Telepsychiatry is becoming a new way of providing psychiatric treatment. However, lack of evidence for real-world clinical setting effectiveness hampers its approval by medical insurance in some countries.
Results
A total of 199 patients participated in this study. After 24 weeks of treatment, two-way video treatment was found to be noninferior to face-to-face treatment regarding SF-36 MCS score (48.50 vs 46.68, respectively; p < 0.001). There were no significant differences between the groups regarding most secondary end points, including all-cause discontinuation, treatment efficacy, and satisfaction.
Conclusion
Two-way video treatment using smartphones and other devices, was noninferior to face-to-face treatment in real-world clinical settings. Modern telemedicine, easily accessible from home, can be used as a form of health care.
Methods
Study Design
Details regarding the study methods and protocols have been previously published.20 This was a multisite, prospective RCT. Patients were assigned in a 1:1 ratio to either a two-way video group (at least 50% of treatment sessions to be conducted by two-way video, with at least one face-to-face session within 6 months) or the face-to-face group (all treatments sessions to be face-to-face). Patients in the two-way video group interacted with their psychiatrists from a private location, such as their home or office, using a smartphone, tablet, or personal computer. Both groups received standard treatment covered by public medical insurance for 24 weeks. The intervals between treatments were determined at the discretion of the psychiatrist in charge.
Participants
Participants were recruited at 19 medical institutions providing psychiatric services in 11 prefectures in Japan between April 2021 and February 2022.
Patients were included based on the following inclusion criteria: (1) met DSM-521 criteria for depressive disorders, anxiety disorders, or obsessive-compulsive disorder (OCD) and related disorders and were outpatients at a participating medical institution; (2) were 18 years or older at the time of obtaining consent; (3) needed continuous treatment for the next 6 months or more (at the discretion of the attending physician); (4) had a smart phone or personal computer as well as access to video-calling over the internet (even if available only with family support); (5) their psychiatric condition was stable enough for them to undergo two-way video treatment, based on the clinical judgment of the attending physician; (6) their psychiatric condition was stable enough for them to have sufficient capacity to provide consent, based on the clinical judgment of the attending physician; and (7) provided written consent to participate in the study. For patients who were minors (younger than 20 years), written consent had to be obtained from the patient and his/her guardian.
The exclusion criteria were as follows: (1) likelihood of requiring unscheduled or urgent treatment at a hospital in addition to regular treatment because of emergent suicidal ideation, anxiety, or agitation; and (2) patients who would have had difficulty in managing an emergency visit by themselves if their psychiatric condition deteriorated (e.g. the hospital was far away).
Randomization
Participating patients were randomly assigned in a 1:1 ratio to either the two-way video group or the face-to-face group for treatment during the study period. To avoid interinstitutional differences and biases among the groups regarding types of disorders, randomization was performed by a blinded, independent third party using a modified minimization method and biased-coin assignment22 balanced for age group (≥60 years or <60 years), sex (male or female), target disorder, and participating institution. Additionally, the allocation results were not disclosed to the central evaluator to minimize bias.
Assessment Schedule
After randomization, participants completed the following assessments through self-rating scales and interviews as baseline assessments and again at weeks 12 and 24.
Primary Outcome
The primary outcome was the 36-Item Short-Form Health Survey Mental Component Summary (SF-36 MCS) score at week 24. The SF-36, a scientifically validated and reliable instrument for assessing health-related quality of life, consists of a self-administered questionnaire23 to which patients in this study responded through a dedicated application. The SF-36 MCS focuses on mental items and was used because the present study targets multiple psychiatric disorders.
