Telephone cognitive-behavioral therapy for adolescents with obsessive-compulsive disorder: a randomized controlled non-inferiority trial.

OBJECTIVE: Many adolescents with obsessive-compulsive disorder (OCD) do not have access to evidence-based treatment. A randomized controlled non-inferiority trial was conducted in a specialist OCD clinic to evaluate the effectiveness of telephone cognitive-behavioral therapy (TCBT) for adolescents with OCD compared to standard clinic-based, face-to-face CBT.
METHOD: Seventy-two adolescents, aged 11 through 18 years with primary OCD, and their parents were randomized to receive specialist TCBT or CBT. The intervention provided differed only in the method of treatment delivery. All participants received up to 14 sessions of CBT, incorporating exposure with response prevention (E/RP), provided by experienced therapists. The primary outcome measure was the Children's Yale-Brown Obsessive-Compulsive Scale (CY-BOCS). Blind assessor ratings were obtained at midtreatment, posttreatment, 3-month, 6-month, and 12-month follow-up.
RESULTS: Intent-to-treat analyses indicated that TCBT was not inferior to face-to-face CBT at posttreatment, 3-month, and 6-month follow-up. At 12-month follow-up, there were no significant between-group differences on the CY-BOCS, but the confidence intervals exceeded the non-inferiority threshold. All secondary measures confirmed non-inferiority at all assessment points. Improvements made during treatment were maintained through to 12-month follow-up. Participants in each condition reported high levels of satisfaction with the intervention received.
CONCLUSION: TCBT is an effective treatment and is not inferior to standard clinic-based CBT, at least in the midterm. This approach provides a means of making a specialized treatment more accessible to many adolescents with OCD. Clinical trial registration information-Evaluation of telephone-administered cognitive-behaviour therapy (CBT) for young people with obsessive-compulsive disorder (OCD); http://www.controlled-trials.com; ISRCTN27070832. Crown

RCT Entities:   Population Interventions Outcomes

Obsessive-compulsive disorder (OCD) in children and adolescents is a chronic disorder that can cause functional impairment across multiple life domains. Estimates suggest that approximately 1 in 100 young persons suffers from OCD. Cognitive-behavioral therapy (CBT) incorporating exposure and response prevention (E/RP) is the recommended psychological treatment for pediatric OCD. However, CBT is not readily available to all who need it because of a variety of factors, and geographical and financial barriers prevent many from receiving treatment.

Telehealth is an area of mental health practice that offers significant potential for improving access to specialized treatments. Telehealth involves the use of telecommunication tools (e.g., telephone, Internet, video-conferencing) as a means for health professionals to provide treatment remotely. Telehealth applications of psychological treatments have grown in popularity, as they have shown effectiveness but also offer additional benefits such as reduced time and cost. A number of recent controlled trials demonstrate efficacy of telehealth interventions for adult mental health disorders; examples include social phobia, depression, and OCD. Increasingly, there is evidence of successful piloting telehealth interventions for childhood disorders, including a pilot study of webcam-delivered CBT for adolescent OCD, tic disorders, and depression. Telehealth treatment typically entails the same components as conventional face-to-face CBT but is simply delivered remotely, via a device.

Within the variety of telehealth methods available, telephone CBT (TCBT) has some important advantages for service providers and users, including relative ease of administration. There is no need for hi-technology equipment, as most clinics have access to telephones and telephone ports, and many service users have a telephone or access to one. For these reasons, plus evidence suggesting that relatively few adolescents with OCD are able to access CBT despite it being the recommended treatment, TCBT was piloted for success and feasibility with adolescents with OCD in a London-based specialist clinic. Results indicated that TCBT could successfully reduce symptoms of OCD and was regarded positively by service users.

The aim of the present study was to determine whether TCBT was as effective as face-to-face CBT for adolescents with OCD. As face-to-face CBT for OCD is a well-established treatment, it is appropriate to use it as a benchmark against which to compare TCBT and to demonstrate non-inferiority. It was hypothesized that:

TCBT would not be inferior to traditional face-to-face CBT in reducing OCD symptoms as measured by the Children’s Yale–Brown Obsessive-Compulsive Scale (CY-BOCS).

TCBT would not be inferior on secondary outcome measures of depressive symptoms, self-report, and parent-report of adolescent OCD symptoms, overall psychological health, global functioning, and parental mental health symptoms.

The changes observed at the end of the treatment period would be maintained over a 12-month follow-up period.

