysis because it facilitated comparison with previous pharmacological meta-analyses for TS (Weisman et al., 2012), and was calculated in Comprehensive Meta-Analysis (CMA)Version 2 (Borenstein et al., 2005). Effect sizes were calculated using change scores because this increases the precision of ES estimators by controlling for pretreatment group differences of tic severity. The mean change in control group from pre-treatment to posttreatment was subtracted from the mean change in the treatment group from pre-treatment to post-treatment and was then divided by the pooled change standard deviation. Effect sizes were standardized so that a positive result indicated that BT performed better than control conditions. 2.2.4. Odds ratio (OR) calculation The OR is the ratio of the odds of an event (treatment response) occurring in one group to the odds of the event in another group. An odds ratio of one indicates that treatment response is equally likely in both groups. The CGI-Improvement rating of “much improved” or “very much improved” was used to classify participants as treatment responders. The number of treatment responders and non-responders was entered into CMA, which calculated the OR for each trial. 2.3. Statistical analyses First, inter-rater agreement of study characteristics and quality ratings was assessed using descriptive statistics and intra-class correlation coefficient (ICC). Second, a random effects model using inverse variance weights examined the SMD of BT in CMA (Borenstein et al., 2005). A random effects model was chosen because the true ES were expected to vary across trials due to different study characteristics (Borenstein et al., 2009). Heterogeneity of ES was assessed using the forest plot, Q statistic, and I2 statistic. Third, a random effects model using inverse variance weights examined the OR for treatment response among BT trials. Treatment response on the CGI-Improvement was used to calculate the number needed to treat (NNT). The NNT is the number of patients with TS that would need to be treated with BT for one patient to respond who would not have responded to the comparison intervention. Fourth, moderator variables were analyzed using method-of-moments meta-regression. Fifth, publication bias was assessed by visual inspection of the funnel plot and Egger’s test for bias. Duval and Tweedie’s trim-and-fill method was used to account for potential publication bias by taking into account unpublished studies within the field, and provided an adjusted summary

Initial search strategies produced 507 potential abstracts/citations, with 14 abstracts/citations being retrieved for detailed review (see Fig. 1). Table 1 displays the eight RCTs that met inclusion criteria, which produced a total sample of 438 participants. Two trials compared BT to a wait-list condition (Azrin and Peterson, 1990; O’Connor et al., 2001), and six trials compared a BT to an active comparison condition [e.g., awareness training (Piacentini et al., 2002), supportive therapy (Deckersbach et al., 2006; Wilhelm et al., 2003), psychoeducation and supportive therapy (Piacentini et al., 2010; Wilhelm et al., 2012), and ERP (Verdellen et al., 2004)]. 3.2. Reliability of coding study characteristics There was excellent inter-rater agreement between the two raters on categorical and continuous study characteristics (100% agreement), as well as overall study methodological quality (ICC ¼ 0.98, 95% CI ¼ 0.91, 0.99). 3.3. Treatment effects of BT As seen in Fig. 2, a random effects meta-analysis identified a medium to large treatment effect of BT compared to all control conditions (SMD ¼ 0.67, 95% CI: 0.22, 1.12, z ¼ 2.91, p ¼ 0.004). Visual inspection of the forest plot, Q statistic, and I2 statistic identified the presence of significant heterogeneity [Q(7) ¼ 29.32, p < 0.001, I2 ¼ 76.13%]. As Verdellen et al., 2004 and Piacentini et al., 2002 used behavioral interventions as comparison conditions, the summary effect was re-calculated with these two trials excluded. Results identified a large treatment effect (SMD ¼ 0.94, 95% CI: 0.54, 1.34, z ¼ 4.58, p < 0.001), with less heterogeneity observed [Q(5) ¼ 12.61, p ¼ 0.03, I2 ¼ 60.34%]. 3.4. Treatment response from BT As seen in Fig. 3, there was a high OR in favor of participants receiving BT experiencing a treatment response (OR ¼ 5.77, 95% CI: 3.23, 10.30, z ¼ 5.92, p < 0.001), with little heterogeneity observed across trials [Q(3) ¼ 2.23, p ¼ 0.53, I2 ¼ 0.00%]. Treatment response on the CGI-Improvement produced a NNT of three across trials (95% CI: 2.3, 4.2). 3.5. Moderators of BT First, when examining comorbidity, no significant association was identified between the percent of study participants with OCD

