This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Mental Health, is properly cited. The complete bibliographic information, a link to the original publication on http://mental.jmir.org/, as well as this copyright and license information must be included.
Face-to-face individual cognitive behavioral therapy (CBT) and internet-based CBT (ICBT) without videoconferencing are known to have long-term effectiveness for obsessive-compulsive disorder (OCD), panic disorder (PD), and social anxiety disorder (SAD). However, videoconference-delivered CBT (VCBT) has not been investigated regarding its long-term effectiveness and cost-effectiveness.
The purpose of this study was to investigate the long-term effectiveness and cost-effectiveness of VCBT for patients with OCD, PD, or SAD in Japan via a 1-year follow-up to our previous 16-week single-arm study.
Written informed consent was obtained from 25 of 29 eligible patients with OCD, PD, and SAD who had completed VCBT in our clinical trial. Participants were assessed at baseline, end of treatment, and at the follow-up end points of 3, 6, and 12 months. Outcomes were the Yale-Brown Obsessive-Compulsive Scale (Y-BOCS), Panic Disorder Severity Scale (PDSS), Liebowitz Social Anxiety Scale (LSAS), Patient Health Questionnaire–9 (PHQ-9), General Anxiety Disorder–7 (GAD-7), and EuroQol-5D-5L (EQ-5D-5L). To analyze long-term effectiveness, we used mixed-model analysis of variance. To analyze cost-effectiveness, we employed relevant public data and derived data on VCBT implementation costs from Japanese national health insurance data.
Four males and 21 females with an average age of 35.1 (SD 8.6) years participated in the 1-year follow-up study. Principal diagnoses were OCD (n=10), PD (n=7), and SAD (n=8). The change at 12 months on the Y-BOCS was −4.1 (
VCBT was a cost-effective way to effectively treat Japanese patients with OCD, PD, or SAD.
University Hospital Medical Information Network Clinical Trials Registry UMIN000026609; https://upload.umin.ac.jp/cgi-open-bin/ctr_e/ctr_view.cgi?recptno=R000030495
Obsessive-compulsive disorder (OCD), panic disorder (PD), and social anxiety disorder (SAD) are mental health illnesses that create severe obstacles for patients in their daily lives [
Telepsychiatry can be delivered to established therapy patients in developed countries where there is wide availability of information and communication devices and internet use is high. Within telepsychiatry, videoconference-delivered cognitive behavioral therapy (VCBT) has proved promising, with the potential to improve the accessibility of specialized care to patients with OCD, PD, and SAD [
VCBT requires a videoconferencing system, thereby making it more expensive compared with face-to-face cognitive behavioral therapy (CBT). For facilities that provide health care services, VCBT is a little more expensive than traditional CBT. However, for patients, VCBT is less burdensome than face-to-face CBT, as there are no travel costs or time costs associated with hospital visits. VCBT puts the burden of cost on the facility; thus, it is particularly important to assess whether its adoption is a worthwhile approach from the perspective of efficient health care policy.
This study’s main objectives were to assess the long-term effectiveness of VCBT for patients with OCD, PD, or SAD and estimate its cost-effectiveness in Japan.