Secondary Outcomes
The following secondary outcomes were assessed: (1) SF-36 Physical Component Summary (PCS) scores; (2) all-cause discontinuation (in the two-way video group, if the patient discontinued two-way video and switched to face-to-face treatment only, the patient was considered to have dropped out of the two-way video group; (3) Working Alliance Inventory (WAI) score (assessed at weeks 12 and 24) as a measure of treatment alliance24; (4) Client Satisfaction Questionnaire (CSQ) score (assessed at weeks 12 and 24) for assessing satisfaction25; (5) adverse events; (6) cost and time (assessed using a self-administered questionnaire on costs and time associated with medical treatments); (7) EuroQol 5 Dimension (EQ-5D) score (assessed at baseline and at weeks 12 and 24) as another measure of health-related quality of life26; (8) degree of anxiety regarding coronavirus disease 2019 (COVID-19); (9) comments about two-way video; (10) for the depressive disorder group, the Hamilton Depression Rating Scale (HAMD) score27; (11) for the anxiety disorder group, the Hamilton Anxiety Rating Scale (HAMA) score28; and (12) for the OCD and related disorders group, the Yale-Brown Obsessive Compulsive Scale (YBOCS) score.29
Sample Size
The sample size was calculated based on previous psychiatric intervention studies (including those involving psychotherapy and electroconvulsive therapy interventions), in which the evaluation period was 6 months.30–35
In previous studies, the mean SF-36 MCS scores ranged from 30 to 50 (SD, 9–14). In the present study, assuming that an SF-36 MCS score of 45 in both the two-way video and face-to-face groups at 6 months (no difference between the two groups), with an SD of 12 and a noninferiority margin of five, the required number of patients in each group would be 92 under the conditions of 80% power and a one-sided significance level of 2.5%.
The all-cause discontinuation rate was expected to be low in this study, because the primary psychiatrist who had been treating a patient until the time of the study would continue to be in charge of the treatment, regardless of whether the patient was in the two-way video or face-to-face group. Assuming an all-cause discontinuation rate of approximately 10%, the total number of required patients was calculated as 200, or 100 in each group.
Data Collection and Management
Data on the SF-36 MCS and SF-36 PCS scores, treatment alliance and satisfaction measures, cost, EQ-5D score, and degree of anxiety about COVID-19 were collected as self-administered patient-reported values. All such electronic patient-reported outcome data were collected through the participants’ smartphones using an electronic data capture system. For the HAMD, HAMA, and YBOCS scores, remote centralized ratings were obtained through two-way video. Evaluators were required to have completed a total of at least 30 h of training on these evaluation items.
Statistical Analyses
The full analysis set (FAS), which included all patients who completed at least one SF-36 MCS assessment during the study period and did not present any serious violation of the study protocol or the ethical research guidelines, was used for the analysis of the primary outcome. The per-protocol set (PPS), which is the supplemental analysis population for the primary outcome, was defined as the population excluding patients in serious violation of the study protocol from among the FAS, i.e. (1) violation of selection/exclusion criteria; (2) violation of discontinuation criteria; (3) violations related to therapies for which concomitant use was prohibited; and (4) lack of follow-up data. The primary analysis was performed for FAS and PPS, and secondary efficacy analyses and exploratory analyses were conducted only for FAS. Safety analysis was performed on the safety analysis set, which was defined as the set of patients enrolled in the study and who underwent at least one SF-36 MCS assessment in addition to that at baseline. As appropriate, χ2 and Fisher exact tests were used for categorical variables, while Wilcoxon rank sum test and t test were employed for continuous variables. In the primary analysis, point estimates and their 95% confidence intervals were estimated for each time point using a mixed-effects model for repeated measures. The correlation structure was assumed to be unstructured. Adjustment factors for allocation were adjusted, a restricted maximum likelihood estimator was used as the estimator of each parameter, and the Kenward-Roger method was used to estimate the variance of the parameter estimators and the degrees of freedom.36 The noninferiority margin was set to −5. The statistical analysis plan was developed by the principal investigator and the biostatistician before the completion of patient recruitment and data fixation. A one-sided P-value <0.025 and a two-sided P-value <0.05 were considered statistically significant. Statistical analyses were performed with SAS version 9.4 (SAS Institute Inc.).
In addition to the analyses described above, the possibility was considered that there might have been a difference in the efficacy of the treatment in the two-way video group that used as many telemedicine visits as possible versus the group that did not. Therefore, as a post hoc analysis, we performed the same analysis for the primary end point in the patients’ group that had 100% of their postbaseline visits performed via two-way video.
Ethical Considerations
This study was approved by the institutional review board of the National Center of Neurology and Psychiatry and the participating medical facilities. The trial was registered with the Japan Registry of Clinical Trials (jRCT1030210037). Written informed consent was obtained from all participants. The study procedures were conducted according to the Declaration of Helsinki.