J.F. McGuire et al. / Journal of Psychiatric Research 50 (2014) 106e112

109

Fig. 1. Study selection and rationale for exclusion.

and ES (B ¼ 0.02, p ¼ 0.64). However, a small negative association existed between the percent of study participants with ADHD and ES (B ¼ 0.05, 95% CI: 0.09,0.002, z ¼ 2.07, p ¼ 0.04). Second, when examining the therapeutic contact, a small positive association between the average number of therapy sessions and ES was identified (B ¼ 0.18, 95% CI: 0.03, 0.33, z ¼ 2.36, p ¼ 0.02). Third, a small positive association was found between the average participant age in each trial and ES (B ¼ 0.05, 95% CI: 0.001, 0.09, z ¼ 2.05, p ¼ 0.04). Fourth, no significant association was found between methodological quality and ES (B ¼ 0.01, p ¼ 0.64). Finally, no significant association was identified between the percent of study participants on tic-influencing medications and ES (B ¼ 0.02, p ¼ 0.26).

3.6. Publication bias and sensitivity analyses Although visual inspection of the funnel plot suggested publication bias may exist, Egger’s test for bias indicated that publication bias was not significant (p ¼ 0.57). When Duval and Tweedie’s trimand-fill method was applied, no studies were trimmed and BT still exhibited a medium to large effect (SMD ¼ 0.67, p ¼ 0.004). Rosenthal’s Fail-safe N calculations identified that at least 62 unretrieved studies with an effect size of zero were needed to reduce the summary ES of BT to a non-significant level. Meanwhile, Orwin’s Fail-safe N calculations identified that only 21 unretrieved studies with a mean effect of zero were needed to bring the summary ES to under 0.16. The analog to ANOVA revealed that

Table 1 Study characteristics and coded variables for studies included in meta-analysis. Author

Control

Azrin and Peterson, 1990

WL

N 10

O0 Connor et al., 2001

WL

50

Piacentini et al., 2002

BTC

25

a

Mean age

% With ADHD

% With OCD

% On tic medication

18

NA

NA

30%

# Of sessions 7

Study quality 25

Effect size 0.48

37

2%

10%

2%

13

17

1.41

11

20%

28%

48%

8

25

0.09

b

Outcome measure

Treatment response

Video observations of tic frequency Video observations of tic frequency and severity (0e100) YGTSS

NA

c

Wilhelm et al., 2003 Verdellen et al., 2004 Deckersbach et al., 2006

ND BTC ND

32 43 30

35 21 35

NA 30% NA

NA 14% 30%

48% 40% 33%

14 10 14

19 21 26

1.38 0.52 1.67

YGTSS YGTSS YGTSSc

Piacentini et al., 2010

ND

126

12

26%

19%

37%

8

40

0.57

YGTSS

Wilhelm et al., 2012

ND

122

32

28%

18%

25%

8

38

0.53

YGTSS

NA

HRT: 6/13 AT: 3/12 NA NA HRT: 10/15 SP: 2/15 CBIT: 32/61 PST: 12/65 CBIT: 24/63 PST: 4/59

Note: WL ¼ Waitlist, BTC ¼ Contains active BT components, ND ¼ Nondirective, ADHD ¼ Attention Deficit Hyperactivity Disorder, OCD ¼ Obsessive Compulsive Disorder, YGTSS ¼ Yale Global Tic Severity Scale, HRT ¼ Habit Reversal Training, AT ¼ Awareness Training, SP ¼ Supportive Psychotherapy, CBIT ¼ Comprehensive Behavioral Intervention for Tics, PST ¼ Psychoeducation and Supportive Therapy, NA ¼ Not Available or Not Applicable. a Although 69 participants with TS participated in the trial, video observation of frequently and severity data available for only 50 participants (40 HRT/CBT, 10 Waitlist control). b Combined effect size for video ratings of frequency and severity. c Evaluator/rater not blind to treatment condition.

