Long-term effect of stepped-care vs in-person cognitive behavioral therapy for pediatric obsessive-compulsive disorder

Abstract

Long-term follow-up data from trials of digital mental health interventions are rare. This study reports 2-year follow-up data from a non-inferiority trial (N = 152) comparing stepped-care (internet-delivered cognitive behavioral therapy [CBT] followed by traditional in-person CBT if needed) vs in-person CBT for pediatric obsessive-compulsive disorder. Both treatment groups had comparable long-term effects, with the majority of participants being responders (stepped-care 66 %; in-person CBT 71 %) 2 years after the end of treatment.

Highlights

• •Long-term follow-ups from trials of digital mental health interventions are rare.
• •2-year follow-up of a non-inferiority trial for pediatric obsessive-compulsive disorder
• •The effects of stepped-care treatment were sustained up to 2 years after treatment.
• •Stepped-care treatment had comparable long-term effects to in-person treatment alone.

Obsessive-compulsive disorder (OCD) is a potentially serious psychiatric disorder affecting about 1–2 % of the population and that typically has an onset in childhood/adolescence (Fawcett et al., 2020; Solmi et al., 2022). Without treatment, the disorder is thought to run a chronic course and be associated with substantial medical morbidity, suicide risk, functional impairment and societal costs (Pérez-Vigil et al., 2018, Lenhard et al., 2021, Micali et al., 2010, Fernández de la Cruz et al., 2022).

Evidence-based treatments for children and adolescents with OCD are serotonin-reuptake inhibitors and cognitive behavior therapy (CBT), and most international guidelines recommend CBT as a first-line treatment due to its favorable side effect profile and patient preference (Geller and March, 2012). CBT is not only an efficacious treatment but also durable, as shown in long-term follow-up studies (Melin et al., 2020; Melin et al., 2018). The main downside of CBT for OCD is that it requires experienced clinicians who are specialized in the treatment of the disorder (Sookman et al., 2021). As one potential solution to improve access to specialist treatment, routine aspects of the treatment can be digitalised. Internet-delivered CBT (ICBT) is a kind of guided self-help, which differs from standard teletherapy in that the therapist provides support asynchronously via a messaging system built in the online platform, instead of actively delivering the treatment in real time; this reduces the required amount of therapist support and also healthcare costs (Vigerland et al., 2016). For youths with OCD, this form of ICBT has been shown to require approximately one third of the therapist time than face-to-face CBT (Lenhard et al., 2017a; Lenhard et al., 2017b), with additional improvements continuing to accrue over a 12-month follow-up period after treatment (Lenhard et al., 2020). A recent trial showed that an ICBT program followed by in-person CBT for non-responders was non-inferior and cost-effective compared to in-person CBT alone for children and adolescents with OCD (Aspvall et al., 2021b; Aspvall et al., 2021a). Here, we report on the planned 2-year long-term evaluation of this stepped-care intervention.

We conducted a long-term follow-up of participants in a previous published randomized non-inferiority trial, where 152 children and adolescents with OCD (age range 8–17 years) were allocated to either stepped-care treatment or in-person CBT alone (Aspvall et al., 2021a). All participants in the stepped-care group first received ICBT for four months (n = 74) and those who were classified as non-responders three months after treatment completion were subsequently offered in-person CBT (n = 34, 45.9 %). Participants in the control group received in-person CBT for four months (n = 78), and non-responders at the 3-month follow-up were also offered additional in-person CBT (n = 23, 29.5 %). The proportion of responders increased to 67.6 % and 66.7 %, respectively, at the 6-month follow-up, which was the primary end point in the original trial (Aspvall et al., 2019; Aspvall et al., 2021a). For an illustration of the ICBT program, see https://vimeo.com/355965105/b3d5d1c439. As pre-specified in our study protocol, naturalistic follow-up assessments were conducted one and two years after the end of treatment (Aspvall et al., 2019). The 1-year follow-up assessments were mainly conducted at the clinic, and when this was not feasible, they were conducted via telephone. The 2-year follow-up assessments were only conducted via telephone. The study was approved by the regional ethical review board in Stockholm, Sweden (DNR 2017/1070-31/1) and registered at ClinicalTrials.gov (Identifier: NCT03263546). All included participants and their caregivers provided a signed informed consent form.