Results
A total of 199 patients were assessed for eligibility, provided consent to participate in the study, and were randomized into either the two-way video or face-to-face group. One hundred five patients were allocated to the two-way video group (53 with a depressive disorder, 34 with an anxiety disorder, and 18 with OCD) and 94 patients were allocated to the face-to-face group (45 with a depressive disorder, 32 with an anxiety disorder, and 17 with OCD). Seven patients in the two-way video group discontinued intervention due to the following reasons: withdrawal of consent (n = 1), failure to meet the inclusion criteria (n = 1), adverse event (n = 1), patient request (n = 1), time commitment challenges (n = 1), and other reasons (n = 2). Four patients in the face-to-face group discontinued intervention due to the following reasons: loss to follow-up (n = 3) and withdrawal of consent (n = 1). The CONSORT (Consolidated Standards of Reporting Trials) diagram for this study is presented in the Fig. 1
Fig. 1
Table 1 presents the baseline demographic and clinical characteristics of the participants. No significant differences were noted between the two groups, including in terms of age, sex, disease duration, and total treatment period. In the two-way video group, the average percentage of two-way video use after baseline was 76.95% ± 22.93%. Of these, there were a total of 38 patients who utilized two-way video 100% of the time.
| Two-way video (n = 98) | Face-to-face (n = 90) | P-value | |
|---|---|---|---|
| Age (years) | 39.7 ± 11.9 | 40.7 ± 11.9 | 0.55 |
| Sex (female), n (%) | 48 (49.0) | 46 (51.1) | 0.88 |
| Psychiatric history | |||
| Duration since the first episode (months) | 105.1 ± 90.3 | 105.7 ± 99.5 | 0.97 |
| Duration since current episode (months) | 82.7 ± 78.5 | 72.2 ± 83.2 | 0.38 |
| Duration of total treatment (months) | 78.5 ± 73.6 | 79.9 ± 78.9 | 0.90 |
| Duration of treatment by a current physician | 38.0 ± 45.5 | 35.4 ± 37.5 | 0.66 |
| Diagnosis, n (%) | |||
| Depressive disorder | 50 (51.0) | 42 (46.7) | |
| Anxiety disorder | 30 (30.6) | 31 (34.4) | |
| OCD and related disorders | 18 (18.4) | 17 (18.9) | |
| Measures | |||
| HAMD-17 (depressive disorders only) | 8.3 ± 1.5 | 6.1 ± 1.5 | 0.08 |
| HAMA (anxiety disorders only) | 10.2 ± 1.9 | 11.5 ± 2.1 | 0.42 |
| YBOCS (OCD and related disorders only) | 14.1 ± 1.8 | 15.6 ± 2.1 | 0.51 |
- Data are mean ± SD unless otherwise indicated.
- EQ-5D, EuroQol 5 Dimension; HAMA, Hamilton Anxiety Rating Scale; HAMD, Hamilton Depression Rating Scale; OCD, obsessive-compulsive disorder; SF-36 MCS, 36-Item Short-Form Health Survey Mental Component Summary; YBOCS, Yale-Brown Obsessive Compulsive Scale.
Primary Outcome
The SF-36 MCS scores at week 24 in the two-way video and face-to-face groups were 48.50 ± 9.57 and 46.68 ± 10.58, respectively. The criteria for noninferiority, for which the margin was set as −5.0, were met (mean between-treatment group difference, 1.82; 95% confidence interval, −1.12 to 4.77; P < 0.0001) (Table 2). With regard to sensitivity analysis based on the PPS, the criteria for noninferiority were also met, with a mean between-treatment group difference of 1.90 points (48.50 ± 9.57 vs 46.60 ± 10.62 in the two-way video and face-to-face groups, respectively [95% confidence interval, −1.06 to 4.86]; P < 0.0001). As a post hoc analysis, only the patients who utilized two-way video 100% of the time were extracted. The SF-36 MCS scores at week 24 for that group were 47.58 ± 9.16, and the criteria for noninferiority in the face-to-face group were also met (Supplementary Table S1).