110

J.F. McGuire et al. / Journal of Psychiatric Research 50 (2014) 106e112 Statistics for each study

Study name

Std diff Standard in means error Variance Azrin & Peterson, 1990 O'Connor et al., 2001 Piacentini et al., 2002 Wilhelm et al., 2003 Verdellen et al., 2004 Deckersbach et al., 2006 Piacentini et al., 2010 Wilhelm et al., 2012

0.48 1.41 0.09 1.38 -0.52 1.67 0.57 0.53

0.64 0.38 0.40 0.42 0.31 0.42 0.18 0.18

0.41 0.15 0.16 0.17 0.10 0.18 0.03 0.03

Std diff in means and 95% CI

Lower Upper limit limit Z-Value p-Value -0.78 0.66 -0.69 0.57 -1.13 0.84 0.22 0.17

1.74 2.16 0.88 2.20 0.09 2.50 0.93 0.89

0.75 3.70 0.23 3.33 -1.68 3.94 3.15 2.87

0.453 0.000 0.818 0.001 0.092 0.000 0.002 0.004

-2.75

-1.38

0.00

Control Condition

1.38

2.75

Behavior Therapy

Fig. 2. Efficacy of BT compared to wait-list and active comparison control conditions for the treatment of tics.

heterogeneous ES were observed between trials using WL (SMD ¼ 1.09), ND (SMD ¼ 0.90) and BTC (SMD ¼ 0.27) [Q(2) ¼ 10.12, p ¼ 0.004]. Follow-up pair-wise comparisons revealed heterogeneous ES between BTC and ND [Q(1) ¼ 9.18, p ¼ 0.002] and BTC and WL [Q(1) ¼ 6.39, p ¼ 0.012], but not between ND and WL [Q(1) ¼ 0.15, p ¼ 0.700]. 4. Discussion This meta-analysis found a medium to large treatment effect for BT across RCTs for TS (SMD ¼ 0.67). This treatment effect was noticeably larger and exhibited less heterogeneity when excluding two trials that used behavioral interventions as comparison conditions (SMD ¼ 0.94). Observed summary ES of BT trials are consistent with effects observed in antipsychotic meta-analyses (SMD ¼ 0.58), and exceed treatment effects identified in metaanalyses of alpha-2 agonists trials (SMD ¼ 0.31) (Weisman et al., 2012). While TS has traditionally been managed with pharmacotherapy, this quantitative synthesis suggests BT presents an alternative treatment option with comparable treatment effects to psychotropic medicationssupporting current treatment recommendations by some professional organizations that BT serve as a first-line treatment for TS (Steeves et al., 2012; Verdellen et al., 2011). When examining treatment response on the CGIImprovement, participants receiving BT were five times more likely to respond to treatment compared to individuals receiving comparison interventions. Calculations identified an NNT of three, which indicates that three patients with TS would need to be treated with BT for one patient to respond who would not have responded to the comparison intervention. Categorical treatment response is an important metric because BT focuses on implementing strategies to manage tics during situations in which tics cause impairment versus a global reduction in Study name

tics. For instance, an adult patient may try to regularly implement competing responses in the workplace and social environments, but may be less inclined to practice these strategies at home where the social costs of tics are less. Although the patient in this example may only exhibit a moderate reduction on tic rating scales, this patient would most likely experience significant overall improvement because of reduced work or social impairment. Indeed, this improved level of tic control may account for findings that 25e35% reductions in tic severity on the YGTSS correspond with ratings of treatment response among youth and adults (Jeon et al., 2013; Storch et al., 2011). Moderator analyses identified that trials with more therapy sessions and older mean participant age exhibited greater ES. The relationship between an increased number of therapy sessions and ES is consistent with findings on the dose-relationships in medication trials among related disorders (e.g., OCD) (Bloch et al., 2010). Although the amount of therapeutic contact likely plays a role in enhancing treatment, an alternative explanation may be that trials with a greater number of therapy sessions also had a longer duration. Indeed, trial duration demonstrated a significant association with ES among alpha-2 agonist trials, with longer trial durations being associated with greater ES (Weisman et al., 2012). Future studies are needed to determine the role and impact of the number of therapeutic contact hours and trial duration on BT treatment effects. There was a marginal, albeit significant, association between older mean participant age and larger treatment effects. This finding may be explained in several ways. First, it may be that older participants are more readily able to implement strategies learned in BT. Second, two BT trials that focused specifically on youth included participants under the age of 10. These participants may have had greater difficulty with awareness training–a core therapeutic component of BT–due to difficulty with premonitory urge

Statistics for each study

Odds ratio and 95% CI

Odds Lower Upper ratio limit limit Z-Value p-Value

Piacentini et al., 2002 2.57 Deckersbach et al., 2006 13.00 Piacentini et al., 2010 4.87 Wilhelm et al., 2012 8.46

0.47 2.07 2.18 2.72

14.10 81.48 10.88 26.33

1.09 2.74 3.86 3.69

0.28 0.01 0.00 0.00 0.01

0.1

1

10

100

Control Condition Behavior Therapy Fig. 3. Treatment Response of BT compared to active comparison conditions.