Method

Recruitment and Inclusion Criteria

Participants were recruited by referral from primary care general practitioners and from mental health professionals within secondary and tertiary care settings within the National Health Service (NHS) to a specialist OCD clinic between 2008 and 2011. Information about the study was conveyed by word of mouth, letters to referring agencies, advertisements published on Web pages of national OCD charities within the UK, and by a research support organisation within the NHS (the Mental Health Research Network). The study protocol was approved by the Joint South London and Maudsley/Institute of Psychiatry Research Ethics Committee (08/H0807/12). Written informed consent was obtained from all parents and participants more than 16 years of age, and informed assent from participants less than 16 years of age, after a detailed description of the study had been given. The trial was registered on the International Standard Randomized Controlled Trial Number Register (ISRCTN27070832).

Inclusion criteria were as follows: having primary OCD according to DSM-IV criteria; having a CY-BOCS score of 16 or greater, indicating moderate to severe impairment; being between the ages of 11 and 18 years; being medication-free or on a stable dose of medication for a period of 12 weeks or more; having no suicidal intent, drug or alcohol abuse, or psychotic symptoms; having no learning disability or pervasive developmental disability; needing and wanting CBT, and agreeable to randomisation; and being agreeable to parental involvement in treatment. Exclusion criteria were the following: current diagnosis of psychosis or current alcohol or substance abuse/dependence; English comprehension too poor to engage in treatment; severe disabling neurological disorder; diagnosed global learning disability or pervasive developmental delay; and characteristics interfering with completion of treatment within trial (e.g., a life-threatening or unstable medical illness).

Measures

Primary Outcome Measure

The primary outcome measure was the clinician-administrated CY-BOCS, which was administered by a blinded rater at all time points.

Secondary Outcome Measures

The Children’s Obsessional Compulsive Inventory–Revised (ChOCI-R) includes child- and parent-rated versions and is aimed at capturing subjective measures of OCD symptom severity and impairment.

The Beck Depression Inventory for Youth (BDI-Y) assesses symptoms of depression in adolescents. Psychometric properties are sound.

The Strengths and Difficulties Questionnaire (SDQ) assesses child mental health symptoms more broadly. We used the total score to provide a measure of general psychological functioning from both a child-rated and parent-rated perspective.

Parental mental health was assessed using the total score of the 42-item version of the Depression, Anxiety and Stress Scales (DASS). The DASS has demonstrated strong psychometric properties in both clinical and community samples.

Diagnostic Assessment, Global Functioning, and Improvement

Blinded raters completed all diagnostic assessments and ratings of global functioning and improvement. Psychiatric diagnoses were established using the Anxiety Disorders Interview Schedule for Children (ADIS-IV-C/P). The Children’s Global Impression Scale (CGAS) is a numeric scale ranging from 1 (needs constant supervision: 24-hour care) to 100 (superior functioning in all areas) and was used to rate the participant’s overall general functioning. The Clinical Global Impression–Improvement (CGI-I) rating scale assessed how much the participant’s symptoms had improved or worsened relative to baseline. Participants were rated on a 7-point scale ranging from 1 (very much improved) to 7 (very much worse).

Assessment of Treatment Credibility, Expectancy, and Satisfaction With Treatment Received

A treatment credibility and expectancy scale (modified from the measure presented by Devilly and Borkovec to reflect use by adolescents with OCD) was used at 2 time points to determine whether participants and their parents viewed the treatments as equally credible. A treatment satisfaction questionnaire was developed to assess participant and parent satisfaction with the intervention; this was given to families posttreatment. Participant and parent satisfaction was also assessed using a qualitative methodology, which is reported separately (Lang et al., in preparation, 2014).

Procedure

Once referral to the specialist clinic was received, young persons and parents were asked to complete the ChOCI, SDQ, BDI (youth only), and the DASS (parents only). A face-to-face assessment was arranged, potential participants completed the CY-BOCS, and a clinical interview with parents was used by experienced psychiatrists/clinical psychologists to establish whether inclusion criteria were met. The study was offered, and information was provided. Families were contacted by telephone 1 week later, and if they were agreeable to the study, a second clinic appointment was arranged approximately 8 weeks after the initial meeting. At this appointment, participants and their parents completed the ADIS-IV-C/P and the CY-BOCS in structured interviews, and self- and parent-report measures were completed again. Families were included in the study if participants remained symptomatic and full inclusion criteria were met. Randomization occurred at this time. After randomization, an appointment was made with the treating therapist, and treatment sessions commenced the following week. All treating therapists were experienced clinical psychologists, and all blinded raters completing assessments were either assistant psychologists or clinical psychologists, all of whom had specialist training and experience in assessing and treating childhood OCD.

