Systematic review methods
An objective synthesis of the evidence to evaluate the safety and effectiveness of digitally enabled non-pharmacological therapy as compared to standard care for treating CTDs and TS in UK clinical practice. This assessment was conducted according to current methodological standards. The methods were pre-specified in a research protocol (www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42024508045).
Identification of studies
An Information Specialist developed a comprehensive literature search strategy to identify relevant published peer-reviewed studies. Major electronic databases were searched, including MEDLINE, EMBASE, Cochrane Library, Web of Science, and Cumulative Index to Nursing and Allied Health Literature (CINAHL). The focus of the search initially was on the approved devices listed in the NICE final scope; search facets defining the population of interest were included. There were no restrictions on the date or language of publication at the time of the search. The reference lists of studies selected for full-text appraisal were screened for additional studies. Major clinical trial registries were searched to identify relevant ongoing trials. Websites of manufacturers of appropriate technologies, professional organisations and regulatory bodies were searched to identify additional relevant reports. Any further information on potentially relevant evidence provided by the manufacturers of the technologies of interest was also considered. All references were exported to Endnote for recording and deduplication. A draft MEDLINE search is detailed in Appendix 1. The MEDLINE search was adapted to search other electronic databases.24
Inclusion and exclusion criteria
The eligibility criteria for the review of clinical effectiveness evidence are summarised in Table 2.
| Population of interest | Children/young people and adults diagnosed with a confirmed primary, CTD |
| Clinical condition | Primary, CTDs, including TSTransient and secondary tic disorders and functional tic-like behaviours were not considered eligible for inclusion |
| Technologies under investigation | ORBIT (MindTech)Neupulse (Neurotherapeutics) |
| Comparator intervention | Standard care, including psychoeducation and face-to-face behavioural therapy |
| Outcome measures | Intermediate outcome measures Intervention-related AEsTreatment satisfaction and engagementIntervention adherence, rates of attrition and completion |
| Clinical outcome measures Measures of symptom severity (self, parental or practitioner reported) using validated instruments such as the Yale Global Tic Severity Scale (YGTSS)Tools for depression and anxiety such as Patient Health Questionnaire for adolescents, Children’s Depression Inventory and Beck Depression InventorySocial, behavioural, and functional outcomesSuicidal thoughts and behaviour | |
| Patient-reported outcome measures Health-related quality of life (HRQoL)Patient’s experience and patient’s satisfactionRates and reasons for attrition | |
| Study design | Clinical studies assessing the efficacy or effectiveness of non-pharmacological treatment delivered remotely or online using digital technologiesWe included randomised controlled trials (RCTs), and comparative non-randomised studies published in English. Articles available in their pre-publication version and relevant reports submitted by the manufacturers of the technologies under investigation were considered for inclusion. Crossover studies (phase before crossover) and evidence from uncontrolled studies (in the absence of evidence from comparative studies) were also considered eligible. Conference abstracts were excluded because they were not considered to provide sufficient information |
| Healthcare setting | Secondary care settings (e.g. CYPMHS)Tertiary care settings (e.g. neurology or neurodevelopmental teams – including neurologists, neuropsychologists, psychiatrists, psychologists, specialist nurses, speech and language therapists) |
Study selection and data extraction
One reviewer, Moira Cruickshank (MC), screened all citations identified by the search strategies. A second reviewer, Miriam Brazzelli (MB), independently screened a random 20% sample (selected using a random number generator, random.org, with selection based on numerical position in the list of citations). Two reviewers (MC, MB) independently assessed each full-text article for eligibility. Reasons for exclusion were documented. The research protocol specified that data extraction had to be conducted by two independent reviewers. However, due to time and resource constraints, one reviewer MC extracted data, which was subsequently cross-checked by a second reviewer MB. A customised Excel data extraction spreadsheet was developed for this assessment. The following information was recorded from each study:
- Characteristics of studies: first author, year of publication, country, language, setting, inclusion and exclusion criteria.
- Characteristics of study participants: age, sex, tic typology, comorbidities, number of enrolled participants, number of participants analysed, number of dropouts and reasons for withdrawal, setting.
- Characteristics of the intervention: digital platform, details of the technology, content of therapy, structure and number of sessions to be completed, duration, type and frequency of contact with a therapist, and therapist’s level of expertise.
- Characteristics of the comparator/control intervention: nature and mode of delivery, duration, type and frequency of contact with a therapist, and therapist’s level of expertise.
- Relevant patient-reported, clinical and intermediate outcome measures, and information related to the use of digital technologies.
At all stages, disagreements were resolved through discussion.24
Data synthesis
The findings of each included study were tabulated and summarised narratively for each outcome of interest. We combined results from two of the included studies that were considered sufficiently similar in terms of intervention/comparator, participants and outcome measures.34,38 We conducted random-effects model meta-analyses to pool unadjusted mean differences and 95% CIs for the Total Tic Severity Score (TTSS) and the impairment score of the YGTSS using the inverse variance method. It is worth noting that the TTSS score of the YGTSS ranges from 0 to 50 and measures the severity of the tic disorders (where 50 indicates higher severity). The YGTSS impairment score (YGTSS-Impairment) also ranges from 0 to 50. A score of 0 indicates that the presence of TS has no negative impact on a person’s daily life. In contrast, a score of 50 indicates considerable interference and disability associated with the presence of TS. Heterogeneity between studies was assessed using the I2 statistic. Statistical analyses were conducted using the Cochrane software for systematic reviews and meta-analysis (RevMan Web). When appropriate, the following subgroups were considered: children with diagnosed comorbidities (ADHD, OCD, ASD, mood disorders, anxiety) and adults with CTDs.