J.F. McGuire et al. / Journal of Psychiatric Research 50 (2014) 106e112

recognition (Woods et al., 2005). Finally, only two trials specifically focused on youth, with one of these trials exhibiting minimal treatment effects (Piacentini et al., 2002). Overall, this marginal moderator effect may be likely attributed to the limited sample of available studies focused on youth. Trials with a greater percentage of participants with ADHD exhibited attenuated BT treatment effects. This is contrary to findings among alpha-2 agonists trials in which studies with a greater proportion of subjects with ADHD reported greater efficacy (Weisman et al., 2012). For individuals with TS and co-occurring ADHD, it may be that symptoms of ADHD impede their ability to engage in therapy sessions, and/or has a negative impact on tic suppression (Lyon et al., 2010). Although providing preliminary evidence that ADHD may attenuate treatment effects of BT, two things should be noted. First, studies with a high percentage of ADHD still exhibited sizeable treatment effects to BT. Second, participant ADHD medication status in BT trials was unknown. It may be that participants whose ADHD symptoms are appropriately managed may still significantly benefit from BT. Future research is needed to clarify whether managed or unmanaged ADHD symptoms influences treatment effects of BT in individuals with TS. Finally, there was no significant association between ES of BT trials and the percent of study participants on tic-influencing medications suggesting that concurrent tic medication does not affect BT response. On balance, these non-significant results should be interpreted with caution given modest power to detect moderator effects (Borenstein et al., 2009). Sensitivity analyses found no evidence of publication bias. Calculations for Rosenthal’s and Orwin’s Fail-safe N suggested that between 21 and 62 un-retrieved studies were needed with nonsignificant effects to make the findings non-significant or trivial, respectively. Although findings from BT trials appear robust to file drawer phenomena, further trials of BT are needed. Additional trials would prove useful to replicate results from existing BT trials, and examine treatment moderators. Given the developmental course of TS, it is critically important to understand the effect of BT in youth and whether moderator effects may differ across TS course. When examining ES across comparison interventions, significant heterogeneity was observed. Follow-up comparisons revealed that trials with BTC as the control treatment had different ESs compared to trials with either ND or WL conditions. Given that trials with BTC incorporated key BT components, it is expected that more modest effects would be observed. Although WL and ND trials may be likened to placebo-controlled trials for medications, comparisons of BT to BTC may be more akin to head-to-head medication trials due to inclusions of active therapeutic ingredients. Thus, BT has been largely compared to active treatments (BTC, ND) and still demonstrated therapeutic benefit, but notably exhibits minimal benefit when compared to related interventions. Several limitations should be considered. First, only eight trials were included and thus this meta-analysis was underpowered to detect moderators (Borenstein et al., 2009). Proposed moderators that did not reach significance should not be interpreted as a conclusive lack of association. Conversely, moderators that were significant suggest that the identified association was robust enough to be detected amidst limited power. Second, this metaanalysis collapsed trials that included related yet differing BT treatment packages. Although all BT trials included awareness and competing response training, it may be that differing treatment packages are more beneficial than others. Third, when drawing comparisons to medication trials, it’s important to consider that participant characteristics may be different between pharmacological and behavioral trials. Future head-to-head trials would prove useful to determine comparative efficacy between behavior therapy and psychotropic medication. Fourth, there were limited