| Weeks | Two-way video | Face-to-face | Difference in mean (95% CI) | P-value | |
|---|---|---|---|---|---|
| Noninferiority tests (margin: −5) | |||||
| SF-36 MCS (FAS) | 24 | 48.50 ± 0.57 (n = 96) | 46.68 ± 10.58 (n = 86) | 1.82 (−1.12 to 4.77) | <0.0001 |
| MMRM | |||||
| SF-36 MCS (FAS) | 0 (baseline) | 47.96 ± 1.88 (n = 98) | 45.44 ± 1.92 (n = 90) | 2.52 (−0.13 to 5.17) | 0.06 |
| 12 | 47.55 ± 1.88 (n = 98) | 46.36 ± 1.92 (n = 89) | 1.19 (−1.47 to 3.84) | 0.38 | |
| 24 | 48.97 ± 1.90 (n = 96) | 47.50 ± 1.95 (n = 86) | 1.48 (−1.33 to 4.29) | 0.30 | |
| Noninferiority tests (margin: −5) | |||||
| SF-36 MCS (PPS) | 24 | 48.50 ± 9.57 (n = 96) | 46.60 ± 10.62 (n = 85) | 1.90 (−1.06 to 4.86) | <0.0001 |
| MMRM | |||||
| SF-36 MCS (PPS) | 0 (baseline) | 48.26 ± 1.88 (n = 96) | 46.13 ± 1.92 (n = 85) | 2.13 (−0.05 to 4.83) | 0.12 |
| 12 | 47.89 ± 1.88 (n = 96) | 46.71 ± 1.92 (n = 85) | 1.18 (−1.50 to 3.86) | 0.39 | |
| 24 | 49.30 ± 1.91 (n = 96) | 47.75 ± 1.95 (n = 85) | 1.55 (−1.27 to 4.37) | 0.28 | |
- Data are mean ± SD.
- CI, confidence interval; FAS, full analysis set; MMRM, mixed-effect model for repeated measure; PPS, per protocol set; SF-36 MCS, 36-Item Short-Form Health Survey Mental Component Summary.
Secondary Outcomes
There was no significant difference between the two-way video and face-to-face groups with regard to the SF-36 MCS score at week 12 (P = 0.38) or the SF-36 PCS scores at weeks 12 and 24 (Table 3). The eight SF-36 subscale scores at weeks 12 and 24 also did not differ significantly between the two groups. All-cause discontinuation was found in one case (1%) in the two-way video group and one case (1.1%) in the face-to-face group, with no statistically significant difference (P = 0.95). There was no significant difference in the WAI or CSQ scores (P = 0.25) between the two groups at weeks 12 and 24. There were four cases of adverse events in the two-way video group and five in the face-to-face group, with no significant difference between the groups with respect to the risk of adverse events (P = 0.7). Most of the adverse events consisted of physical illnesses such as cholecystitis, posterior longitudinal ligament ossification, and abdominal pain, which were not related to the intervention.
| Weeks | Two-way video | Face-to-face | Difference in mean (95% CI) | P-value | |
|---|---|---|---|---|---|
| SF-36 PCS | 0 (baseline) | 48.19 ± 1.72 (n = 98) | 48.35 ± 1.77 (n = 90) | −0.17 (−2.63 to 2.30) | 0.90 |
| 12 | 47.92 ± 1.71 (n = 98) | 47.06 ± 1.75 (n = 89) | 0.86 (−1.51 to 3.24) | 0.48 | |
| 24 | 49.48 ± 1.75 (n = 96) | 47.49 ± 1.79 (n = 86) | 2.00 (−0.60 to 4.60) | 0.13 | |
| WAI | 12 | 71.7 ± 2.8 (n = 98) | 68.1 ± 2.9 (n = 87) | 3.6 (−0.1 to 7.4) | 0.06 |
| 24 | 71.6 ± 2.8 (n = 96) | 69.5 ± 2.9 (n = 85) | 2.1 (−1.9 to 6.0) | 0.31 | |
| CSQ | 12 | 26.9 ± 0.9 (n = 98) | 26.2 ± 0.9 (n = 87) | 0.7 (−0.5 to 2.0) | 0.24 |
| 24 | 27.3 ± 0.9 (n = 96) | 26.5 ± 1.0 (n = 85) | 0.8 (−0.6 to 2.1) | 0.25 | |
| EQ-5D | 0 (baseline) | 0.811 ± 0.029 (n = 98) | 0.823 ± 0.030 (n = 89) | −0.011 (−0.052 to 0.029) | 0.58 |
| 12 | 0.807 ± 0.030 (n = 98) | 0.822 ± 0.031 (n = 88) | −0.014 (−0.062 to 0.033) | 0.55 | |
| 24 | 0.828 ± 0.029 (n = 96) | 0.829 ± 0.030 (n = 86) | −0.001 (−0.0045 to 0.043) | 0.97 | |