Randomization and Assessment Points

Participants were randomly allocated to CBT or TCBT in a 1:1 ratio using a computer-generated randomization sequence, prepared before the study commenced, to which only the principal investigator (C.T.) had access. No other study personnel had access to the allocation; the allocation sequence was kept separate from all trial materials and personnel. Trial enrollment occurred at the time of the second assessment. The integrity of the randomization table was maintained and allocation concealment was ensured, with allocation provided only to treating therapists and allocated participants. There were no instances of delayed or prevented enrollment. There were no restrictions or matching.

A repeated-measures design was used, and assessments were conducted immediately before treatment (i.e., baseline), mid-treatment (i.e., session 7), immediately after treatment (i.e., posttreatment), and at follow-up points scheduled at 3 months, 6 months, and 12 months posttreatment. At each time point (except session 7), participants completed the CY-BOCS, ChOCI, SDQ, and BDI-Y, and parents completed the ChOCI, SDQ and DASS. At session 7, all participants in both groups were invited to the clinic for a blinded face-to-face CY-BOCS assessment; the results of this assessment were fed back to treating therapists to assist in planning future sessions. Treatment credibility and expectancy and a measure of therapeutic alliance were completed by participants and parents after session 2 and session 7.

Masking Procedures

To ensure that blindness was maintained, treating therapists and participants/parents were instructed not to discuss the treatment that they had received with the blinded assessors, and reminder cards were placed in the interview rooms. After completing the assessments, blinded raters noted whether blindness had been broken (for example, if the patient had inadvertently revealed his/her treatment group) and, if not, made a guess regarding the patient’s group allocation. A research assistant managed the allocation of blinded assessors and monitored blindness rating forms. Where blindness was inadvertently broken, outcome assessors were changed.

Monitoring of Treatment Integrity

Treatment integrity was ensured in 3 ways. First, a detailed treatment manual was used (Turner, 2006, unpublished), and, regardless of condition, participants received a workbook with clinical information and monitoring sheets. Second, all treating therapists received supervision by more senior clinical psychologists who were specialists in CBT for OCD. Finally, therapists were asked to audio-record all sessions (whenever possible) and to complete an integrity-to-protocol rating form developed for this study. A random sample of 225 recorded sessions (25%) were then audited and independently rated for integrity to protocol by a doctorate-level trainee in clinical psychology. The rate of adherence to the manual was 94% as measured by therapist report and was 93% as measured by independent audit. This is considered a high rate of treatment fidelity. There were no differences in adherence ratings between conditions.

Treatment

Treatment consisted of 14 sessions of CBT delivered by 6 experienced clinical psychologists. Treatment was identical within conditions except that participants randomized to TCBT received all treatment sessions via telephone. Sessions 1 and 2 consisted of psychoeducation, sessions 3 to 12 consisted of gradual exposure with response prevention (E/RP) and incorporated various cognitive strategies as appropriate, and sessions 13 and 14 consisted of relapse prevention and ongoing symptom management (if required). The treatment protocol incorporated 10 minutes of parental discussion at the end of each treatment session. Homework E/RP tasks were assigned between sessions, and participants were encouraged to complete daily E/RP. The treatment protocol has been validated in previous trials. All 14 sessions were required to be completed within 17 weeks, allowing accommodation for illness, missed appointments, and holidays.

Power Analysis

A noninferiority margin of 5 points on the CY-BOCS was selected as the acceptable mean difference between conditions, based on clinical judgment and the adult trial of TCBT conducted by Lovell et al. We estimated that with a sample size of 33 in each condition, a 2-group 0.05 1-sided test would have 80% power to reject the null hypothesis that TCBT is inferior to CBT (i.e., the difference in means is 5.0 or farther from 0 in the same direction) in favor of the alternative hypothesis that TCBT is not inferior, assuming that the expected difference in means is 0.0 and the common standard deviation is 8.0. To allow for drop-outs, we sought to recruit 72 participants (n = 36 in each condition).

Statistical Analysis

Little’s Missing Completely at Random (MCAR) test was used to establish that missing data was missing at random. Missing values on all outcome measures for all follow-up occasions were imputed via Multiple Imputation using Chained Equations (MICE) in Stata 11, with the ice command. Ten imputed data sets were created for the CY-BOCS, and up to 100 imputed data sets were created for secondary outcome measures depending on the amount of missing data. The missing data at each occasion were imputed on the basis of observed values of each measure at all other time points. Because many of the measures were non-normally distributed, missing values were imputed through predictive mean matching, which replaces missing values with observed values of similar cases. This avoids the potential for imputed values, which fall outside the observed range of the variable.