111

factors available for extraction across BT trials. Although theoretically driven variables were selected, there may be unexamined factors (e.g., homework compliance) omitted from these reports that may influence treatment effects. In summary, findings suggest that BT is efficacious in reducing tic severity and yields comparable treatment effects to pharmacotherapy trials. Individuals receiving BT are more likely to exhibit a treatment response in comparison to control interventions. Although participants’ ADHD medication status remains unknown, moderator analyses suggest that trials that include greater percentages of participants with co-occurring ADHD may exhibit attenuated treatment effects to BT. Meanwhile, BT trials that have an older mean participant age and an elevated therapeutic contact hours are likely to exhibit greater treatment effects. Contributors J. McGuire designed the meta-analysis, analyzed data, and wrote the first draft of the manuscript. J. Piacentini assisted with the design of the meta-analysis, data collection and manuscript preparation. E. Brennan assisted with data collection and manuscript preparation. A. Lewin assisted with the design of the meta-analysis and manuscript preparation. T. Murphy assisted with the design of the meta-analysis and manuscript preparation. B Small assisted in the design of the meta-analysis, data analysis, and manuscript preparation. E. Storch assisted with the design of the meta-analysis and manuscript preparation. Disclosures Mr. McGuire and Ms. Brennan report no relevant disclosures. Dr. Piacentini has received support from NIMH, the Tourette Syndrome Association (TSA), and the Obsessive Compulsive Foundation, book royalties from Guilford Press and Oxford University Press, including the Comprehensive Behavioral Intervention for Tics manual discussed in this paper, and serves on the speakers’ bureau for the TSA. Dr. Lewin has served as a consultant for Otsuka America Pharmaceutical and ProPhase, Inc. He has received grant support from International Obsessive Compulsive Disorder Foundation; National Alliance for Research on Schizophrenia and Depression; University of South Florida Research Foundation, Inc. He has received travel support from University of South Florida Research Foundation, Inc. Dr. Murphy has received research funding from NIH/NIMH, CDC, Otsuka Pharmaceuticals, NARSAD, IOCDF, Ortho-McNeil Janssen Pharmaceuticals, Shire Pharmaceuticals, Pfizer, Inc. and Indevus Pharmaceuticals. She has received travel support from the Tourette Syndrome Association and honorarium from grand rounds lectures. Dr. Small has received support from the National Institute of Aging (NIA), National Cancer Institute (NCI) and the American Cancer Society (ACS). Dr. Storch has received grant funding in the last 2 years from the National Institutes of Health, Centers for Disease Control, Agency for Healthcare Research and Quality, National Alliance for Research on Schizophrenia and Affective Disorders, International OCD Foundation, Tourette Syndrome Association, and Janssen Pharmaceuticals. He receives textbook honorarium from Springer publishers, American Psychological Association, and Lawrence Erlbaum. Dr. Storch has been an educational consultant for Rogers Memorial Hospital. He is a consultant for Prophase, Inc. and CroNos, Inc., and is on the Speaker’s Bureau and Scientific Advisory Board for the International OCD Foundation. He receives research support from the All Children’s Hospital Guild Endowed Chair.