Assessment of risk of bias
The Cochrane risk-of-bias tool (version 2) was used to assess the risk of bias for the primary outcome measured by the Yale Global Tic Severity Scale – Total Tic Severity Score (YGTSS-TTSS) in the randomised controlled trials (RCTs) included in the review.39 One reviewer (MC) assessed each included study, and a second reviewer (MB) cross-checked the assessments. Disagreements were resolved by consensus.
Results of the assessment of clinical effectiveness
Results of the literature searches
The literature searches identified 379 titles/abstracts. Three further studies were identified: one40 was identified from a reference list,36 one was provided by the corresponding author of the Maiquez 2023 study41 and the third42 was identified from the website search. A total of 40 publications were selected for full-text screening, of which 14 met our inclusion criteria, reporting a total of 3 studies. The Andrén 2024 study was out with our search dates but was included as a secondary publication to Andrén 202238 for completeness. A Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram detailing the process of study selection is presented in Figure 1.
FIGURE 1 Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram. (Figure not reproduced here; see the original article.)
Characteristics of included studies
A total of 3 RCTs reported in 14 publications were included in the review of clinical effectiveness evidence. Characteristics of the included studies are presented in Table 27, Appendix 2. Two studies compared ORBIT with psychoeducation; one was conducted in the UK34 and the other in Sweden.38 The third study was UK-based and compared Neupulse active stimulation with sham stimulation; this trial also included an open-label waitlist (treatment as usual) control condition, which is reported here for completeness.36 All three were published in English and full-text form. The two ORBIT studies recruited children and young people aged 9–17 years.34,38 The Neupulse study recruited people aged 12 years upwards.36 A total of 566 participants were randomised (UK ORBIT: 224, Swedish ORBIT: 221; Neupulse: 121).34,36,37 Longest follow-up periods in the ORBIT studies were 18 months (UK ORBIT)34 and 12 months (Swedish ORBIT).37
Evidence gap: evidence was not available to compare the interventions under investigation and face-to-face behavioural therapy, the current standard of care. The comparator in the two ORBIT studies was psychoeducation, while the Neupulse active stimulation arm was compared to sham stimulation and a waitlist group. Both ORBIT trials did not include a non-active control group (e.g. waitlist).
Characteristics of participants
Baseline characteristics of participants in the included studies are reported in Table 3. The mean age of participants was between 12.0 years and 12.4 years in the two ORBIT studies34,38 and between 23.5 years and 24.4 years in the Neupulse study.36 The proportion of male participants ranged from 64.0% to 80.4% in the ORBIT studies and 59.0% to 63.4% in the Neupulse study. Regarding tic typology, most participants in the UK ORBIT study had both motor and vocal tics (92% and 95% in the intervention and comparator groups, respectively), while 8% and 5%, respectively, had motor tics only, and none had only vocal tics.34 In the Swedish ORBIT study, most participants had TS (93.7% and 89.1% in the intervention and comparator groups, respectively).38 A further 6.3% and 8.2%, respectively, had chronic motor tic disorder and 2.7% of participants in the comparator group had chronic vocal tic disorder. Tic typology in the Neupulse study was not reported. Comorbidities reported by the studies included ADHD (ranging from 12.7% to 24.4%),36,38 anxiety disorder (ranging from 13.6% to 30.8%)36,38 and OCD (ranging from 3% to 41.5%).34,36
| Study ID | Intervention group (n analysed) | Age, years, mean (SD) | Male sex, % | White ethnicity, % | Tic typology, n (%) | Comorbidities, n (%) |
|---|---|---|---|---|---|---|
| Hollis 2021 34 (ORBIT UK) | ERP (n = 112) | 12.2 (2.0) (age of tic onset: NR) | 80.4 | 86 | Both motor and vocal tics: 103 (92)Motor tics only: 9 (8)Vocal tics only: 0 (0) | Anxiety disorder: 34 (30)ADHD: 26 (23)Oppositional defiant disorder: 26 (24) (n = 110)ASD: 9 (8) (n = 111)OCD: 8 (7)Major depression: 2 (2)Conduct disorder: 3 (3) (n = 110) |
| Psychoeducation (n = 112) | 12.4 (2.1) (age of tic onset: NR) | 77.7 | 88 | Both motor and vocal tics: 106 (95)Motor tics only: 6 (5)Vocal tics only: 0 (0) | Anxiety disorder: 27 (24)ADHD: 25 (22)Oppositional defiant disorder: 23 (21) (n = 111)ASD: 4 (4)OCD: 2 (3)Major depression: 6 (5)Conduct disorder: 2 (2) (n = 111) | |
| Andrén 2022 38 (ORBIT Sweden) | ERP (n = 111) | 12.0 (2.3) (age of tic onset: 5.7) | 64.0 | NR | TS: 104 (93.7)CTD motor: 7 (6.3)CTD vocal: 0 | Any: 44 (39.6)ADHD: 20 (18.0)Anxiety disorder: 16 (14.4)OCD: 11 (9.9)Depression: 1 (0.9)Other: 7 (6.3) |
| Education (n = 110) | 12.1 (2.3) (age of tic onset: 6.2) | 73.6 | NR | TS: 98 (89.1)CTD motor: 9 (8.2)CTD vocal: 3 (2.7) | Any: 40 (36.0)ADHD: 14 (12.7)Anxiety disorder: 15 (13.6)OCD: 6 (5.5)Depression: 3 (2.7)Other: 3 (2.7) | |
| Maiquez 202336 (Neupulse) | Active stimulation (n = 41) | 23.5 (12.6) (age of tic onset: 7.0) | 63.4 | NR | NR | ADHD: 10 (24.4)OCD: 17 (41.5)ASD: 8 (19.5)Anxiety disorder: 9 (22.0) |
| Sham stimulation (n = 39) | 24.0 (13.4) (age of tic onset: 8.4) | 59.0 | NR | NR | ADHD: 9 (23.1)OCD: 8 (20.5)ASD: 9 (23.1)Anxiety disorder: 12 (30.8) | |
| Waitlist (n = 41) | 24.4 (12.6) (age of tic onset: 7.5) | 63.4 | NR | NR | ADHD: 8 (19.5)OCD: 12 (29.3)ASD: 2 (4.9)Anxiety disorder: 11 (26.8) |
Evidence gap: the Neupulse study was the only study included in the review to recruit adults. The maximum age of participants in the two ORBIT trials was 17 years and the materials were developed accordingly. However, with adaptation, ORBIT may be suitable for adults.