Data on outcome measures at each time point after baseline were modeled using linear regression techniques with treatment condition as the key covariate. All models were also estimated separately, excluding cases with imputed values of the outcome variable as suggested by von Hippel. However, as this did not substantively alter any of the conclusions, the results reported here use the imputed outcome values.

Non-inferiority is established when the confidence interval for the difference between treatment conditions excludes a prespecified margin of inferiority, typically chosen to represent a clinically significant difference. As noted above, the margin of inferiority for the primary outcome was set at 5 points on the CY-BOCS scale using a 95% confidence interval. TCBT would be deemed non-inferior to CBT if the upper limit of the 95% confidence interval for the difference between TCBT and CBT is less than 5, meaning that we are 95% confident that the “true” value of the difference is not worse than 5 points on the CY-BOCS scale. For secondary outcome measures, the margin of inferiority was set for each measure at a difference of 1 standard deviation from the CBT condition mean for each assessment point. All primary and secondary outcomes are reported using an intent-to-treat sample of N = 72. Last available observation data were used to determine the number of participants who could be classified as responders (i.e., CY-BOCS reduction of 35% or more) or remitters (i.e., CY-BOCS ≤12).

Results

Participant Characteristics

Participants were 72 adolescents aged 11 through 18 years and their parents. All participants were outpatients with a primary diagnosis of OCD. Eleven young persons (15%) had a first-degree relative diagnosed with OCD. A total of 21 adolescents (29%) had received a previous trial of CBT for OCD. Comorbid diagnoses included another anxiety disorder (n = 39, 54.16%), depression or dysthymia (n = 7, 9.7%), a tic disorder or Tourette syndrome (n = 6, 8.3%), attention-deficit/hyperactivity disorder (ADHD; n = 1, 1.4%), oppositional-defiant disorder (ODD; n = 1, 1.4%), body dysmorphic disorder (BDD; n = 1, 1.4%), and an eating disorder (n = 1, 1.5%). Table 1 presents additional demographic information and pretreatment clinical information; there were no significant differences between groups. Participants in each condition received on average 12 of the 14 CBT sessions offered.

Table 1

Pretreatment Clinical and Demographic Variables for Participants With Each Condition

Variable	CBT (n = 36)		TCBT (n = 36)
Age, y	14.50	(2.19)	14.19	(2.07)
Age at OCD Onset	10.97	(3.12)	11.00	(3.61)
Female Gender, %	47.2		44.4
Male Gender, %	52.8		55.6
On stable SSRI Medication, %	22.6		17.9
Previous CBT Treatment, n (%)	8	(22.2)	13	(36.1)
CY-BOCS Obsessions	11.67	(2.30)	12.36	(2.20)
CY-BOCS Compulsions	12.44	(2.05)	13.28	(2.02)
CY-BOCS Total	24.11	(4.02)	25.64	(3.86)
CHOCI-C Total	23.21	(8.70)	24.11	(7.01)
CHOCI-P Total	22.65	(8.81)	24.53	(9.12)
BDI-Y	14.44	(8.77)	14.58	(8.73)
FAS-M Total	20.33	(12.75)	21.45	(11.32)
FAS-F Total	13.36	(7.48)	13.18	(7.82)
DASS-M Total	17.68	(18.63)	15.53	(14.04)
DASS-F Total	15.33	(17.11)	8.56	(5.50)
SDQ-P Total	17.21	(3.86)	16.26	(4.37)
SDQ-C Total	16.63	(3.37)	17.13	(3.71)
CGAS	50.53	(7.06)	50.00	(7.49)

Note: Values represent mean and SD unless otherwise specified. BDI-Y = Beck Depression Inventory for Youth; CBT = cognitive-behavioral therapy; CGAS = Clinical Global Assessment Scale; CHOCI-C = Children’s Obsessional Compulsive Inventory Revised–Child; CHOCI-P = Children’s Obsessional Compulsive Inventory Revised–Parent; CY-BOCS = Children’s Yale–Brown Obsessive Compulsive Scale; DASS-F = Depression Anxiety Stress Scale–Father; DASS-M = Depression Anxiety Stress Scale–Mother; FAS-F = Family Accommodation Scale–Father; FAS-M = Family Accommodation Scale–Mother; OCD = obsessive-compulsive disorder; SDQ-C = Strengths and Difficulties Questionnaire–Child Version; SDQ-P = Strengths and Difficulties Questionnaire–Parent Version; SSRI = selective serotonin reuptake inhibitor; TCBT = telephone cognitive-behavioral therapy.