| Degree of anxiety about COVID-19 (VAS) | 0 (baseline) | 46.3 ± 5.3 (n = 98) | 49.0 ± 5.5 (n = 89) | −2.7 (−10.7 to 5.2) | 0.50 |
| 12 | 44.3 ± 5.4 (n = 98) | 47.0 ± 5.5 (n = 87) | −2.8 (−11.0 to 5.4) | 0.50 | |
| 24 | 43.2 ± 5.3 (n = 96) | 44.2 ± 5.4 (n = 86) | −0.9 (−8.7 to 6.9) | 0.82 | |
| HAMD-17 (depressive disorders only) | 0 (baseline) | 8.3 ± 1.5 (n = 50) | 6.1 ± 1.5 (n = 42) | 2.2 (−0.3 to 4.6) | 0.09 |
| 12 | 8.8 ± 1.5 (n = 50) | 6.0 ± 1.6 (n = 41) | 2.8 (0.2 to 5.4) | 0.03 | |
| 24 | 7.9 ± 1.6 (n = 48) | 5.9 ± 1.7 (n = 39) | 2.0 (−1.0 to 5.0) | 0.18 | |
| HAMA (anxiety disorders only) | 0 (baseline) | 10.2 ± 1.9 (n = 30) | 11.5 ± 2.1 (n = 31) | −1.3 (−4.7 to 2.0) | 0.42 |
| 12 | 9.7 ± 2.0 (n = 29) | 12.2 ± 2.1 (n = 31) | −2.5 (−6.1 to 1.0) | 0.15 | |
| 24 | 8.7 ± 1.8 (n = 29) | 9.0 ± 1.9 (n = 30) | −0.3 (−2.8 to 2.2) | 0.81 | |
| YBOCS (OCD and related disorders only) | 0 (baseline) | 14.1 ± 1.8 (n = 18) | 15.6 ± 2.1 (n = 17) | −1.5 (−6.2 to 3.2) | 0.52 |
| 12 | 14.0 ± 1.8 (n = 18) | 15.0 ± 2.1 (n = 16) | −1.0 (−5.8 to 3.8) | 0.67 | |
| 24 | 12.9 ± 1.8 (n = 18) | 14.0 ± 2.1 (n = 16) | −1.1 (−5.7 to 3.5) | 0.62 |
- Data are mean ± SD.
- CI, confidence interval; CSQ, Client Satisfaction Questionnaire; EQ-5D, EuroQol 5 Dimension; HAMA, Hamilton Anxiety Rating Scale; HAMD, Hamilton Depression Rating Scale; OCD, obsessive-compulsive disorder; SF-36 PCS, 36-Item Short-Form Health Survey Physical Component Summary; VAS, visual analog scale; WAI, Working Alliance Inventory; YBOCS, Yale-Brown Obsessive Compulsive Scale.
Regarding the time required for hospital visits, the two-way video group spent less time than the face-to-face group (42.9 ± 40.8 min in the two-way video group and 79.2 ± 61.6 min in the face-to-face group, respectively; P < 0.001). Regarding the costs incurred for hospital visits (including communication costs for two-way video), nonparametric tests were used because there was a large variation in costs and high outliers due to the variety of forms of hospital visits and the readiness of the communication environment. As a result, the two-way video group paid less than the face-to-face group (median, 168.9 [interquartile range, 0–793.3] Japanese yen in the two-way video group and 500.0 [interquartile range, 140.0–1266.7] yen for the face-to-face group, respectively, P = 0.01) (Table 4). The number of work days missed for treatment sessions averaged 1.5 ± 2.5 in the two-way video group and 2.6 ± 7.1 in the face-to-face group, with no significant difference between the two groups (P = 0.15). The EQ-5D scores did not differ significantly between the two groups (P = 0.97).
| Two-way video (n = 98) | Face-to-face (n = 90) | P-value | |
|---|---|---|---|
| Number of hospital visit during the study period | |||
| Mean ± SD | 6.3 ± 2.8 | 5.7 ± 2.5 | 0.12 |
| 95% CI | 5.7–6.8 | 2.1–6.2 | |
| Time required per hospital visit (minutes) | |||
| Mean ± SD | 42.9 ± 40.8 | 79.2 ± 61.6 | <0.0001 |
| 95% CI | 34.7–51.1 | 66.3–92.1 | |
| Cost per hospital visit (Japanese yen) | |||
| Median | 168.9 | 500.0 | 0.0104 |
| IQR | 0.0–793.3 | 140–1266.7 | |
| Number of work days missed for hospital visits | |||
| Mean ± SD | 1.5 ± 2.5 | 2.6 ± 7.1 | 0.15 |
| 95% CI | 1.1–2.0 | 1.2–4.1 | |
- CI, confidence interval; IQR, interquartile range.