Uncertainty: the proportion of participants with TS in the UK ORBIT trial was not reported. The mean baseline tic severity measured using the YGTSS-TTSS scores was slightly higher in the UK ORBIT study than in the Swedish study; however, both trials described the participants’ severity of tic disorders as moderate to severe.
Uncertainty: it was unclear whether the participants in the two ORBIT trials had access to psychoeducation prior to recruitment into the trials. Access to psychoeducation was not an inclusion criterion for either trial.
Risk-of-bias assessments
Figure 2 presents a summary of the risk-of-bias assessments of the three included studies for the YGTSS-TTSS primary outcome.
FIGURE 2 Summary of risk-of-bias assessments for the three included studies. (Figure not reproduced here; see the original article.)
Risk-of-bias assessments of individual studies is presented in Figure 3.
FIGURE 3 Risk-of-bias assessments of individual studies. (Figure not reproduced here; see the original article.)
According to the Cochrane risk-of-bias tool (version 2), the overall risk of bias was low for the three trials.34,36,38 All three were appropriately randomised, there was no evidence of deviations from the intended interventions, missing outcome data were low, accounted for and balanced across groups, the outcome and its assessment were appropriate and there was no evidence of selected reporting.
Clinical effectiveness results
Clinical outcomes
The primary outcome of all three studies was total tic severity assessed by the TTSS of the YGTSS scale in the two ORBIT studies34,38 and by the TTSS of the YGTSS revised in the Neupulse study.36 Tables 4 and 5 present the YGTSS-TTSS outcome data from the two ORBIT studies and the Neupulse study, respectively. The improvements in tic severity in the intervention group in the Hollis 2021 study were described by the authors as clinically important, considering an average of 0.5 of a standard deviation (SD) between the intervention and comparator as being clinically important.34 The authors of the Neupulse study defined a clinically meaningful reduction in tic severity as reduction of 25 percentiles or greater in YGTSS-TTSS scores, and reported a substantially greater proportion of the active stimulation group (59.0%) as responders than the sham stimulation group [33.3%; odds ratio (OR) 2.9, 95% CI 1.1 to 7.2].
| Study ID | Time point | ORBIT | Psychoeducation | Difference between groups at time point |
|---|---|---|---|---|
| Orbit | ||||
| Hollis 202134 | Baseline | 28.4 (7.7) | 28.4 (7.1) | N/A (not applicable) |
| 3 months | 23.9 (8.2) | 26.8 (7.3) | Estimated difference: −2.29 (95% CI −3.86 to −0.71)Effect size −0.31 (−0.52 to −0.10) | |
| 6 months | 21.5 (8.8) | 25.0 (7.6) | Estimated difference: −2.64 (95% CI −4.56 to −0.73)Effect size −0.36 (−0.62 to −0.10) | |
| 12 months | 21.7 (8.8) | 24.9 (7.3) | Estimated difference −2.64 (95% CI −4.48 to −0.79)Effect size −0.36 (95% CI −0.61 to −0.11) | |
| 18 months | 21.5 (9.0) | 23.9 (8.4) | Estimated difference −2.01 (95% CI −3.86 to −0.15)Effect size −0.27 (95% CI −0.52 to −0.02) | |
| Andrén 202238 | Baseline | 22.25 (5.60) | 23.01 (5.92) | N/A |
| 3 months | 16.17 (6.82) | 17.72 (7.11) | Intention-to-treat linear quantile mixed-model coefficient −0.53 (95% CI −1.28 to 0.22)Effect size 0.11 (95% CI −0.09 to 0.30), p = 0.17 | |
| 6 months | 16.06 (6.98) | 17.23 (8.18) | NR (not reported) | |
| 12 months | 14.93 (7.70) | 16.73 (8.30) | Interaction between treatment and time coefficient −0.38 (95% CI −1.11 to 0.35), p = 0.30Effect size 0.13 (95% CI −0.12 to 0.37) | |
| Study ID | Time point | Active stimulation | Sham stimulation | Waitlist | Difference between groups at time point |
|---|---|---|---|---|---|
| Neupulse | |||||
| Maiquez 202336 | Baseline | 40.1 (7.0) | 39.5 (6.3) | 38.9 (6.9) | N/A (not applicable) |
| 4 weeks | Mean (SD) reduction 7.13 (1.1) | Mean (SD) reduction 2.13 (0.32) | Mean (SD) reduction 2.26 (0.34) | Active vs. sham: Observed difference −5, effect size −0.47 (95% CI −0.94 to −0.02), p = 0.02Active vs. waitlist: Observed difference −5, effect size −0.48 (95% CI −0.97 to −0.04), p = 0.02 | |
Tables 6 and 7 report further clinical outcomes from the two ORBIT studies and the Neupulse study, respectively.
