A mixed-methods study design with quantitative and qualitative descriptive data [19] was used to provide a deep understanding of the studied phenomena and the descriptive approach. In this study, the quantitative data were used to describe how adolescents used the e-diary, while the qualitative data were used to identify the contents adolescents discussed in their e-diaries. We assumed that using quantitative and qualitative approaches together will allow complementary information to arise—the quantitative data augmented the qualitative data, which were analyzed with the method of content analysis [20]. The criteria of Tong et al [21] were followed for reporting the qualitative data.

Participants (N=89) in this study were adolescents who were between 15 and 17 years of age at the time of obtaining informed consent. This study is a part of a two-arm randomized controlled trial (Depis.Net), and participants of this paper were adolescents who were randomized solely into an intervention group. The inclusion criteria for the randomized controlled trial were as follows: those aged 15-17 (at the time of recruitment) years and able to speak Finnish. Adolescents with serious mental disorders, such as psychotic depression, bipolar disorder, substance abuse, or a primary eating disorder, were excluded; those admitted to psychiatric hospital wards or involved in a brief intervention at an outpatient clinic (≤3 appointments) were also excluded. Participants in the intervention group were given personal identifiers (IDs) for logging into the Depis.Net program, which included an e-diary. They could use the program from any computer with internet access. We included in the analysis all potential participants of the intervention group who had an opportunity to use the e-diary.

Participants’ information, including their age, gender, education, previous use of mental health services, previous depression, and medication use during the last 6 months, was collected from medical records by nurses at the outpatient clinics. Outpatient clinic nurses assessed the overall severity of disturbance using the Children’s Global Assessment Scale (C-GAS) [22]. Depressive symptoms were identified using the Beck Depression Inventory (BDI-21) [23], which is a self-assessment instrument that has been validated for adolescents [24].

The Depis.Net intervention took 6 weeks—a 1-week introduction and 5 weeks during which themes changed weekly (well-being, home and family, adolescents’ rights and responsibilities, adolescent depression, and treatment of adolescents’ depression). The intervention also included health information, self-monitoring, and the use of e-diary where adolescents could reflect on their own life situation. The e-diary included technical instructions as well as personal writings that other adolescents could not see. Adolescents were given ideas related to the weekly theme, which aimed to support them in the use of the e-diary. They were asked to describe their current thoughts, feelings, and moods at least once a week, whenever they had time and the willingness to do so. They were also encouraged to use extra characters in the text, such as emoticons, pictures, song lyrics, or comics. They had access to material from all weeks and the e-diary during the 6-week program. It was possible to use the e-diary at any time within any week because the weeks did not close after they had been activated.

To support adolescents in their use of the e-diary, supportive short message service (SMS) text messages were sent every Monday. They informed participants of the theme of that week. Feedback based on previous week was also offered to participants. In addition, those who had not used the e-diary by Thursday received an additional SMS text message to encourage them to write in it. Those who had used the e-diary during the week received feedback from the trained tutor through the program. Trained tutors read adolescents’ e-diary entries daily. If any severe concern was identified in an adolescent’s text, such as a threat of suicide, the trained tutors informed it to the responsible nurse or physician.

Descriptive quantitative statistical methods (frequencies, averages) were used [25], and qualitative data were analyzed using content analysis [20]. First, using descriptive statistics, the content of e-diaries was described by how often and how much adolescents wrote and used pictures (eg, emoticons, combination of signs, photos, or drawings). The number of words was calculated, and the pictures were checked. For example, I should do lots of schoolwork at some point (I hate school >:/ )....i.e. it seems in fact, that I have quite a lot of stress! 8O So typical >_<“” included 27 words and 3 pictures (>:/, 8O, >_<“”). After this, an Excel table was generated to gather detailed information from 1 week at a time. The table was used to obtain a general overview of the 6-week period. The e-diary program made one automatic timestamp when an adolescent logged in. Automatic timestamps and the time of logging in were calculated. Background information on adolescents’ baseline characteristics and BDI-21 and C-GAS scores were analyzed by frequencies and percentages.