The primary outcome measure was the Children's Yale-Brown Obsessive-Compulsive Scale (CY-BOCS; Scahill et al., 1997). Secondary outcome measures were the Clinical Global Impression-Severity and Improvement scales (CGI-S/I), the Children's Global Assessment Scale, the Obsessive-Compulsive Inventory-Child Version, the Children's Obsessional Compulsive Inventory-Revised-Parent version, the Family Accommodation Scale-Self Rated, the Work and Social Adjustment Scale-Youth and Parent versions, the Mood and Feelings Questionnaire-Child and Parent versions, the Insomnia Severity Index, and the Child Health Utility 9D. The clinician-rated measures were administered by a clinician who had not been directly involved in the treatment of the participants. Self-rated and parent-rated measures were only administered up to the 1-year follow-up.

Between-group differences on the outcome measures were investigated using mixed-effect regression analyses with all time-points included. The model included fixed effects of group, time and an interaction effect of group x time, as well as random intercept and random slope. All randomized participants were included in the analysis. No data imputation method was used. Treatment response was defined as ≥35 % reduction on the CY-BOCS and a CGI-I score of 1 or 2 (Mataix-Cols et al., 2016). Remission was defined as a score of ≤12 on the CY-BOCS and a CGI-S score of 1 or 2 (Mataix-Cols et al., 2016). Proportions of responders and remitters were compared with logistic regression.

In total, 143/152 participants (94.1 %) completed the 1-year follow-up and 133 (87.5 %) the 2-year follow-up. No significant between-group differences were observed on any of the outcome measures (Table 1, Fig. 1). The number (proportion) of responders at the 1-year follow-up was 50 (70.4 %) in the stepped-care group and 57 (79.2 %) in the in-person CBT group (OR = 0.63 [95 % CI 0.29 to 1.34]; P = .23). The corresponding figures at the 2-year follow-up were 43 (66.2 %) and 48 (70.6 %), respectively (OR = 0.81 [95 % CI 0.39 to 1.69]; P = .58). At the 1-year follow-up, 46 (64.8 %) in the stepped-care group were classified as being in remission compared to 49 (68.1 %) in the in-person CBT group (OR = 0.86 [95 % CI 0.43 to 1.73]; P = .68). The corresponding figures at the 2-year follow-up were 43 (66.2 %) and 48 (70.6 %), respectively (OR = 0.81 [95 % CI 0.39 to 1.69]; P = .58).

Table 1.

Observed means and estimated effect sizes with the CY-BOCS (N = 152).