| Study ID | Time point | ORBIT | Psychoeducation | Difference between groups at time point |
|---|---|---|---|---|
| YGTSS-Impairment, mean (SD) | ||||
| Hollis 202134 | Baseline | 23.8 (10.3) | 22.9 (9.9) | N/A |
| 3 months | 16.7 (10.4) | 19.1 (10.9) | Estimated difference (95% CI) −2.24 (−4.82 to 0.33)Effect size −0.22 (−0.48 to 0.03) | |
| 6 months | 14.7 (10.7) | 17.0 (10.5) | Estimated difference (95% CI) −1.95 (−4.68 to 0.78)Effect size −0.19 (−0.46 to 0.08) | |
| 12 months | 14.8 (11.6) | 17.5 (11.1) | Estimated difference (95% CI) −2.41 (−5.35 to 0.53)Effect size −0.24 (−0.53 to 0.05) | |
| 18 months | 15.8 (11.5) | 16.9 (12.1) | Estimated difference (95% CI) −0.97 (−3.93 to 1.99)Effect size −0.10 (−0.39 to 0.20) | |
| Andrén 202238 | Baseline | 18.38 (7.08) | 18.73 (7.79) | N/A |
| 3 months | 7.68 (8.82) | 8.70 (8.10) | ITT linear quantile mixed model coefficient −0.26(−1.70 to 1.18), p = 0.72Effect size (95% CI) 0.05 (−0.34 to 0.44) | |
| 6 months | 6.85 (7.81) | 7.84 (8.97) | NR | |
| 12 months | 6.54 (8.14) | 6.14 (8.12) | Interaction between treatment and time coefficient 0.16 (95% CI −0.55 to 0.86), p = 0.67Effect size 0.03 (95% CI −0.14 to 0.20) | |
| Parent Tic Questionnaire | ||||
| Hollis 202134 | Baseline | 54.7 (29.9) | 53.1 (26.1) | N/A |
| 3 months | 34.7 (26.4) | 45.7 (25.5) | Estimated difference (95% CI) −9.44 (−15.37 to −3.51)Effect size −0.34 (−0.55 to −0.13)* | |
| 6 months | 31.1 (21.6) | 40.6 (24.3) | Estimated difference (95% CI) −8.60 (−14.43 to −2.77)Effect size −0.31 (−0.51 to −0.10)* | |
| 12 months | 30.7 (23.8) | 43.0 (25.3) | Estimated difference (95% CI) −9.89 (−16.01 to −3.77)Effect size −0.35 (−0.57 to −0.13)* | |
| 18 months | 28.1 (19.1) | 35.9 (25.6) | Estimated difference (95% CI) −2.15 (−8.83 to 4.53)Effect size −0.08 (−0.31 to 0.16) | |
| Andrén 202238 | Baseline | 34.33 (19.06) | 38.04 (23.27) | N/A |
| 3 months | 19.84 (17.92) | 23.51 (18.14) | ITT linear quantile mixed model coefficient 0.13 (95% CI −1.43 to 1.68)Effect size (95% CI) −0.01 (−0.22 to 0.19) p = 0.87 | |
| 6 months | 18.17 (16.18) | 24.18 (20.08) | NR | |
| 12 months | 16.76 (15.97) | 20.76 (17.04) | Interaction between treatment and time coefficient −0.10 (95% CI −1.63 to 1.44), p = 0.90, effect size 0.02 (95% CI −0.22 to 0.26) | |
| Clinical Global Impressions – Improvement Scale (CGI-I) | ||||
| Hollis 202134 | Baseline | NR | NR | N/A |
| 3 months | 2.96 (1.1) | 3.37 (1.1) | Estimated difference (95% CI) −0.41 (−0.71 to −0.11)Effect size −0.37 (−0.64 to −0.10)* | |
| 6 months | 2.8 (1.3) | 3.1 (1.1) | Estimated difference (95% CI) −0.31 (−0.66 to 0.03)Effect size −0.29 (−0.61 to 0.03) | |
| 12 months | 2.67 (1.09) | 3.07 (0.9) | Estimated difference (95% CI) −0.43 (−0.75 to −0.10)Effect size −0.43 (−0.74 to −0.12)* | |
| 18 months | 2.49 (1.36) | 2.86 (1.1) | Estimated difference (95% CI) −0.38 (−0.71 to −0.05)Effect size −0.35 (−0.66 to −0.04)* | |
| Clinical Global Impression Severity scale (CGI-S) | ||||
| Andrén 202238 | Baseline | 4.08 (0.74) | 4.19 (0.72) | N/A |
| 3 months | 3.24 (0.92) | 3.49 (0.90) | ITT linear quantile mixed model coefficient −0.36(95% CI −0.67 to −0.04), p = 0.03Effect size (95% CI) 0.71 (0.05 to 1.37)* | |
| 6 months | 3.10 (0.93) | 3.32 (1.11) | NR | |
| 12 months | 2.97 (0.96) | 3.25 (1.13) | Interaction between treatment and time coefficient 0.03 (95% CI −0.16 to 0.21), p = 0.28Effect size −0.11 (95% CI −0.80 to 0.58) | |
| CGAS | ||||
| Hollis 202134 | Baseline | 70.7 (13.7) | 72.1 (11.8) | N/A |
| 3 months | 75.9 (12.6) | 75.2 (12.6) | Estimated difference (95% CI) 0.96 (−1.48 to 3.41)Effect size 0.08 (−0.12 to 0.27) | |