Attrition

Figure 1 shows the participant flow throughout the trial.

Figure 1

Consolidated Standards of Reporting Trials (CONSORT) diagram. Note: CBT = cognitive-behavioral therapy; ITT = intention to treat; TCBT = telephone cognitive-behavioral therapy.

Primary Outcome

Figure 2 presents the 95% confidence interval for the CY-BOCS mean difference between CBT and TCBT groups, and Figure 3 presents the mean scores (with standard error) for both conditions across time. For all assessment points through to the 6-month follow-up, the difference between conditions was nonsignificant, and the 95% confidence interval lay below the 5-point difference margin, indicating that TCBT was not inferior to CBT. For the 12-month follow-up point, the 95% confidence interval included the margin of difference, and, based on the guidance provided by Piaggio et al. for the Consolidated Standards of Reporting Trials (CONSORT) group, we conclude that the difference was nonsignificant; however, non-inferiority of TCBT could not conclusively be demonstrated at this time point.

Figure 2

Ninety-five percent CI for the difference in Children’s Yale-Brown Obsessive Compulsive Scale (CY-BOCS) between telephone cognitive-behavioral therapy (TCBT) and cognitive-behavioral therapy (CBT) conditions.

Figure 3

Total Children’s Yale–Brown Obsessive Compulsive Scale (CY-BOCS) mean scores (with standard error) for the cognitive-behavioral therapy (CBT) and telephone cognitive-behavioral therapy (TCBT) conditions. Note: M-FU = months of follow-up; post-tx = posttreatment.

As shown in Table 2, there were no group differences in the proportion of the sample who could be defined as treatment responders (defined as participants with a ≥35% reduction on the CY-BOCS) and remitters (defined as participants with CY-BOCS scores ≤12).

Table 2

Cognitive-Behavioral Therapy (CBT) Sessions, and Percentage of Sample Classified as CBT Respondersa and Remitters,b Based on Last Available Observation

	CBT	TCBT	χ2	Df	P
Responders, %
Posttreatment	90.6	87.5	0.160	1	.689
3-mo follow-up	92.0	89.7	0.088	1	.767
6-mo follow-up	84.0	92.0	0.758	1	.384
12-mo follow-up	94.4	88.9	0.364	1	.546
Remitters, %
Posttreatment	60.6	58.8	0.022	1	.882
3-mo follow-up	51.7	58.1	0.243	1	.622
6-mo follow-up	53.8	77.8	3.38	1	.066
12-mo follow-up	83.3	77.8	0.177	1	.674

Note: Mo = month; TCBT = telephone cognitive-behavioral therapy.

CBT responders ≥35% Children’s Yale–Brown Obsessive Compulsive Scale (CY-BOCS) reduction.

Remitters ≤12 on CY-BOCS.

Secondary Outcomes

Table S1 (available online) presents the means, standard deviations, and 95% confidence interval data for all outcome measures. Based on the predetermined criteria of 1 standard deviation difference between the groups, analysis shows that the upper bound of the confidence intervals falls within the non-inferiority margin, confirming the non-inferiority hypothesis on all measures. For the CGAS, because higher scores reflect a greater level of overall functioning, the lower-bound confidence interval is the critical value, again confirming non-inferiority.

Clinician-Administrated Measure of Global Improvement and Functioning

There were no group differences in clinician-rated global functioning using the CGAS with non-inferiority of TCBT demonstrated, and no group differences with regard to global improvement using the CGI-I scale. Participants in both conditions improved over the course of the study, and improvements were maintained over the follow-up period (Table S1, available online).

Treatment Credibility and Expectancy

Credibility ratings were provided at the end of session 2. There were no significant differences in treatment credibility ratings provided by adolescents in the TCBT group (mean = 23.26, SD = 5.62) compared with those in the face-to-face group (mean = 22.12, SD = 4.73); t41 = 0.48, p = .31. Similarly, there were no significant differences in credibility ratings provided by parents in the TCBT group (mean = 23.61, SD = 4.64) compared with those in the face-to-face group (mean = 23.08, SD = 5.41); t39 = 0.74, p = .96.