In this study, we included data from our previous clinical trials and follow-ups [
In March 2018, the Cognitive Behavioral Therapy Center at Chiba University Hospital implemented a prospective observational study involving all patients who participated in VCBT (reference number: G28038, UMIN000026609) [
All participants had received face-to-face treatment from the attending physician (psychiatrist) during a previous clinical trial period [
The following Japanese version of three scales were used to assess the severity of the three disorders. The Yale-Brown Obsessive-Compulsive Scale (Y-BOCS) was used to measure OCD by identifying the patient’ contents of obsessions and compulsions on the symptom checklist and assessing their severity in 4 stages using responses to 10 items on the symptom severity scale [
We also assessed depression and general anxiety associated using responses to the Patient Health Questionnaire–9 (PHQ-9) and Generalized Anxiety Disorder 7 (GAD-7). PHQ-9 has 9 questions related to depression status set [
To calculate responsiveness to VCBT treatment and remission rates after the VCBT, we used criteria employed by previous studies regarding the severity rating scales of the three disorders (Y-BOCS, PDSS, and LSAS). Regarding OCD, treatment response was defined as a 35% or greater reduction in the total Y-BOCS score, and remission was defined as a 12-month Y-BOCS≤14 [
We calculated the total VCBT cost using the sum of the costs of implementing the intervention: (1) health care costs (¥3500-¥4800 × 16 sessions) and (2) costs of videoconferencing (license fee ¥1490 per month × 4 months in Webex (Cisco), ¥300 × 16 sessions in curon (MICIN, Inc) [
We did not assume that hardware would have to be newly purchased in order to access VCBT. This was because, as reported by the Ministry of Internal Affairs and Communications in 2017, ownership of information communication equipment in Japan was at 94.8% for mobile devices in general and 72.5% for PCs and because the penetration rate of information and communication equipment and the internet was at more than 80.9% for all households [
Statistical analysis and reporting were performed in accord with the CONSORT-EHEALTH guidelines [
The main analysis compared the baseline assessment scores with those obtained at the 12-month posttreatment follow-up. The differences were estimated using mixed-model analysis of variance (ANOVA) on all patients displaying symptoms in each scale (Y-BOCS, PDSS, LSAS, PHQ-9, and GAD-7), taking into account missing values, individual variance, and multiple measurement points.
Analysis of secondary outcomes was performed in an identical fashion to that of the primary analysis. To analyze cost-effectiveness using the EQ-5D-5L, QALY scores were estimated via area-under-the-curve analysis, which involved summing the areas of the distribution shapes for utility scores over the study period [
The amount of change in QALY was calculated from the difference between QALY and the actually observed utility value assuming no change from the utility value of EQ-5D-5L at baseline. We calculated a summary statistic for the change in QALY and performed a paired
Cost-effectiveness of the VCBT was analyzed as follows. The additional consumption of health care resources was divided by the benefits (such as QALY) gained from the health care intervention to calculate an incremental cost-effectiveness ratio (ICER). When the ICER, such as cost per QALY, was less than a predetermined threshold, the intervention was considered cost-effective [
Cost-effectiveness of VCBT = WTP − cost of VCBT
Cost of VCBT = (videoconference system costs) + traditional CBT costs
WTP = increased QALYs × threshold in Japan (¥5 million)
Calculated cost-effectiveness greater than one indicated that VCBT was a cost-effective intervention. WTP was calculated by multiplying the increase in QALY between baseline and 12-month follow-up after VCBT using the Japanese cost-effectiveness threshold (¥5 million). Incremental cost-effectiveness ratio per QALY was calculated by dividing the increase in QALY between baseline and 12-month follow-up after VCBT using total cost of VCBT.
The sample comprised 4 males and 21 females, aged 20 to 54 years (mean 35.1 [SD 8.6] years) with 12 to 18 years of education (mean 14.72 [SD 1.90] years). Except for their principal diagnoses, participants’ demographic and diagnostic data are described in
Participant clinical and demographic characteristics.
Characteristics | Overall (n=25) | OCDa (n=10) | PDb (n=7) | SADc (n=8) | |
Age in years, mean (SD) | 35.1 (8.6) | 37.7 (6.9) | 36.1 (9.3) | 30.9 (9.4) | |
Gender (female), n (%) | 21 (84) | 8 (80) | 7 (100) | 6 (75) | |
Employed, n (%) | 14 (56) | 3 (12) | 5 (71) | 6 (75) | |
Pharmacotherapy (yes), n (%) | 9 (36) | 5 (50) | 3 (43) | 1 (13) | |
|
|
|
|
|
|
|
Depression | 3 (12) | 1 (10) | 0 (0) | 2 (24) |
|
Panic/agoraphobia | 2 (11) | 2 (20) | 0 (0) | 0 (0) |
|
PTSDd | 1 (4) | 1 (10) | 0 (0) | 0 (0) |
|
Alcohol dependence | 1 (4) | 0 (0) | 0 (0) | 1 (13) |
aOCD: obsessive-compulsive disorder.
bPD: panic disorder.
cSAD: social anxiety disorder.
dPTSD: posttraumatic stress disorder.