| 6 months | 77.5 (14.7) | 76.8 (12.3) | Estimated difference (95% CI) 0.60 (−2.24 to 3.44)Effect size 0.05 (−0.17 to 0.27) | |
| 12 months | 77.4 (13.3) | 75.0 (12.9) | Estimated difference (95% CI) 2.85 (0.15 to 5.56)Effect size −0.22 (−0.43 to −0.01)* | |
| 18 months | 79.3 (13.5) | 77.3 (12.6) | Estimated difference (95% CI) 3.18 (0.47 to 5.90)Effect size −0.25 (−0.46 to −0.04)* | |
| Andrén 202238 | Baseline | 60.60 (6.59) | 60.71 (6.46) | N/A |
| 3 months | 66.83 (8.64) | 65.72 (7.44) | Intention to treat linear quantile mixed model coefficient 0.67 (95% CI −0.15 to 1.49), p = 0.11, effect size 0.12 (−0.05 to 0.29) | |
| Strengths and difficulties questionnaire | ||||
| Hollis 202134 | Baseline | 18.0 (6.5) | 16.3 (6.2) | N/A |
| 3 months | 14.7 (6.1) | 14.2 (6.3) | Estimated difference (95% CI) −0.38 (−1.62 to 0.85)Effect size −0.06 (−0.25 to 0.13) | |
| 6 months | 15.3 (6.2) | 13.3 (6.1) | Estimated difference (95% CI) 0.57 (−0.93 to 2.07)Effect size 0.09 (−0.15 to 0.32) | |
| 12 months | 14.4 (5.6) | 14.6 (6.4) | Estimated difference (95% CI) −0.86 (−2.31 to 0.58)Effect size −0.13 (−0.36 to 0.09) | |
| 18 months | 13.6 (6.1) | 13.8 (5.4) | Estimated difference (95% CI) −0.71 (−2.26 to 0.84)Effect size −0.11 (−0.35 to 0.13) | |
| Mood and feelings questionnaire | ||||
| Hollis 202134 | Baseline | 16.3 (11.3) | 15.9 (1.5) | N/A |
| 3 months | 10.7 (11.1) | 12.6 (11.1) | Estimated difference (95% CI) −1.36 (−3.75 to 1.02)Effect size −0.12 (−0.33 to 0.09) | |
| 6 months | 11.4 (12.1) | 11.4 (11.2) | Estimated difference (95% CI) −0.61 (−3.85 to 2.64)Effect size −0.05 (−0.34 to 0.23) | |
| 12 months | 11.4 (10.4) | 14.3 (11.6) | Estimated difference (95% CI) −2.93 (−5.77 to −0.09)Effect size −0.26 (−0.51 to −0.01)* | |
| 18 months | 10.9 (10.0) | 16.0 (14.6) | Estimated difference (95% CI) −4.87 (−8.00 to −1.75)Effect size −0.43 (−0.70 to −0.15)* | |
| Spence child anxiety scale | ||||
| Hollis 202134 | Baseline | 32.9 (20.0) | 30.5 (17.9) | N/A |
| 3 months | 27.2 (19.0) | 28.2 (18.3) | Estimated difference (95% CI) −2.80 (−6.52 to 0.93)Effect size −0.15 (−0.34 to 0.05) | |
| 6 months | 25.7 (19.6) | 25.9 (18.7) | Estimated difference (95% CI) −5.10 (−9.70 to −0.50)Effect size −0.27 (−0.51 to −0.03)* | |
| 12 months | 25.3 (17.1) | 29.9 (19.1) | Estimated difference (95% CI) −6.11 (−10.41 to −1.81)Effect size −0.31 (−0.53 to −0.08)* | |
| 18 months | 24.3 (18.6) | 32.6 (20.4) | Estimated difference (95% CI) −9.41 (−14.11 to −4.70)Effect size −0.49 (−0.74 to −0.25)* | |
| Study ID | Time point | Active stimulation, mean (SD) | Sham stimulation, mean (SD) | Waitlist, mean (SD) | Difference between groups at time point |
|---|---|---|---|---|---|
| YGTSS-Impairment | |||||
| Maiquez 202336 | Baseline | 25.5 (13.7) | 29.8 (13.5) | 30.1 (12.9) | N/A |
| 4 weeks | NR | NR | NR | Active vs. sham: Observed difference 1.07, effect size 0.06 (95% CI −0.39 to 0.52); p = 0.6Active vs. waitlist: Observed difference 1.86, effect size 0.11 (95% CI −0.35 to 0.58); p = 0.7 | |
| YGTSS-motor | |||||
| Maiquez 202336 | Baseline | 21.1 (3.2) | 20.4 (3.5) | 20.8 (3.1) | N/A |
| 4 weeks | NR | NR | NR | Active vs. sham: Observed difference −2.03, effect size −0.4 (95% CI −0.85 to 0.04); p = 0.04*Active vs. waitlist: Observed difference −2.15, effect size −0.45 (95% CI −0.88 to −0.01); p = 0.03* | |
| YGTSS-phonic | |||||
| Maiquez 202336 | Baseline | 19.0 (4.7) | 19.1 (4.7) | 18.1 (4.7) | N/A |
| 4 weeks | NR | NR | NR | Active vs. sham: Observed difference −3, effect size −0.42 (95% CI −0.87 to 0); p = 0.04*Active vs. waitlist: Observed difference −2.72, effect size −0.42 (95% CI −0.85 to 0.02); p = 0.04* | |