There were no significant differences in disease severity between the two groups at 12 and 24 weeks as assessed based on the HAMD, HAMA, and YBOCS scores, except that the HAMD score at 12 weeks in the two-way video group was higher than that in the face-to-face group (P = 0.03). Data regarding other secondary outcomes, including the changes from the respective baseline values in the two groups, are presented in the Supplementary Tables S2 and S3.
Discussion
Here, we report the results of a large-scale, long-term study comparing two-way video and face-to-face treatment in the real-world clinical setting. The most important feature of this pragmatic trial is that it adapted relatively broad inclusion criteria, namely depressive disorder, anxiety disorders, OCD and related disorders, and examined the effect of long-term treatment over 6 months. The treatment provided was the same as that in general outpatient care with no restrictions on the number of visits or treatment content. In other words, the psychiatrists in the study provided the best insurance-covered treatment they considered appropriate in a two-way video or face-to-face setting. In addition, the study was validated in a modern telemedicine setting, where patients easily accessed and received treatment from a psychiatrist at home or in the office using smartphones, tablets, or personal computers. Most of the telemedicine RCTs to date have been relatively short-term trials for a single disorder, often with some form of specific treatment. To our knowledge, there are very few pragmatic RCTs validating two-way video treatment that incorporate multiple psychiatric disorders.9 In addition, each trial design has limitations, such as a limited follow-up period of less than 6 months11 or, in the case of long-term follow-up studies, the number of participants is limited to a few dozen12 to 140.10 Through the COVID-19 pandemic, two-way video appointments became established as a common method of psychiatric care delivery. Patients can now easily see their psychiatrists remotely using their smartphones and other devices from home. The fact that two-way video was determined noninferior to face-to-face treatment in this study is an important finding, given that this type of health care will continue to be used around the world.
Another important aspect of this trial is that it is the first pragmatic two-way video trial in Japan, a country with universal health insurance where people can choose their preferred medical facilities and receive medical care at a relatively low cost. The flip side to low health care prices is the extremely busy treatment environment where health care providers have to see many patients in a short period of time. There was some concern that two-way video treatment would be difficult to implement in such an environment. Since Japan is considered a relatively restrictive country for telemedicine,3, 13 it was important to verify that two-way video treatment is equally effective in light of existing individual and cultural regulatory environments in order to appropriately promote its use.
The results of this study are consistent with those of previous meta-analyses that have reported comparable efficacies of two-way video and face-to-face treatment.9, 37, 38 In the evaluation of the primary outcome, treatment via two-way video was noninferior to face-to-face treatment. Although the difference did not reach significance, numerically, the SF-36 MCS was higher in the two-way video group at 24 weeks. There were no significant differences between the two groups on the gold standard rating scales for each disease. The only exception was that the HAMD score at 12 weeks for patients with depression was significantly higher in the two-way video group than in the face-to-face group. The reason for this is unclear, but both psychiatrists and patients may have been unfamiliar with two-way video and may have had some difficulty with the initial treatment. However, the baseline HAMD score of the face-to-face group was originally higher than that of the two-way video group at the trend level, and the difference may have been significant only incidentally. At the final 24-week time point, the significant difference between the two groups disappeared. This study had a low dropout rate, and there were no differences between the two groups in terms of measures related to treatment alliance, such as the WAI and CSQ scores. Reflecting such patients’ positive attitudes toward two-way video, the percentage of two-way video use was relatively high. In this study, the two-way video group was supposed to use two-way video for more than 50% of visits, but the average rate of two-way video use after baseline was approximately 77%. Furthermore, approximately one-third of patients in the two-way video group received only two-way video treatment, demonstrating the noninferiority of two-way video compared with face-to-face treatment in this group as well. At the same time, the reasons for not using two-way video 100% of the time should have been examined in detail. This study did not collect detailed data on the reasons why the two-way video group chose face-to-face care for some visits, and this is an issue for future studies.