Participant flow.
Mixed-model ANOVA results regarding the long-term effectiveness of VCBT showed statistically significant improvement in participant symptoms (
Mixed-model analysis of variance results on changes in participant symptomology.
Characteristics | n | Score mean (SD) | Min-maxa | ||
|
— | — | — | .04 | |
|
Baseline | 10 | 23.3 (6.5) | 15 (36) | — |
|
Posttreatment | 10 | 17.1 (9.9) | 2 (34) | — |
|
3-month | 10 | 19.4 (7.5) | 9 (32) | — |
|
6-month | 10 | 18.6 (8.1) | 7 (32) | — |
|
12-month | 10 | 19.2 (8.4) | 8 (29) | — |
|
— | — | — | .01 | |
|
Baseline | 7 | 8.9 (3.8) | 5 (16) | — |
|
Posttreatment | 7 | 5.3 (6.7) | 0 (19) | — |
|
3-month | 7 | 5.4 (4.9) | 2 (13) | — |
|
6-month | 6 | 4.5(6.1) | 0 (16) | — |
|
12-month | 6 | 4.5 (3.6) | 0 (10) | — |
|
— | — | — | .01 | |
|
Baseline | 8 | 96.6 (27.3) | 53 (132) | — |
|
Posttreatment | 8 | 57.4 (34.7) | 21 (128) | — |
|
3-month | 8 | 62.6 (34.4) | 20 (112) | — |
|
6-month | 6 | 57.3 (31.9) | 7 (85) | — |
|
12-month | 7 | 65.7 (43.8) | 10 (118) | — |
|
— | — | — | .007 | |
|
Baseline | 25 | 8.8 (6.2) | 0 (23) | — |
|
Posttreatment | 25 | 6.8 (7.0) | 0 (22) | — |
|
3-month | 25 | 7.2 (5.8) | 0 (24) | — |
|
6-month | 22 | 6.6 (6.1) | 0 (19) | — |
|
12-month | 23 | 6.1 (5.7) | 0 (20) | — |
|
— | — | — | .009 | |
|
Baseline | 25 | 8.8 (5.3) | 0 (20) | — |
|
Posttreatment | 25 | 5.5 (5.1) | 0 (16) | — |
|
3-month | 25 | 7.2 (4.6) | 0 (19) | — |
|
6-month | 25 | 6.3 (5.0) | 0 (21) | — |
|
12-month | 23 | 5.8 (4.5) | 0 (14) | — |
amin-max: minimum to maximum.
bY-BOCS: Yale-Brown Obsessive-Compulsive Scale.
cPDSS: Panic Disorder Severity Scale.
dLSAS: Livobitz Social Anxiety Scale.
ePHQ-9: Patient Health Questionnaire–9.
fGAD-7: Generalized Anxiety Disorder–7.
Participants’ changes in symptomology.
To investigate the predictive effects that symptoms of patients’ depression at pretreatment may have had on the treatment response change posttreatment, multiple regression analyses in simultaneous forced entry were performed. The treatment response percentage change was set as a dependent variable in multiple regression analyses. We set depressive symptoms due to PHQ-9 as independent variables. The treatment response percentage change was calculated by dividing the total baseline score with the score difference between baseline and 12-month. The treatment response percentage change in this study was the decline in baseline Y-BOCS, PDSS, or LSAS score.
The degree of change (in percentages) in the treatment response was analyzed as a continuous variable. Statistical analysis was performed using SPSS Statistics version 24.00 (IBM Corp). Multiple regression analysis showed that the effects of depression on therapeutic response rates were not significant across the data (
At the 12-month follow-up assessment, treatment response rate was 32% (8/25) and remission rate was 40% (10/25;
Participant response and remission rates at each follow-up end point.