| Premonitory Urge for Tics Scale – Revised (PUTS-R) | |||||
| Maiquez 202336 | Baseline | 17.9 (8.8) | 19.3 (8.5) | 17.6 (8.6) | N/A |
| 4 weeks | NR | NR | Active vs. sham: Observed difference at 4w active vs. sham: −0.59, effect size −0.05 (95% CI −0.5 to 0.39); p = 0.4Active vs. waitlist: Observed difference −2.98, effect size −0.24 (95% CI −0.72 to 0.19); p = 0.14 | ||
Patient-reported outcomes
Health-related quality-of-life (HRQoL) outcomes in terms of Child and adolescent Gilles de la Tourette syndrome quality of life scale (C&A-GTS-QOL) scores reported by the two ORBIT trials are presented in Table 8 and those from the Neupulse study are presented in Table 9.
| Study ID | Time point | ORBIT, mean (SD) | Psychoeducation, mean (SD) | Difference between groups at time point |
|---|---|---|---|---|
| Hollis 202134 | Baseline | 36.6 (16.4) | 35.0 (17.2) | N/A |
| 3 months | 25.7 (18.0) | 31.8 (17.7) | Estimated difference (95% CI) −4.81(−8.79 to −0.83)Effect size −0.29(−0.52 to −0.05)* | |
| 6 months | 27.4 (16.5) | 28.9 (18.3) | Estimated difference (95% CI) −2.91(−7.60 to 1.78)Effect size −0.17(−0.45 to 0.11) | |
| 12 months | 25.5 (16.8) | 32.2 (16.8) | Estimated difference (95% CI) −5.79(−10.28 to −1.30)Effect size −0.34(−0.61 to −0.08)* | |
| 18 months | 26.0 (16.6) | 36.8 (21.1) | Estimated difference (95% CI) −9.00(−13.98 to −4.01)Effect size −0.53(−0.83 to −0.24)* | |
| Andrén 202238 | Baseline | 29.1 (15.1) | 30.5 (16.5) | N/A |
| 3 months | 19.8 (16.3) | 20.1 (15.7) | Coefficient 0.46(95% CI −1.63 to 2.55), effect size −0.04 (−0.24 to 0.16), p = 0.67 | |
| 6 months | 18.3 (15.2) | 21.2 (16.7) | NR | |
| 12 months | 20.7 (17.5) | 20.0 (15.2) | Coefficient 0.18(95% CI −0.98 to 1.33), effect size −0.04 (−0.31 to 0.23), p = 0.77 |
| Study ID | Time point | Neupulse, mean (SD) | Sham stimulation, Mean (SD) | Waitlist, mean (SD) |
|---|---|---|---|---|
| Maiquez 202336 | Baseline | 54.9 (24.6) | 52.8 (22.4) | 56.7 (24.1) |
| 3 months | 46.7 (25.2) | 40.3 (23.6) | 52.1 (24.0) | |
| Difference within group at time point | t = 2.64, p < 0.015(baseline vs. 3 months) | t = 3.9, p < 0.0005(baseline vs. 3 months) | t = 1.6, p = 0.13(baseline vs. 4 weeks) | |
Intermediate outcomes
Adverse events
The two ORBIT studies reported AEs. AEs were not reported in the Neupulse study. In the UK study, 78.6% of the intervention group and 84.8% of the comparator group experienced an AE.34 It was unclear if they were related to the relevant intervention. In addition, two participants of the comparator group experienced a serious AE (one collapse and one tic attack), both of which were deemed unrelated to trial participation. In the Swedish study, 44 treatment-related AEs were reported: depressed mood (n = 11), irritability (n = 8), anxiety/worry/stress (n = 8), conflicts with family/peers (n = 5), increased tics (n = 4), pain (n = 3), tiredness/fatigue/drowsiness (n = 2), increased isolation (n = 1), restless (n = 1), other (n = 1).38 Twenty-one AEs were reported in the comparator group: depressed mood (n = 7), anxiety/worry/stress (n = 4), increased tics (n = 2), irritability (n = 2), other (n = 2), bullying (n = 1), conflicts with family/peers (n = 1), physical harm/injury (n = 1), hopelessness (n = 1). One serious AE unrelated to treatment was reported in the comparator group (meningitis requiring hospitalisation).
Evidence gap: Reporting of safety data was limited in the ORBIT trials, particularly with respect to long-term AEs. Additionally, the Neupulse study did not report any AEs. Collecting long-term safety data would be valuable to better assess the overall risk profile of these interventions.