Satisfaction With Treatment Received

Adolescents reported high levels of satisfaction, with 94.4% of youth reporting satisfaction with the help received, and no significant differences between groups (TCBT = 96.3%, CBT = 92.6%, χ2[1, n = 54] = 0.35, p = .552). The majority of adolescents (59.3%) reported that they were very happy with the treatment condition to which they were allocated. There was a significant difference between groups, with 77.8% of adolescents in the TCBT condition being very happy with their group allocation compared to 40.7% of adolescents in the CBT condition: χ2(3, n = 54) = 8.89, p = .031. Parents similarly reported high levels of satisfaction, with no significant difference between treatment groups (TCBT = 93.1%, CBT = 88.9%, χ2[3, n = 56] = 2.11, p = .349).

Assessment of Blinding

Treatment condition was inadvertently revealed to blinded assessors in a small number of cases at each time point as follows: 2 patients at session 7; 3 patients at session 14; 1 patient at 3-month follow-up; 2 patients at 6-month follow-up; and 4 patients at 12-month follow-up. Excluding those, the blind raters’ guesses regarding treatment condition were no better than chance at session 7 (κ = 0.259, p = .061), session 14 (κ = −0.003, p = .984), 3-month follow-up (κ= 0.005, p = .967), 6-month follow-up (κ = −0.164, p = .227), or 12-month follow-up (κ = −0.064, p = .654).

Discussion

This study demonstrates that CBT for child and adolescent OCD delivered by telephone is not inferior in efficacy to the traditional method of clinic-based, face-to-face CBT. Blinded assessor ratings of OCD symptom severity demonstrated non-inferiority of TCBT from posttreatment through to 6-month follow-up. The results have significant potential to increase the accessibility of CBT for adolescents with OCD.

At 12-month follow-up, although the confidence intervals for the mean CY-BOCS score in the TCBT group fell outside the non-inferiority margin, no significant difference was observed between the conditions with respect to clinician-rated OCD symptoms, and importantly, according to child- and parent-reported OCD symptoms, TCBT was non-inferior to face-to-face CBT at all follow-up points. Similarly, non-inferiority was established at all time points for all other secondary outcome measures including child depressive symptoms, child global functioning, parental accommodation of OCD symptoms, and parental psychopathology.

Participant and parent responses to the telephone condition were positive. Adherence rates for both conditions were high, and participants and parents from both conditions reported high levels of satisfaction with the treatment received. Of note, a small number of families declined to participate in the study, as they did not want to be allocated to the TCBT condition (n = 8). Conversely, there was a small number of families (n = 3) who declined participation because they did not want to be allocated to the CBT condition, and the reason given was that the travel to the clinic was a deterrent. Interestingly, there was a significant difference between groups with regard to the number of young persons who reported satisfaction with their allocated condition, and a greater number of young persons reported being satisfied with the TCBT condition. Participants and parents within each condition reported equivalent levels of alliance/engagement with their treating therapist, and both conditions were perceived as credible.

Given that a non-inferiority design was used, it is important to consider whether face-to-face CBT was as effective as in previous trials. We found that face-to-face CBT was associated with a 51% reduction in OCD symptom severity over the acute phase of treatment and a 67% reduction from pretreatment to 12-month follow-up, which is commensurate with findings from previous controlled trials. Furthermore, at posttreatment, around 90% were classified as treatment responders and 61% as remitters, rates that are consistent with or better than previous studies. Face-to-face CBT in the current study can therefore be considered to be an appropriate benchmark.

The current findings suggest that TCBT has promise as a mode of treatment delivery and supports the general move toward integrating telehealth into mental health services. TCBT has the potential to extend the availability of CBT to remote areas and/or regions with workforce shortages, thereby reducing geographical inequalities in the availability of CBT. More generally, TCBT may be preferred to traditional face-to-face CBT by some families and services, because of its convenience and its potential to reduce demands on resources (e.g., travel, clinic space). A previous trial conducted among adults with OCD found that TCBT sessions were overall 50% shorter than face-to-face sessions, yet the 2 treatments were associated with equivalent clinical outcomes, suggesting a possible economic benefit of TCBT. Further research is needed to establish whether TCBT may be more cost-effective than face-to-face CBT in pediatric OCD.

There are several strengths to this study, including the following: conformity to CONSORT guidelines for non-inferiority trials; adequate power to test the hypotheses; the inclusion of participants/parents who were seeking help through the UK NHS; inclusion/exclusion criteria designed to maximize generalizability; good retention (i.e., low attrition) rates; use of blinded outcome assessors; and use of a treatment protocol shown to be effective. TCBT was therefore assessed for non-inferiority against a gold-standard treatment.