Characteristics | Overall (n=25), n (%) | OCDa (n=10), n (%) | PDb (n=7), n (%) | SADc (n=8), n (%) | |
|
|
|
|
|
|
|
Posttreatment | 12 (48) | 4 (40) | 4 (57) | 4 (50) |
|
3-month | 10 (40) | 2 (20) | 5 (71) | 3 (38) |
|
6-month | 8 (32) | 2 (20) | 3 (43) | 3 (38) |
|
12-month | 8 (32) | 2 (20) | 2 (29) | 4 (50) |
|
|
|
|
|
|
|
Posttreatment | 12 (48) | 4 (40) | 6 (86) | 2 (25) |
|
3-month | 11 (44) | 4 (40) | 5 (71) | 2 (25) |
|
6-month | 10 (40) | 3 (30) | 5 (71) | 2 (25) |
|
12-month | 10 (40) | 3 (30) | 5 (71) | 2 (25) |
aOCD: obsessive-compulsive disorder.
bPD: panic disorder.
cSAD: social anxiety disorder.
EuroQol 5-Dimension 5-Level index each end point.
Characteristics | n | Mean | SD | SE | |
|
|
|
|
|
|
|
Baseline | 25 | 0.7206 | 0.14 | — |
|
Posttreatment | 25 | 0.7677 | 0.20 | — |
|
3-month | 25 | 0.7350 | 0.17 | — |
|
6-month | 22 | 0.7207 | 0.24 | — |
|
8-month | 20 | 0.7760 | 0.15 | — |
|
12-month | 23 | 0.7503 | 0.15 | — |
|
|
|
|
|
|
|
6-month | 25 | 0.7342 | 0.23 | — |
|
8-month | 25 | 0.7669 | 0.15 | — |
|
12-month | 25 | 0.7530 | 0.15 | — |
|
|
|
|
|
|
|
6-month | 25 | 0.7075 | — | 0.05 |
|
8-month | 25 | 0.7651 | — | 0.03 |
|
12-month | 25 | 0.7564 | — | 0.03 |
aLOCF: last observation carried forward.
bMICE: multivariate imputation by chained equations.
Paired
Characteristics | n | Mean | SE | 95% CI | |
Complete cases | 19 | 0.0379 | 0.01 | 0.0085-0.0674 | .02 |
LOCFa | 25 | 0.0214 | 0.01 | 0.0067-0.0495 | .13 |
MICEb | 25 | 0.0187 | 0.01 | 0.0093-0.0466 | .19 |
aLOCF: last observation carried forward.
bMICE: multivariate imputation by chained equations.
The quality-adjusted life years (QALYs) observed at follow-up and QALY in complete cases. Note: Estimated QALYs was calculated in terms of effective value without videoconference-delivered cognitive behavioral therapy at each time point as a baseline, with the area under the curve assuming no change. Increased QALY was calculated as the difference between the measured utility value and the estimated QALY.
Quality-adjusted life years at 12 months after videoconference-delivered cognitive behavioral therapy.
Characteristics | n | Mean | SE | 95% CI |
Complete cases | 19 | 0.7469 | 0.04 | 0.6728-0.8210 |
LOCFa | 25 | 0.7420 | 0.03 | 0.6839-0.8001 |
MICEb | 25 | 0.8343 | 0.04 | 0.7565-0.9121 |
aLOCF: last observation carried forward.
bMICE: multivariate imputation by chained equations.
Results of a cost-utility analysis.