	Mean (SD)		Between-group d (95 % CI)a	Estimated absolute difference (CI)	P value
	Stepped-care (n = 74)	In-person CBT (n = 78)
Clinician-ratedb
CY-BOCS
Pretreatment	22.96 (3.64)	22.95 (3.70)
6-month FU	11.57 (6.40)	10.57 (7.57)
1-year FU	9.62 (6.90)	8.86 (6.92)	0.10 (−0.27 to 0.46)	0.68 (−1.83 to 3.19)	.60
2-year FU	8.82 (8.65)	8.35 (8.87)	0.02 (−0.44 to 0.40)	0.20 (−2.86 to 3.26)	.90
CGAS
Pretreatment	55.05 (5.59)	56.29 (7.16)
6-month FU	66.26 (11.46)	67.99 (12.28)
1-year FU	69.21 (12.30)	70.29 (11.37)	0.03 (−0.29 to 0.35)	−1.16 (−5.18 to 2.86)	.57
2-year FU	73.55 (13.06)	74.59 (14.55)	0.12 (−0.27 to 0.51)	−0.20 (−4.94 to 4.55)	.93
Self-ratedb
OCI-CV
Pretreatment	18.70 (7.52)	20.35 (6.42)
6-month FU	8.14 (6.68)	8.82 (7.75)
1-year FU	8.60 (7.66)	9.32 (8.21)	0.12 (−0.21 to 0.45)	−1.03 (−3.91 to 1.84)	.48
WSAS-Y
Pretreatment	15.12 (7.96)	17.19 (8.66)
6-month FU	6.10 (5.50)	7.46 (7.71)
1-year FU	6.33 (7.23)	6.87 (7.53)	0.21 (−0.19 to 0.62)	−0.15 (−3.19 to 2.88)	.92
MFQ-C
Pretreatment	8.93 (5.62)	9.91 (6.26)
6-month FU	4.84 (4.92)	4.97 (5.71)
1-year FU	5.15 (5.93)	5.75 (6.42)	0.12 (−0.24 to 0.48)	−0.43 (−2.78 to 1.93)	.72
ISI
Pretreatment	6.99 (5.68)	5.78 (4.88)
6-month FU	4.76 (4.45)	4.03 (5.01)
1-year FU	5.03 (4.95)	3.94 (4.13)	0.05 (−0.24 to 0.33)	1.27 (−0.47 to 3.01)	.15
CHU9D
Pretreatment	19.93 (5.71)	19.97 (5.45)
6-month FU	16.05 (5.15)	14.76 (4.85)
1-year FU	15.08 (5.29)	14.89 (5.00)	0.12 (−0.25 to 0.49)	0.42 (−1.56 to 2.40)	.68
Parent-ratedb
ChOCI-R-P
Pretreatment	27.70 (6.61)	28.51 (7.77)
6-month FU	13.26 (7.91)	13.49 (9.65)
1-year FU	11.62 (9.10)	10.96 (8.03)	0.12 (−0.22 to 0.45)	0.76 (−2.65 to 4.18)	.66
FAS-SR
Pretreatment	19.53 (14.43)	21.44 (15.99)
6-month FU	5.23 (7.44)	8.17 (10.66)
1-year FU	4.31 (5.94)	5.88 (9.52)	0.09 (−0.26 to 0.44)	−0.33 (−4.11 to 3.45)	.86
WSAS-P
Pretreatment	16.59 (8.45)	17.59 (8.31)
6-month FU	7.47 (7.41)	7.81 (8.51)
1-year FU	5.40 (6.52)	5.82 (6.18)	0.12 (−0.24 to 0.48)	−0.10 (−3.01 to 2.80)	.95
MFQ-P
Pretreatment	9.18 (5.63)	10.29 (5.89)
6-month FU	4.68 (4.73)	5.57 (6.04)
1-year FU	4.74 (4.83)	4.75 (5.91)	0.18 (−0.17 to 0.52)	0.06 (−1.93 to 2.06)	.95

Abbreviations: CBT, cognitive behavioral therapy; CGAS, Children's Global Assessment Scale; ChOCI-R-P, Children's Obsessional Compulsive Inventory-Revised-Parent version; CHU9D, Child Health Utility 9D; CY-BOCS, Children's Yale-Brown Obsessive-Compulsive Scale; FAS-SR, Family Accommodation Scale-Self Rated; FU, follow-up; ISI, Insomnia Severity Index; MFQ-C, Mood and Feelings Questionnaire-Child version; MFQ-P, Mood and Feelings Questionnaire-Parent version; OCI-CV, Obsessive-Compulsive Inventory-Child Version; WSAS-P, Work and Social Adjustment Scale-Parent version; WSAS-Y, Work and Social Adjustment Scale-Youth version.

^aFrom pretreatment to 1-year FU and 2-year FU respectively. The between-group effect sizes were calculated with the m_effectsize script in Stata (freely available by typing ‘net install m_effectsize, from [http://www.imm.ki.se/biostatistics/stata] replace’ in Stata).

^bNumber of participants who completed the 1-year FU: Clinician-rated 71 (95.9 %), self-rated 60 (81.1 %) and parent-rated 65 (87.8 %) in the stepped-care group and clinician-rated 72 (92.3 %), self-rated 63 (80.8 %) and parent-rated 68 (87.2 %) in the in-person CBT group. Number of participants who completed the 2-year FU: 65 (87.8 %) in the stepped-care group and 68 (87.2 %) in the in-person CBT group.

Fig. 1.

