We recruited subjects from the day center and outpatient clinic of Roozbeh Psychiatric Hospital and some private clinics in Tehran. Our inclusion criteria were patients with a DSM-5 BD, diagnosed by a psychiatrist, age 15-65 years old, being able to access a computer and the internet, a minimum educational level of the 6th grade in primary school, no previously receipt or completion of any PE courses, and full or partial remission at the time of the study as indicated by their psychiatrist. We excluded patients with severe visual impairment, intellectual disabilities, or neurodevelopmental disorders.

The first session of each course (1st module) was held in person to familiarize each patient with other participants, the research team, and the overall process of the project and to obtain informed consent. Patients were informed regarding the importance of PE in the management of psychiatric conditions and more specifically BD. In each round of the interventions, we invited patients via a phone call, and we reminded them of the session time via text message 1 day before the meeting. In this first session, personal username and password were allocated to each patient to fill in their data. After this session, we invited the patients to participate in our study for 7 consecutive weeks and to expose themselves to the aforementioned PE video clips on the website. Figure 1 demonstrates the front page of the website where the patients were requested to enter their own username and password to access their individual profile. The Farsi language was selected for the website so a language barrier would not become an issue for our patients who all were Iranian.

Figure 2 shows the main page of the website, which showed 8 round circles at the top of the page as a manifestation of the course progress and the week the patients were in at any time, 2 main sections in the middle of the page, and some ancillary tabs on the right side.

One of the 2 main sections linked to the weekly streaming PE video clips (Figure 3; Multimedia Appendix 1), and the other section showed a weekly timetable of medications to take and allowed them to indicate their pattern of actual medication administration during the week (Figure 4).

Initially, this section was supposed to be a reminder to regulate the patients’ patterns of taking medication as well as give them a sense of mastery over their treatment plan; however, we thought a reminder might be negatively viewed by the patients as being controlled and might even decrease their cooperation. Therefore, this section remained optional, although strongly suggested, for the patients to use.

It is worth mentioning that during the intervention period, patients received other mental health services as well, which included medications and if indicated, psychotherapy and rehabilitation services provided by the hospital. For cases for whom we suspected a relapse, patients were referred for a medication review, and afterwards if the patient was still interested in participating in the study, we asked him or her to update his or her medication table on the website and continue the course. If the patient was not still interested, he or she was excluded from the study.

Our primary objective was to determine the level of feasibility and acceptability of this intervention. Feasibility outcomes were assessed via the recruitment rate, adherence to the website training course protocol, participant drop-out rate (participants who did not attend ≥4 weeks during their 8-week course). In addition, we conducted qualitative interviews with the team members and patients at the end of study to discuss obstacles and facilitators of the implementation platform and how to address them. Acceptability was assessed via the frequency and pattern of website use by participants, which was designed to be reported as an Excel file by the programmer of the website, as well as patients’ satisfaction rate, including whether the program was user-friendly, comprehensibility of the content, usefulness of the content, and usefulness of reminders. These variables were rated on a Likert scale, ranging from 1 to 5, for each item by the participants in terms of an interview conducted by 1 of the authors. SPSS v20 and Excel 2016 were used for statistical calculations and analyses as needed.

Among the patients referred to the day center, the recruitment rate was 3% (2/70). It was 63% (21/33) for the outpatient and private clinics.

We defined a dropout as being absent in 4 or more of the 8 sessions of the whole course. There was a 100% dropout rate during the first round of the study, which was held for the patients referred to the day center of Roozbeh Hospital, as none of the patients attended online meetings. Among all patients, 14 patients dropped out during the study. Of the 9 patients who completed the course, only 5 (22%) attended all the sessions, and 4 (17%) patients were absent in some sessions, but they completed the course. Furthermore, the level of commitment of all patients to participate in each week is illustrated in Table 2.

As mentioned, among 23 patients who entered the study, 14 dropped out, which accounts for 61% of the total sample who attended the first in-person meeting. Of these 14 participants, 50% (7/14) provided reasons that could be considered as a “lack of enough motivation.” Table 3 shows the possible reasons for nonadherence (dropout) that were obtained from interviews with the participants.

This includes the number of patient logins to the website and the entire time that was spent on the website by patients. As shown in Table 4, an average frequency of 1-2 logins per week, each lasting about 21 minutes, by patients who completed the course indicates the relative acceptability of this course, while for noncompleters, this variable is meaningfully lower (Table 4). While 14 (14/23, 60%) watched at least one PE video, only 5 (5/23, 22%) patients watched all of the videos.

To assess participants’ experiences using the online platform, we interviewed 13 participants who at least watched one PE video to gain answers to satisfaction questions, with responses rated from 1 to 5 on a Likert scale; the results are presented in Table 5. The average range of satisfaction with each item shows the relative acceptability of the intervention.

After holding a number of sessions, the research team reached consensus about influential factors in creating the website and sorted these factors into 2 categories: factors related to the individuals involved in the study and factors related to the intervention.

For the web design team, the main obstacle was that our research team did not have any previous experience in designing a website, so we had to outsource this stage to a reliable web design firm. Since the web design firm did not have any professional knowledge regarding our subject, we had to hold a number of sessions to align our expectations and knowledge with their experience in designing an online platform.

Regarding the research team, a limited number of professional staff and insufficient team members for implementing different stages of the study and follow-ups meant one person carried all these responsibilities, which could increase the number of errors.

For patients, first, we should state that a noticeable proportion of patients did not have enough motivation or skills to participate in the current study, and some patients thought that participating in this study may lead to a relapse of their symptoms. Several patients also preferred face-to-face sessions and did not identify enough with the proposition of online platforms.

First, we can point to the novel feature of this intervention in our setting, which can result in some trial and error. Furthermore, we faced some unfortunate incidences such as filtering the website in the middle of the study, which happened accidently by the government. The main limitation was the necessity to have 1 face-to-face session at the initiation of the study for patients. This obstacle manifested as the participation gap between the potential 45 volunteers at the preliminary invitation and the ~50% attendance (n=23) at the first in-patient session. Another limitation was specifying a 1-week period for each module; according to some patients, they had a problem attending some modules due to a lack of spare time, sickness, or travelling, and sometimes, they lost their motivation in continuing the course when they were absent for 1 module. Another issue was that watching a video clip or answering some questions did not take more than 1 or 2 hours per week, and this limited involvement during the week could lead to decreased motivation to continue participation. We could have tackled this issue by providing daily tasks for participants to complete and revive their motivation on a regular basis.

Our results are consistent with the outcome of the qualitative study conducted by Poole and colleagues [18]; they analyzed the feasibility, acceptability, and impact of an internet-based PE platform, “beating bipolar.” They found it feasible to deliver and acceptable for patients to use via a computer. This intervention had a satisfying impact on insight concerning illness, health behavior, personal habits, and positive attitude toward medication [18]. However, in their study, thematic analysis was employed to describe the participants’ experience. In the current study, we used a more quantitative approach. Hidalgo-Mazzei et al [12] designed a simple smartphone application to provide PE contents for patients with BD. Consistent with our results, this type of online intervention was an efficient approach to improve self-management by patients with BD. However, their small sample size (51 patients) made generalization difficult, as we can also say about this study [12].