Characteristics | Value | |
|
|
|
|
CBTa by a nurse | ¥56,000 |
|
CBT by a medical doctor | ¥76,000 |
|
Videoconferencing in Webex | ¥5960 |
|
Videoconferencing in Curon | ¥4800 |
|
Total cost | ¥60,800-¥81,960 |
|
|
|
|
Complete case | 0.7469 |
|
LOCFc | 0.742 |
|
MICEd | 0.8343 |
|
|
|
|
Complete case | 0.0379 |
|
LOCF | 0.0214 |
|
MICE | 0.0187 |
|
|
|
|
Complete case | ¥1,604,222 to ¥2,162,533 |
|
LOCF | ¥2,841,122 to ¥3,829,907 |
|
MICE | ¥3,251,337 to ¥4,382,888 |
|
|
|
|
Complete case | ¥189,500 |
|
LOCF | ¥107,000 |
|
MICE | ¥93,500 |
aCBT: cognitive behavioral therapy.
bQALY: quality adjusted life year.
cLOCF: last observation carried forward.
dMICE: multivariate imputation by chained equations.
We investigated the long-term effectiveness and cost-effectiveness of VCBT in 25 patients with OCD, PD, or SAD in a 12-month observational study. The principal symptomology of OCD, PD, and SAD significantly decreased and the QALY significantly improved. The therapeutic response rate was 32% (8/25) and remission rate was 40% (10/25) at the 12-month postintervention follow-up assessment. The total cost of providing VCBT was ¥60,800 to ¥81,960 per patient; in contrast, the threshold using WTP was ¥189,500. Therefore, our results suggested that VCBT was a cost-effective intervention for this sample of patients with OCD, PD, or SAD in Japan.
In a previous study on VCBT provided to 10 adult OCD patients, 2 patient scores were below the Y-BOCS cutoff (<14) after treatment, but just one patient was below the cutoff 3 months later [
In a study that provided VCBT to 11 adult PD patients, 82% (9/11) had improved symptoms after the intervention and 91% (10/11) had improved symptoms after 6 months and no panic attacks [
In a study of VCBT in 24 adult patients with SAD, 54% (13/24) experienced remission after treatment, and symptoms that had decreased were maintained at that lower level 6 months later [
Several studies have reported that internet-delivered cognitive behavioral therapy (ICBT) provided to patients with depression saved on direct medical costs more than providing just the usual care [
This study provides the world’s first empirical knowledge about the cost-effectiveness of VCBT. VCBT costs totaled ¥60,800 to ¥81,960, which was far below the ¥189,500 threshold based on WTP calculated using the QALY. In other words, under the Japanese insurance system in 2018 [
This study has some limitations. First, there was no statistical control over the relationship between VCBT and pharmacological therapy during our previous trial and this follow-up study. Studies have suggested that combining therapeutic approaches with drug therapy is particularly effective in panic disorder prognoses [
Our results suggest that VCBT for patients with OCD, PD, and SAD was effective in improving symptoms over 12 months and was a cost-effective approach in Japan.
analysis of variance
cognitive behavioral therapy
EuroQol 5-Dimension 5-Level
Generalized Anxiety Disorder–7
internet-delivered cognitive behavioral therapy
incremental cost-effectiveness ratio
last observation carried forward
Liebowitz Social Anxiety Scale
multivariate imputation by chained equations
National Institute for Health and Care Excellence
obsessive-complusive disorder
panic disorder
Panic Disorder Severity Scale
Patient Health Questionnaire–9
quality-adjusted life year
randomized controlled trial
social anxiety disorder
videoconference-delivered cognitive behavioral therapy
willingness-to-pay
Yale-Brown Obsessive-Compulsive Scale
We express our gratitude and respect for the dedication and contribution of research participants. This study was supported by the Japan Society for the Promotion of Science KAKENHI Grant-in-Aid for Scientific Research C (Nos. 18K03130 and 18K17313). The funding sources had no role in the design and conduct of the study.
KM contributed to design of research, data collection and analysis, and the development of the manuscript. SH contributed to data collection. KN contributed to the design and statistical analysis in this study. CS contributed to management of this study. AN contributed to data collection. ES contributed to the drafting and planning of this study and provided interpretations based on the results.
None declared.