Treatment satisfaction/engagement
The two ORBIT studies reported indicators of treatment satisfaction or engagement. In the intervention arm of one study, the median number of logins by the young person was 19 as compared to 9 in the comparator group.34 The median score for the young person’s perception of treatment suitability and credibility was 7 in the intervention group and 6 in the comparator group. Mean child satisfaction scores with the ORBIT intervention were 24.8 out of 32, with higher scores indicating higher levels of satisfaction. In the other study, both children and parents were more satisfied with the ORBIT intervention than the comparator treatment (children: coefficient 3, 95% CI 1.13 to 4.87, p = 0.003; parents: coefficient 4, 95% CI 2.33 to 5.66, p < 0.001).38 The treatment credibility score was identical for children and parents (coefficient 1, 95% CI 0.38 to 1.62, p = 0.002).
Intervention adherence
In the study by Andrén 2022, treatment adherence, as measured by the internet intervention Patient Adherence Scale, showed no differences between the groups [mid + post-treatment summarised to one score: quantile regression 1 (95% CI −1.28 to 3.28, p = 0.39)].38 The Neupulse study did not report adherence data.36
Rates and reasons for attrition
In the intervention arm of the UK ORBIT study, 23/112 (20.5%) of participants were lost to follow-up at 18 months (13 withdrew, 10 were uncontactable).34 In the comparator group, 22/112 (19.6%) were lost to follow-up at 18 months (10 withdrew, 12 were uncontactable). In the Swedish ORBIT study, at the 3-month follow-up, three in the intervention arm and two in the comparator arm were lost to follow-up; no reasons were reported.38 In the Neupulse study, a total of ten participants withdrew during the initial training as they found the stimulation uncomfortable.36 Reasons for other withdrawals from the study were not requested for ethical reasons but some were volunteered, such as insufficient time to complete the study or because they were planning to be away during the period of the trial and could not commit to participation (personal communication with Stephen Jackson from Neupulse).
Intervention completion
In the two ORBIT studies, treatment completion was defined as completion of at least the first four child chapters. As such, in the UK study, 88.4% of the intervention group and 93.8% of the comparator group were classed as treatment completers.34 In the Swedish study, 100% of the intervention group and 94.6% of the comparator group were treatment completers.38
Evidence gap: although there were statistically significant improvements in YGTSS-TTSS scores over time in the intervention groups as compared to the control groups, changes in the reported secondary outcomes and QoL scores were less consistent. In particular, the YGTSS-Impairment scores did not show any improvement. Therefore, it is unclear if improvements in tic severity scores translate into improvements in people’s daily lives.
Evidence gap: published evidence was not available for all outcomes specified in the scope (Table 10), namely, behavioural outcomes, functional outcomes, suicidal thoughts and behaviour, and information about digital technology. Clinical measures were reported by all three included studies, albeit using various outcomes and time points, precluding a more extensive pooling of outcomes.
| Outcome specified in scope | Study/ies in which reported |
|---|---|
| Intermediate outcomes | |
| Intervention-related AEs | ORBIT UK,34 ORBIT Sweden38 |
| Treatment satisfaction and engagement | ORBIT UK,34 ORBIT Sweden38 |
| Intervention adherence, rates of attrition and completion | ORBIT Sweden38ORBIT UK,34 ORBIT Sweden,38 Neupulse36ORBIT UK,34 ORBIT Sweden38 |
| Clinical outcome measures | |
| Measures of symptom severity (self, parental or practitioner reported) | ORBIT UK,34 ORBIT Sweden,38 Neupulse36 |
| Social, behavioural, and functional outcomes | ORBIT UK34Not reportedNot reported |
| Suicidal thoughts and behaviour | Not reported |
| Patient-reported outcome measures | |
| HRQoL | ORBIT UK,34 ORBIT Sweden,38 Neupulse36 |
| Patient’s experience and patient’s satisfaction | ORBIT UK,34 ORBIT Sweden38 |
| Rates and reasons for attrition | ORBIT UK,34 ORBIT Sweden,38 Neupulse36 |
| Digital technology information | Not reported |
Relevant subgroups
There were insufficient data available to conduct statistical analyses of the specified subgroups. The UK ORBIT study conducted an unplanned post hoc analysis of the effect of the intervention on the primary outcome by common comorbidities (anxiety disorder, ADHD). The number of participants with ADHD in each intervention group was 22 while the number of participants without ADHD in the ERP group and psychoeducation group were 79 and 78, respectively. The number of participants with anxiety disorder in the ERP group and psychoeducation group was 30 and 23, respectively, and that without anxiety disorder was 71 and 77, respectively. There was no difference in the effect of the intervention on participants with or without ADHD (p = 0.906) or anxiety disorder (p = 0.204).
The Swedish ORBIT study conducted post hoc analyses on the impact of age and sex on the primary outcome. The median age of 11 years was used to split the sample; in the 9–11 years age group, there were 124 participants and there were 97 participants in the 12–17 years group. There was an interaction effect between group (ERP and comparator) and time (baseline to 3-month follow-up) on the older group [coefficient (95% CI) −1.21 (−2.14 to −0.02); p = 0.05] but not the younger group. More participants in the older group responded to ERP (n = 26; 51%) than the comparator (n = 10; 23%). The equivalent analysis in the younger group showed no difference between the groups. The sample was also split by gender, with 152 boys and 68 girls (and one participant not included in the analysis due to identifying as non-binary). There was a significant interaction effect for boys [coefficient (95% CI) −1.06 (−2.09 to −0.03); p = 0.04] but not for girls. Significantly more boys responded to ERP (n = 35; 51%) than the comparator (n = 21; 26%) at the 3-month follow-up [OR 2.94 (95% CI 1.48 to 5.84); p = 0.002]. The equivalent analysis among girls showed no between-group difference.