The findings should also be considered in the context of the study’s limitations. First, the design did not include a no-treatment or placebo treatment control condition. Second, the noninferiority threshold of 5 could be considered as high, and future studies may consider adopting a more stringent threshold to allow for the potential detection of small between-group differences. Third, the study was conducted by experienced CBT therapists, and it cannot be assumed that the findings would necessarily translate to less experienced therapists or other service settings, although previous research has demonstrated that manualized CBT for pediatric OCD is transportable and effective when delivered in community settings by nonspecialist therapists. Fourth, although every effort was made to complete an E/RP task within each session among sessions 3 through 12, there may have been some instances in which this was not possible (e.g., because of participant reluctance or avoidance). In addition, we do not know whether homework compliance differs between CBT and TCBT. Finally, the current study included young persons aged 11 through 18 years, and although this is comparable to previous pediatric trials, the findings may not generalize to a younger population.

In summary, this is the first study to demonstrate that TCBT is as effective as face-to-face CBT for treatment of adolescent OCD, in both the short and medium term, and that it is associated with high levels of patient satisfaction. Further research is needed to explore the longer-term outcomes of TCBT as compared to face-to-face CBT and to examine potential predictors of outcome. Nevertheless, these findings suggest that telephone CBT could be a promising method for overcoming geographical barriers in accessing CBT.

Table S1

Means (Standard Deviations), Mean Differences, and 95% CIs for Secondary Measures of Outcome