As expected, two-way video was also found to reduce the burden of hospital visits for the patients; patients in the two-way video group spent less time in hospital visits and had fewer expenses than patients in the face-to-face group. The value that telemedicine can provide to patients is significant, not to mention the time and cost-savings associated with hospital visits. These include the provision of medical care in medically underserved areas, access to highly specialized psychiatrists, and easier access for patients who may have difficulty seeing a psychiatrist due to symptoms and/or stigma.
The following limitations of this study should be noted. First, we targeted only three disorder groups, namely depressive disorders, anxiety disorders, and OCD and related disorders. Although the three disorder groups considered in this study can be assumed to represent a large number of patients in psychiatric outpatient clinics, they do not cover all psychiatric disorders, such as schizophrenia, bipolar disorder, substance-related disorder, and neurodevelopmental disorder. Thus, a comparison between two-way video and face-to-face treatment for diseases not covered in this study remains a subject for future research. Second, it was not possible to blind psychiatrists or patients in this study comparing two-way video to face-to-face. However, our study implemented a centralized rating, and we were able to implement blinding of the raters who performed the HAMD, HAMA, and YBOCS assessments. This is part of the design advantage of this study over other studies, but, despite this, it cannot be ruled out that the fact that physicians as well as patients knew the assignments may have worked in favor of telemedicine when, for example, expectations for telemedicine were high. Third, although the present study followed patients for a relatively long period (6 months), there is still room for further evaluation of the effects over even longer follow-up periods, as psychiatric disorders often have a long course. While long-term use of two-way video is likely to reduce the financial burden on patients, it may take longer to establish a good rapport or reduce the quality of an established rapport, compared with face-to-face treatment.39 Future research should examine the usefulness of two-way video for longer periods of time and the desirable methods of operation.
Conclusion
The study showed that two-way video treatment over a 6-month period was no less effective than face-to-face treatment in patients with depressive disorders, anxiety disorders, OCD and related disorders. Many of the patients accessed their psychiatrists from home using smartphones, which is meaningful in that the study demonstrated the effectiveness of a modern form of telemedicine. In addition, this was the first RCT conducted in a real-world clinical setting in Japan, and the results indicate that two-way video is a practical option in Japan and can be used equally with face-to-face treatment. In future studies, longer follow-up and further validation of the usefulness of disease-specific two-way video treatment will be desirable.
Acknowledgments
The authors are grateful to the other members of J-PROTECT (Japanese Project for Telepsychiatry Evaluation During COVID-19: Treatment Comparison Trial) collaborators: Dr Kazunari Yoshida, Dr Toshiro Horigome, Dr Mayu Fujikawa, Ms Kumiko Hiza, Ms Hiromi Mikami, M. Saki Hattori, Mr Masao Takaishi, Mr Satoshi Tsujimura, Mr Hajime Tamura, Ms Junko Suzuki, Ms Keiko Komiyada, Dr Anri Watanabe, Dr Yoshihiro Matsumoto, Ms Satoko Kimura, Ms Haruka Okamoto, Dr Kyosuke Sawada, Ms Yuka Oba, Mr Satoshi Tsujimura, Ms Shii Sagae, Mr Kiyoji Nagao, Mr Ryuhei Terashi, Ms Sumako Onishi, Ms Mayumi Hiruma, Ms Junko Kannari, Ms Kanako Sasao, Ms Ayumi Konishi, Ms Nobuko Haga, Mr Nobuhiko Noguchi, Mr Kosuke Hino, Mr Yuya Igarashi, and Dr Takaharu Azekawa for their support with designing the study, data collection, and management protocols used in this study. The authors would like to express their sincere gratitude to the late Dr. Yoshinori Watanabe for his invaluable help in conducting the study.