Meta-analysis: Yale Global Tic Severity Scale – Total Tic Severity Score
The two ORBIT trials at low risk of bias reported YGTSS-TTSS scores at 3 and 12 months and were combined in a meta-analysis.34,38 The Neupulse study was not included in the meta-analyses as it is a different type of technology than ORBIT. Furthermore, the Neupulse study used sham stimulation and waitlist as comparator interventions, whereas the ORBIT studies used psychoeducation as the comparator intervention. The Swedish ORBIT study did not report comparison data at 6 months and, therefore, a meta-analysis at that time point was not possible.37 Figures 4 and 5 show that the TTSS scores of the YGTSS were significantly lower for the ERP group as compared to the psychoeducation group at 3 months (mean difference −2.12, 95% CI −3.52 to −0.73, p = 0.0003) and 12 months (mean difference −2.45, 95% CI −4.05 to −0.85, p = 0.0003). In both meta-analyses, there was no evidence of heterogeneity between studies (I2 = 0%).
FIGURE 4 Meta-analysis for YGTSS-TTSS at 3 months. (Figure not reproduced here; see the original article.)
FIGURE 5 Meta-analysis for YGTSS-TTSS at 12 months. (Figure not reproduced here; see the original article.)
Meta-analysis: Yale Global Tic Severity Scale-Impairment
The mean difference in the impairment score of the YGTSS at 3 and 12 months were pooled from the same two ORBIT trials. The YGTSS-Impairment score indicates whether the presence of tic disorders had a negative impact on people’s personal life. The YGTSS-Impairment summary score was lower for the ERP group as compared to the psychoeducation group at 3 months (mean difference −1.53, 95% CI −3.32 to 0.26, p = 0.09) and 12 months (mean difference −0.90, 95% CI −3.90 to 2.10, p = 0.56), but the differences were not statistically significant (Figures 6 and 7). There was no evidence of heterogeneity between studies in the 3-month meta-analysis (I2 = 0%) and a moderate degree of heterogeneity (I2 = 57%) in the 12-month meta-analysis. The Swedish ORBIT study did not report comparison data at 6 months, and, therefore, a meta-analysis at that time point was not possible.37
FIGURE 6 Meta-analysis for YGTSS-Impairment at 3 months. (Figure not reproduced here; see the original article.)
FIGURE 7 Meta-analysis for YGTSS-Impairment at 12 months. (Figure not reproduced here; see the original article.)
Summary of clinical effectiveness section
A total of three RCTs investigating two types of digitally enabled therapy, namely ORBIT and Neupulse, for CTDs and TS in children, young people and adults were included in the review of clinical effectiveness. Meta-analysis results of the two ORBIT studies showed that tic severity in terms of YGTSS-TTSS scores was significantly lower in children and young people in the ERP groups than those in the psychoeducation groups at 3 months and 12 months. However, no significant improvements in tic-related impairment and distress measured using the impairment score of the YGTSS were observed between treatment groups. In each ORBIT study, secondary clinical outcome measures such as the CGAS, the Clinical Global Impression Improvement (CGI-I), and the Clinical Global Impression Severity (CGI-S) did not show a consistently greater response in the ERP group compared to the psychoeducation group at all assessed time points. In the UK ORBIT study, CGI-I showed a greater response in the ERP group at 3, 12 and 18 months but not at 6 months. In the Swedish ORBIT study, CGI-S showed a difference in favour of the ERP group at 3 months, but not at 6 months. In both studies, the CGAS showed no differences between intervention groups at 3 months and a positive difference in favour of the ERP group at 12 and 18 months in the UK ORBIT study. The estimated mean difference in the Parent Tic Questionnaire (PTQ) favoured the ERP group in the UK ORBIT study at 3, 6 and 12 months but not at 18 months and was not significant at 3 and 12 months in the Swedish ORBIT study. Similarly, in the UK ORBIT study, other measures evaluating anxiety and mood, emotional, and behavioural functioning did not show a consistent pattern of response at all assessed time points. It is worth noting that in Phase 2 of the UK ORBIT study (the naturalistic follow-up beyond 6 months), a substantial proportion of data were missing for most secondary outcome measures – including the PTQ, Mood and Feeling Questionnaire, Spence Child Anxiety Scale, Strengths and Difficulties Questionnaire and GTS-QoL at both 12- and 18-month follow-ups. Missing data rates ranged from 41% to 61%, significantly limiting the reliability and interpretability of these findings. Notably, the YGTSS-TTSS and YGTSS-Impairment scales had considerably lower rates of missing data (19–21%).
The Neupulse study reported significantly lower YGTSS-TTSS scores at 4 weeks in the active stimulation group compared to the sham stimulation group. Greater reductions in YGTSS motor and phonic scores were also observed among participants receiving active stimulation than among those receiving sham stimulation. No significant differences between treatment groups were observed for the YGTSS-Impairment score or the Premonitory Urge for Tics Scale – Revised. In general, participants’ engagement with the interventions was reported to be good.