Variable	CBT Mean (SD)		TCBT Mean (SD)		Coefficient Mean Diff.	95% CI Lower Bound	95% CI Upper Bound
CY-BOCS Obsessions
Baseline	11.67	(2.30)	12.36	(2.20)
Post-tx (T3)	6.12	(3.10)	6.24	(4.23)	0.06	−1.76	1.87
3-mo FU (T4)	6.12	(3.40)	6.70	(4.40)	0.38	−1.56	2.31
6-mo FU (T5)	5.46	(3.31)	6.06	(4.31)	0.52	−1.34	2.37
12-mo FU (T6)	4.00	(3.84)	5.71	(4.69)	1.60	−0.49	3.70
CY-BOCS Compulsions
Baseline	12.44	(2.05)	13.28	(2.02)
Post-tx (T3)	5.55	(3.19)	6.82	(4.45)	1.09	−0.80	2.99
3-mo FU (T4)	6.04	(3.43)	6.64	(4.43)	0.54	−1.44	2.51
6-mo FU (T5)	5.57	(3.31)	5.83	(4.89)	0.16	−1.87	2.19
12-mo FU (T6)	4.39	(3.80)	5.53	(4.75)	1.41	−0.79	3.60
CY-BOCS Total
Baseline	24.11	(4.02)	25.64	(3.86)
Post-tx (T3)	11.72	(6.06)	12.99	(8.56)	1.06	−2.61	4.73
3-mo FU (T4)	12.23	(6.48)	13.32	(8.67)	0.74	−3.06	4.53
6-mo FU (T5)	10.91	(6.07)	11.90	(9.10)	0.35	−3.45	4.14
12-mo FU (T6)	8.07	(7.36)	10.94	(9.17)	2.78	−1.38	6.94
CHOCI-C Total
Baseline	23.21	(8.70)	24.11	(7.01)
Post-tx (T3)	14.36	(8.21)	13.14	(10.77)	−1.15	−5.96	3.65
3-mo FU (T4)	13.85	(8.74)	14.09	(10.05)	−0.65	−5.30	3.99
6-mo FU (T5)	11.53	(9.49)	13.47	(9.84)	0.67	−4.20	5.54
12-mo FU (T6)	11.95	(9.84)	13.16	(10.97)	0.23	−4.90	5.36
CHOCI-P Total
Baseline	22.65	(8.81)	24.53	(9.12)
Post-tx (T3)	12.26	(9.45)	12.77	(11.87)	−0.24	−3.29	2.81
3-mo FU (T4)	11.61	(11.59)	12.98	(10.01)	0.73	−2.65	4.12
6-mo FU (T5)	11.64	(10.83)	13.06	(9.08)	0.78	−2.53	4.08
12-mo FU (T6)	10.30	(11.52)	11.93	(10.68)	1.68	−1.63	4.99
BDI-Y
Baseline	14.44	(8.77)	14.58	(8.73)
Post-tx (T3)	10.98	(10.16)	11.08	(11.28)	0.11	−4.77	4.99
3-mo FU (T4)	11.38	(10.85)	8.16	(6.88)	−3.67	−8.29	0.95
6-mo FU (T5)	10.04	(9.38)	6.55	(5.02)	−3.81	−7.60	0.30
12-mo FU (T6)	8.22	(6.30)	6.89	(6.95)	−1.71	−4.95	1.52
DASS-M Total
Baseline	17.68	(18.63)	15.53	(14.04)
Post-tx (T3)	13.55	(19.95)	14.58	(16.44)	2.19	−6.86	11.24
3-mo FU (T4)	16.28	(21.10)	8.96	(9.87)	−6.69	−14.85	1.47
6-mo FU (T5)	13.20	(13.08)	13.55	(14.27)	0.83	−6.27	7.92
12-mo FU (T6)	10.60	(14.88)	13.75	(18.53)	4.27	−3.59	12.13
DASS-F Total
Baseline	15.33	(17.11)	8.56	(5.50)
Post-tx (T3)	7.25	(8.92)	13.26	(19.29)	7.85	−3.66	19.35
3-mo FU (T4)	16.33	(18.50)	9.50	(12.11)	−5.79	−15.52	3.95
6-mo FU (T5)	14.40	(19.41)	10.25	(15.57)	−1.26	−10.11	7.58
12-mo FU (T6)	12.79	(11.34)	6.18	(5.78)	−5.04	−13.39	3.31
SDQ-P Total
Baseline	17.21	(3.86)	16.26	(4.37)
Post-tx (T3)	14.37	(3.55)	15.13	(4.24)	0.89	−1.06	2.85
3-mo FU (T4)	13.89	(4.72)	14.24	(4.04)	0.52	−1.76	2.8
6-mo FU (T5)	13.73	(3.39)	13.78	(3.43)	0.06	−1.92	2.04
12-mo FU (T6)	13.84	(4.60)	13.29	(3.74)	−0.34	−2.56	1.89
SDQ-C Total
Baseline	16.63	(3.37)	17.13	(3.71)
Post-tx (T3)	14.64	(3.36)	14.97	(3.90)	0.29	−1.56	2.13
3-mo FU (T4)	14.08	(3.20)	15.08	(3.91)	0.87	−0.94	2.70
6-mo FU (T5)	13.60	(2.90)	15.01	(3.64)	1.17	−0.44	2.78
12-mo FU (T6)	13.64	(3.07)	14.48	(3.59)	0.83	−0.94	2.60
CGAS
Baseline	50.53	(7.06)	50.00	(7.49)
Post-tx (T3)	74.28	(14.23)	71.37	(17.61)	−2.56	−10.18	5.06
3-mo FU (T4)	72.10	(15.01)	70.74	(17.26)	0.31	−7.38	8.00
6-mo FU (T5)	71.50	(10.30)	73.22	(16.94)	3.07	−4.02	9.16
12-mo FU (T6)	76.92	(15.02)	74.25	(14.06)	−2.17	−9.59	5.25
CGI-I
Baseline	4.12	(0.59)	4.10	(0.33)	−0.02	−0.25	0.21
Post-tx (T3)	1.84	(0.67)	1.98	(1.09)	0.13	−0.31	0.58
3-mo FU (T4)	2.14	(1.01)	1.91	(0.92)	−0.24	−0.75	0.26
6-mo FU (T5)	1.97	(0.91)	1.87	(1.09)	−0.11	−0.64	0.42
12-mo FU (T6)	1.61	(1.09)	1.81	(1.01)	0.20	−0.45	0.84

Note: BDI-Y = Beck Depression Inventory for Youth; CBT = cognitive-behavioral therapy; CGAS = Clinical Global Assessment Scale; CGI-I = Clinical Global Impression Improvement Scale; CHOCI-C = Children’s Obsessional Compulsive Inventory Revised–Child; CHOCI-P = Children’s Obsessional Compulsive Inventory Revised–Parent; CY-BOCS = Children’s Yale–Brown Obsessive Compulsive Scale; DASS-F = Depression Anxiety Stress Scale–Father; DASS-M = Depression Anxiety Stress Scale–Mother; FAS-F = Family Accommodation Scale–Father; FAS-M = Family Accommodation Scale–Mother; FU = follow-up assessment; Post-tx = posttreatment; SDQ-C = Strengths and Difficulties Questionnaire–Child Version; SDQ-P = Strengths and Difficulties Questionnaire–Parent Version; TCBT = telephone cognitive-behavioral therapy.