Obsessive-compulsive symptom severity over the total study period.

Effect of treatment on the primary outcome measure (CY-BOCS). Scores are shown at pretreatment, posttreatment, 3-month follow-up, 6-month follow-up (primary end point), 1-year follow-up and 2-year follow-up. The circles represent the mean values and the whiskers the 95 % confidence intervals.

Abbreviations: CBT, cognitive behavioral therapy; CY-BOCS, Children's Yale-Brown Obsessive-Compulsive Scale.

Twenty participants (30.8 %) in the stepped-care group and 24 (35.3 %) in the in-person CBT group reported receiving additional treatment for OCD (i.e., new or increased dosage in pharmacological treatment and/or additional CBT sessions) during the naturalistic follow-up period between the 6-month and 2-year follow-up, a non-significant difference (X²(1) = 0.31; P = .58). A post-hoc analysis excluding these individuals from the models did not alter the overall results. At the 2-year follow-up, the mean (SD) CY-BOCS score was 7.79 (SD = 8.29) in the stepped-care group and 7.02 (SD = 8.09) in the in-person CBT group (estimated mean difference of 0.89 [95 % CI, −2.74 to 4.52]; P = .63), and 32 (71.1 %) participants in the stepped-care group and 33 (75.0 %) in the in-person CBT group were classified as being both responders and remitters. The participants receiving additional help for their OCD during the follow-up period had slightly higher average CY-BOCS severity scores compared to the participants who did not receive additional help; 6-month FU: 14.5 (SD = 6.8) versus 9.2 (SD = 6.6; t(131) = −4.3, P < .001), 1-year FU: 12.4 (SD = 7.0) versus 7.6 (SD = 6.4; t(131) = −3.9, P < .001), and 2-year FU: 11.1 (SD = 9.6) versus 7.5 (SD = 8.2; t(129) = −2.3, P < .05).

The main finding was that the effects of ICBT followed by in-person CBT for non-responders were comparable to the effects of in-person CBT alone up to two years after the end of treatment. The between-group difference on the primary outcome measure was below the noninferiority margin of 4 points used in the original trial at both follow-up assessments (Aspvall et al., 2021a). The proportion of responders and remitters was somewhat lower to that of other trials (Melin et al., 2020; Melin et al., 2018), but we employed stricter definitions of response and remission established through international expert consensus (Mataix-Cols et al., 2016). One potential risk of bias was the naturalistic design where participants were free to seek additional treatment during the follow-up period. We found that a similar proportion (up to a third) of individuals in both groups had sought additional treatments. Furthermore, post-hoc exclusion of these participants did not alter the results. The assessing clinicians were not blinded to treatment allocation, but no between-group differences emerged on any of the clinician-rated, self-rated or parent-rated measures. Also, the format of administration of the 1-year (mainly in-person) and the 2-year (telephone) assessments differed. Overall, the results provide additional evidence that ICBT delivered in a stepped-care fashion is a safe, effective and durable treatment model for young people with OCD. Implementation of ICBT as a first-step treatment option for young people with OCD in routine healthcare is now feasible. Implementation should preferably be done in the context of specialist services with expertise in treating OCD. This would ensure the possibility to offer seamless access to in-person CBT and other evidence based alternatives, such as medication, in case of non-response or patient preference. Future studies should evaluate ICBT in other countries and heath care contexts to ensure generalizability, and also evaluate if it is possible to identify early predictors of clinical response to each delivery mode.

CRediT authorship contribution statement

EA, KM, DMC, FL, ES and KA contributed to the initial study design. KOL, LN, VE and KA contributed to acquisition of data. KOL and KA conducted the statistical analysis. KOL, EA, DMC and KA drafted the manuscript. All authors have read and approved the final version of the manuscript. KA had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Prof Mataix-Cols receives royalties for contributing articles to UpToDate, Inc., outside the submitted work. Dr Andersson receives royalties from Natur & Kultur for a self-help book on health anxiety.

Data availability

The dataset analyzed during the current study is not publicly available due to Swedish and EU legislation, but can be made available from the corresponding author on a reasonable request on a case by case basis, according to the current legislation and ethical permits.