112

J.F. McGuire et al. / Journal of Psychiatric Research 50 (2014) 106e112

Role of the funding source None. Acknowledgments The authors would like to express their gratitude to Dr. Kieron O’Connor and Dr. Cara Verdellen for their assistance in obtaining values not published in their original articles. Additionally, the authors would like to express their sincere appreciation to Dr. Michael Brannick for his statistical expertise and feedback. References Azrin NH, Nunn RG, Frantz SE. Habit reversal vs negative practice treatment of nervous tics. Behav Ther 1980;11:169e78. Azrin NH, Peterson AL. Treatment of Tourette Syndrome by habit reversal: a waiting-list control group comparison. Behav Ther 1990;21:305e18. Bate KS, Malouff JM, Thorsteinsson ET, Bhullar N. The efficacy of habit reversal therapy for tics, habit disorders, and stuttering: a meta-analytic review. Clin Psychol Rev 2011;31:865e71. Bloch MH, Leckman JF. Clinical course of Tourette Syndrome. J Psychosomatic Res 2009;67:497e501. Bloch MH, McGuire J, Landeros-Weisenberger A, Leckman JF, Pittenger C. Metaanalysis of the dose-response relationship of SSRI in obsessive-compulsive disorder. Mol Psychiatry 2010;15:850e5. Borenstein M, Hedges L, Higgins J, Rothstein H. Comprehensive meta-analysis (version 2). Englewood, New Jersey: Biostat; 2005. Borenstein M, Hedges LV, Higgins JPT, Rothstein HR. Introduction to meta-analysis. Wiley; 2009. Conelea CA, Woods DW, Zinner SH, Budman C, Murphy TK, Scahill LD, et al. Exploring the impact of chronic tic disorders on youth: results from the Tourette Syndrome Impact Survey. Child Psychiatry Hum Dev 2011;42:219e42. Conelea CA, Woods DW, Zinner SH, Budman CL, Murphy TK, Scahill LD, et al. The impact of Tourette Syndrome in adults: results from the Tourette Syndrome Impact Survey. Community Ment Health J 2013;49:110e20. Centers for Disease Control and Prevention. Prevalence of diagnosed Tourette Syndrome in persons aged 6e17 years e United States, 2007. Morbidity and Mortality Weekly Report, Department of Health and Human Services, Centers for Disease Control 2009;58:581e5. Cubo E, Gonzalez M, Singer H, Mahone M, Scahill L, Muller-Vahl KR, et al. Impact of placebo assignment in clinical trials of tic disorders. Mov Disord 2013;28: 1288e92. Deckersbach T, Rauch S, Buhlmann U, Wilhelm S. Habit reversal versus supportive psychotherapy in Tourette’s disorder: a randomized controlled trial and predictors of treatment response. Behav Res Ther 2006;44:1079e90. Freeman RD, Fast DK, Burd L, Kerbeshian J, Robertson MM, Sandor P. An international perspective on Tourette Syndrome: selected findings from 3,500 individuals in 22 countries. Dev Med Child Neurol 2000;42:436e47. Guy W, Bonato R, editors. CGI: clinical global impressions. Chevy Chase, MD: National Institute for Mental Health; 1970. Harcherik DF, Leckman JF, Detlor J, Cohen DJ. A new instrument for clinical studies of Tourette’s syndrome. J Am Acad Child Psychiatry 1984;23:153e60. Jalenques I, Galland F, Malet L, Morand D, Legrand G, Auclair C, et al. Quality of life in adults with Gilles de la Tourette Syndrome. BMC Psychiatry 2012;12:109. Jeon S, Walkup JT, Woods DW, Peterson A, Piacentini J, Wilhelm S, et al. Detecting a clinically meaningful change in tic severity in Tourette Syndrome: a comparison of three methods. Contemp Clin Trials 2013;36:414e20. Leckman JF, Riddle MA, Hardin MT, Ort SI. The Yale Global Tic Severity Scale: initial testing of a clinician-rated scale of tic severity. J Am Acad Child Adolesc Psychiatry 1989;28:566e73. Lyon GJ, Samar SM, Conelea C, Trujillo MR, Lipinski CM, Bauer CC, et al. Testing tic suppression: comparing the effects of dexmethylphenidate to no medication in children and adolescents with attention-deficit/hyperactivity disorder and Tourette’s disorder. J Child Adolesc Psychopharmacol 2010;20:283e9. McGuire JF, Kugler BB, Park JM, Horng B, Lewin AB, Murphy TK, et al. Evidencebased assessment of compulsive skin picking, chronic tic disorders and trichotillomania in children. Child Psychiatry Hum Dev 2012;43:855e83.