Disclosure statement
The authors declare the following financial interests/personal relationships that may be considered as potential competing interests: Taishiro Kishimoto has received grants from Sumitomo, Otsuka, and Micin; royalties or licenses from Sumitomo and FRONTEO; consulting fees from TechDoctor and FRONTEO; speaker’s honoraria from Sumitomo, Boehringer Ingelheim, Takeda, Astellas, Meiji Seika, and Janssen; and stock from i2medical and TechDoctor. Takeshi Asami has received speaker’s honoraria from Otsuka, Meiji Seika, Sumitomo, Lundbeck, Yoshitomiyakuhin, Eisai, MSD, and Takeda; and scholarship grants from Otsuka and PDRadiopharma. Akira Suda has received grants from Otsuka; speaker’s honoraria from Eisai, Lundbeck, Meiji Seika, MSD K.K., and Takeda. Toshiaki Kikuchi has received speaker’s honoraria from Takeda, Lundbeck, Viatris, Sumitomo, and MSD. Mitsuhiro Sado has received research support from the Japan Research Institute; royalties from Igaku-Shoin, Seiwa-Shoten, Sogensha, and Kongo-Shuppan; and lecture fees from Mochida, Takeda, Sumitomo, and Viatris. Masaru Mimura has received grants from Daiichi Sankyo, Eisai, FRONTEO, Shionogi, Takeda, Tanabe Mitsubishi, and Tsumura; speaker’s honoraria from Biogen Japan, Byer Pharmaceutical, Daiichi Sankyo, Sumitomo Pharma, Demant Japan, Eisai, Eli Lilly, Fuji Film RI Pharma, Hisamitsu, H.U. Frontier, Janssen, Mochida, MSD, Mylan EPD, Nippon Chemiphar, Novartis, Ono Yakuhin, Otsuka, Pfizer, Shionogi, Takeda, Teijin, and Viatris. Kengo Nagashima has received consulting fees from SENJU, Toray, and Kowa. Takashi Nakamae has received a telemedicine system from Medley, and speaker’s honoraria from AstraZeneca and Lundbeck. Yoshinari Abe has received a telemedicine system from Medley. Tetsufumi Kanazawa has received speaker’s honoraria from EA Pharma Co., Ltd., Eisai Co., Ltd., Janssen Pharmaceutical K.K., Lundbeck, Meiji Seika Pharma Co., Ltd., MSD K.K., Otsuka Pharmaceutical Co., Ltd., Pfizer Japan Inc., Sumitomo Dainippon Pharma Co., Ltd., Takeda Pharmaceutical Company Limited, and Viatris Inc.; received grants from Eisai Co., Ltd., Otsuka Pharmaceutical Co., Ltd., and Sumitomo Dainippon Pharma Co., Ltd.; and received a supervision fee for advertising materials from Teijin-Pharma Co., Ltd. Hiroaki Tomita has received grants from Daiichi-Sankyo, Eisai, Otsuka, and Sumitomo Pharma; speaker’s honoraria from Daiichi-Sankyo, EA Pharma, Eisai, Janssen, Lundbeck, Meiji Seika, Mochida, MSD, Mylan EPD, Otsuka, Pfizer, Sumitomo Pharma, Takeda, and Viatris. Hiroshi Kimura has received speaker’s honoraria from Otsuka, Meiji Seika, Janssen, Teijin, Takeda, MSD, and Sumitomo pharma. Aiko Sato has received speaker’s honoraria from Takeda and Lundbeck. Naoya Sugiyama has received consulting fees from Hisamitsu, and speaker’s honoraria from Sumitomo, Otsuka, Janssen, UCB, Shionogi, EA, Eisai, and Viatris. Masaru Mimura is a vice editor in chief of Psychiatry and Clinical Neurosciences and a co-author of this article. Hiroaki Tomita is an editorial board member of Psychiatry and Clinical Neurosciences and a co-author of this article. To minimize bias, they were excluded from all editorial decision-making related to the acceptance of this article for publication.
Author contributions
All authors were involved in designing the study. Akitoyo Hishimoto, Takeshi Asami, Akira Suda, Shogyoku Bun, Toshiaki Kikuchi, Akihiro Takamiya, Takashi Nakamae, Yoshinari Abe, Tetsufumi Kanazawa, Yasuo Kawabata, Hiroaki Tomita, Koichi Abe, Seiji Hongo, Hiroshi Kimura, Aiko Sato, Hisashi Kida, Kei Sakuma, Michitaka Funayama, Naoya Sugiyama, Kousuke Hino, Toru Amagai, Maki Takamiya, Hideyuki Kodama, Kenichi Goto, and Shuichiro Fujiwara contributed to the recruiting of patients and collecting the data. Yasunori Sato, Ryo Takemura, and Kengo Nagashima analyzed the data. Taishiro Kishimoto and Shotaro Kinoshita prepared the original draft. All authors contributed to the final draft. All authors have read and approved the final manuscript.
Funding Information
This research is supported by the Japan Agency for Medical Research and Development (AMED) under grant number JP22dk0307098. The funding source did not participate in the design of this study and did not have any hand in the study’s execution, analyses, or submission of results.