Then, data were analyzed using inductive content analysis made from the manifest content. The sentence was used as a unit of analysis because, with smaller (word) or larger analysis units, the overall vision or some essential information could go unnoticed. For the inductive content analysis, printed e-diary entries were read a number of times to understand their content clearly. After this, the sentences that corresponded to the question “what issues have been addressed in the diaries” were manually highlighted [20]. These sentences were transferred into a separate document and categorized by relevant phrases [20,26], which were coded (a coding phase) [20]. The content of the e-diaries was then exported to the qualitative data analysis software, NVivo, where the coding of the sentences continued [20,27]. The codes and their content were compared to identify differences and similarities; subcategories were further formed. In the next step, these subcategories were abstracted and organized into categories based on the similarity of their content. A total of 7 subcategories were formed, which were further combined into 3 categories and named through reflection and discussions with the cowriters. [20]

To increase the validity of the analysis during the analysis process, the data were coded independently by 2 coders (KM and MK) and categories were formed by 3 authors (KM, MK, and MA) [21]. Comments and thoughts were written in a notebook, which aimed to make the internal thinking visible. The notes also allowed the ideas related to the interpretation of the data to be revisited. Conclusions in all the phases were compared with the original data [26].

The ethical approval was received from the ethics committee (#R08075H). Hospital administrators granted permissions to conduct the study. Good scientific practices [28] and the principles for medical research and legislation [29] were adhered to. The research assistants provided verbal and written information about the purpose and process of the study to adolescents. Adolescents’ participation in the trial was voluntary, and they gave written informed consent. It was possible to withdraw from the study without giving any reason, and it did not affect treatment. If there were severe concerns, such as suicidal ideation, a research assistant informed the responsible nurse or physician to ensure the safety of the adolescent. Anonymity was ensured by having personal access to the program only with a username and password. Unidentifiable codes (IDs) were used, and only the research assistants had access to the program. Anonymity was guaranteed in all phases of the study and reporting of the results. Printed e-diaries included only the unidentifiable IDs. They were stored in a locked space, and the e-diary texts were protected by passwords.

Of the 89 participants, 47 used the e-diary. Almost all e-diary users were females. The majority of them were in high school compared to nonusers, many of whom were in a vocational school. More e-diary users had had previous episodes of depression. They had also used mental health services and medication more frequently than nonusers. According to BDI-21, more e-diary users had moderate or severe depression symptoms than nonusers. However, nonusers had slightly more serious problems in functioning (C-GAS) than did e-diary users. Characteristics of the e-diary users and nonusers are described in Table 1.

Overall, 53% (47/89) of the participants used the e-diary during data collection. Most of the adolescents (16/47, 34%) logged into the e-diary within 2 weeks, and 21% (10/47) logged into the e-diary at least once during each of the 5 weeks. The adolescents logged into the program a total of 147 times (range 1-13 times; median 2; mean 3.1, SD 2.3). Most adolescents (39/47, 83%) logged into the e-diary during the first week, and over half (27/47, 57%) of the e-diary logs were made during the 2 first weeks. In addition, 21% (10/47) of the adolescents relogged into their e-diary and modified or added their output later. Nearly all (22766/22855, 99.61%) of the e-diary logs were made using text form (words), while pictures were rarely used (89/22855, 0.39%). The most logged themes, based on the number of log-ins, were “well-being” and “adolescents rights and responsibilities” (Table 2). The number of words used ranged from 8 to 1442 per each log (mean 174.2, SD 204.4).

Sundays to Tuesdays were the most common days for writing in the e-diary, and evenings were the most common time of the day (Figure 1). Adolescents did not use the e-diary at all after 2 am. Logs made in the program at night time (from 23 pm to 2 am) or in the mornings (from 09 am to 12 pm) were also very limited.

In this study, 3 main categories were formed to represent the topics of the adolescents’ e-diaries—mental disorder, relationship, and identity (Table 3).

In this study, we assumed that e-diary writing might be a usable tool for reflecting thoughts for those who have depressive symptoms. Indeed, adolescents included different topics in their e-diaries. They described and discussed negative and positive events in their lives as well as difficulties and good experiences. They also pointed out their hopes and concerns [6] about their future and recovery from mental disorder. Despite the structure and content of the weekly program, adolescents had the chance to bring up concerns and other issues that were important to them, regardless of the week session. We can therefore assume that the instruction of the e-diary supported, but did not restrict, adolescents in dealing with important issues in e-diaries.