McGuire JF, Nyirabahizi E, Kircanski K, Piacentini J, Peterson AL, Woods DW, et al. A cluster analysis of tic symptoms in children and adults with Tourette Syndrome: clinical correlates and treatment outcome. Psychiatry Res 2013;210: 1198e204. Miltenberger RG, Fuqua RW, Woods DW. Applying behavior analysis to clinical problems: review and analysis of habit reversal. J Appl Behav Anal 1998;31: 447e69. Moncrieff J, Churchill R, Drummond DC, McGuire H. Development of a quality assessment instrument for trials of treatments for depression and neurosis. Int J Met Psychiatric Res 2001;10:126e33. Murad MH, Montori VM. Synthesizing evidence: shifting the focus from individual studies to the body of evidence. J Am Med Assoc 2013;309:2217e8. Murphy TK, Lewin AB, Storch EA, Stock S. the AACAP Committee on Quality Issues. Practice parameter for the assessment and treatment of children and adolescents with Tourette’s disorder and chronic tic disorders. J Am Acad Child Asolesc Psychiatry 2013;52:1341e59. O’Connor KP, Brault M, Robillard S, Loiselle J, Borgeat F, Stip E. Evaluation of a cognitive-behavioural program for the management of chronic tic and habit disorders. Behav Res Ther 2001;39:667e81. Orwin RG. A fail-safe N for effect size in meta-analysis. J Educ Statistics 1983;8: 157e9. Piacentini J, Chang S, Barrios V, McCracken J. Habit reversal training for childhood tic disorders: a randomized controlled trial. In: Paper presented at the Association for the Advancement of Behavior Therapy, Reno, NV 2002. Piacentini J, Woods DW, Scahill L, Wilhelm S, Peterson AL, Chang S, et al. Behavior therapy for children with Tourette disorder: a randomized controlled trial. J Am Med Assoc 2010;303:1929e37. Piacentini JC, Chang SW. Behavioral treatments for tic suppression: habit reversal training. Adv Neurol 2006;99:227e33. Rosenthal R. Meta-analytic procedures for social research. Rev. ed. Thousand Oaks, CA: Sage Publications, Inc.; 1991. Scahill L, Erenberg G, Berlin CM, Budman C, Coffey BJ, Jankovic J, et al. Contemporary assessment and pharmacotherapy of Tourette Syndrome. NeuroRx: J Am Soc Exp Neuro Ther 2006;3:192e206. Shapiro AK, Shapiro E. Controlled study of pimozide vs. placebo in Tourette’s syndrome. J Am Acad Child Psychiatry 1984;23:161e73. Steeves T, McKinlay BD, Gorman D, Billinghurst L, Day L, Carroll A, et al. Canadian guidelines for the evidence-based treatment of tic disorders: behavioural therapy, deep brain stimulation, and transcranial magnetic stimulation. Can J Psychiatry/La Revue Can de psychiatrie 2012;57:144e51. Storch EA, De Nadai AS, Lewin AB, McGuire JF, Jones AM, Mutch PJ, et al. Defining treatment response in pediatric tic disorders: a signal detection analysis of the Yale Global Tic Severity Scale. J Child Adolesc Psychopharmacol 2011;21:621e7. Storch EA, Merlo LJ, Lack C, Milsom VA, Geffken GR, Goodman WK, et al. Quality of life in youth with Tourette’s syndrome and chronic tic disorder. J Clin Child Adolesc Psychol 2007;36:217e27. Verdellen C, van de Griendt J, Hartmann A, Murphy T. European clinical guidelines for Tourette Syndrome and other tic disorders. Part III: behavioural and psychosocial interventions. Eur Child Adolesc Psychiatry 2011;20:197e207. Verdellen C, Keijsers GPJ, Cath DC, Hoogduin CAL. Exposure with response prevention versus habit reversal in Tourettes’s syndrome: a controlled study. Behav Res Ther 2004;42:501e11. Walkup JT, Rosenberg LA, Brown J, Singer HS. The validity of instruments measuring tic severity in Tourette’s syndrome. J Am Acad Child Adolesc Psychiatry 1992;31:472e7. Weisman H, Qureshi IA, Leckman JF, Scahill L, Bloch MH. Systematic review: pharmacological treatment of tic disorders e efficacy of antipsychotic and alpha-2 adrenergic agonist agents. Neurosci Biobehav Rev 2012;37:1162e71. Wile DJ, Pringsheim TM. Behavior therapy for Tourette Syndrome: a systematic review and meta-analysis. Curr Treat Opt Neurol 2013;15:385e95. Wilhelm S, Deckersbach T, Coffey BJ, Bohne A, Peterson AL, Baer L. Habit reversal versus supportive psychotherapy for Tourette’s disorder: a randomized controlled trial. Am J Psychiatry 2003;160:1175e7. Wilhelm S, Peterson AL, Piacentini J, Woods DW, Deckersbach T, Sukhodolsky DG, et al. Randomized trial of behavior therapy for adults with Tourette Syndrome. Archives Gen Psychiatry 2012;69:795e803. Woods DW, Piacentini J, Chang SW, Deckersbach T, Ginsburg GS, Peterson AL, et al. Managing Tourette Syndrome: a behavioral intervention for children and adolescents. New York: Oxford University Press; 2008. Woods DW, Piacentini J, Himle MB, Chang S. Premonitory Urge for Tics Scale (PUTS): initial psychometric results and examination of the premonitory urge phenomenon in youths with Tic disorders. J Dev Behav Pediatrics 2005;26: 397e403.