On the other hand, 47% (42/89) of the participants did not use the e-diary at all. About one-third (19/47, 40%) logged into the e-diary each week. Välimäki et al [7], in their systematic review and meta-analysis, have identified high dropout rates in Web-based interventions for adolescents. Reasons for low usage in Web-based interventions remain unknown. First, we can assume that adolescents simply may not like to use Web-based interventions because they are out of date or they do not meet their expectations. Second, there is some evidence that young girls in particular prefer to use the internet and smartphone apps for social communication and entertainment but not for purposes related to mental health [9]. Third, adolescents who did not log into the e-diary may have been reluctant to do so if they did not experience any advantage in writing or if they had a fear of disclosing painful or difficult topics [11]. Fourth, as shown in a previous study [12], adolescents with depression may also have difficulties with motivation and concentration, and they may have found Web-based intervention rather demanding. On the other hand, we found in this study that e-diary users more often had a history of depression and previous experience of mental health care services, and their depressive symptoms were more severe compared with those of nonusers. We can therefore assume that, despite burdening depressive symptoms, young people may find an e-diary useful.

We still found that boys were in a clear minority (2/47, 4%) when it came to writing in the e-diary. This is surprising because boys are active users of information technology. On the other hand, diary keeping is a more common habit for girls, especially at the age of 14-15 years [15]. To satisfy boys’ needs for technology use, more action and gaming elements might be more usable for young male participants. We assumed at the beginning of the study that adolescents might be reluctant to continue the intervention. Therefore, we sent them reminder SMS text messages and supportive feedback to encourage and motivate their program use [10]. Despite the SMS text message reminder, we found that both participation and e-diary entries decreased after the first week. We can, therefore, assume that SMS text messages did not have much value in supporting adolescents’ participation, although a more recent study with adults has showed that participants preferred receiving SMS text message reminders on Mondays and that they preferred slightly humorous SMS text messages [30]. It is also possible that the feedback and SMS text messages were not supportive enough. Gender may also be one explanation since intervention dropouts have previously more often been among women rather than men [31]. While participants’ engagement is an important factor for Web-based interventions [23], more studies are still needed to determine whether participation in electronic health applications could be increased [22,24,32]. In future, it would be important to explore whether enhancing real-time support [23] can increase engagement in the Web-based interventions.

This study has limitations that should be taken into consideration. First, the data were limited because the number of nonusers increased during the follow-up. Second, the data were collected during 2008-2010, and it could be questioned whether the data are still up to date. However, writing diary entries has a long history, and generally, as in this study, it includes descriptions or comments about events that occur or will occur at about the same time as they are written about [16]. Therefore, we can assume that these data provide us relevant knowledge about topics that concern the minds of adolescents.

On the other hand, technology use has been developed during the last decade, especially mobile technology. If a mobile version of the e-diary was used in this study, the number and time of logs could be different [33]. Third, e-diary entries were written in Finnish, and they represent the quite homogenous nature of Finnish culture. Therefore, the results of this study may not be culturally generalizable. Even so, writing in itself is beneficial for ethnically diverse groups and with a broad spectrum of people, including those who are less socially, academically, or physically advantaged [34]. In general, the results can therefore be valuable in different lingual and cultural regions when engaging Web-based interventions are being designed for young people with mental health problems.

Not all adolescents in this study were willing to use the e-diary to express their emotions or feelings [7]. We also found that adolescents who used the e-diary were more often girls and had current or previous experiences of depression and mental health care services. These adolescents were willing to describe their thoughts about mental health problems and their need for recovery. They also had the ability to reflect on their thoughts in writing, which was supported by specific themes. For future purposes, it is important to explore the reasons for nonuse and suspensions as well as how to support engagement in Web-based interventions.

We are already aware that Web-based interventions with structured monitoring are effective for adolescents with depression [7,9]. Based on this study, we can only assume that an e-diary with an element of expressive writing may have an additional value for some young people, especially for girls, to help them reflect their own thoughts and feelings about their burdening situation. On the other hand, the methods could not reach young boys, who still need more engaging technology where they can share their thoughts, feelings, and moods. In general, participant engagement is an important factor for Web-based interventions [23]. More studies are still needed to identify how participant engagement in Web-based interventions could be supported [22,24,32]. Results of this study can still be used to develop user-centered electronic health applications, which can allow users to express their own thoughts and feelings in ways other than systematic mood monitoring.