Nonsuicidal self-injury (NSSI) is common in adolescence and associated with several long-term risks. NSSI refers to the “direct and deliberate destruction of body tissue in the absence of any observable intent to die” [1]. NSSI is a symptom of borderline personality disorder but is also present in the absence of borderline personality disorder [2,3], together with several other disorders [4,5], and in nonclinical populations [6]. NSSI is common, especially in midadolescence [7], with a pooled prevalence of approximately 17% [8]. In fact, the prevalence rate seems to have increased in recent years [9,10]. NSSI in adolescence is worrying, as it is associated with an increased risk of incidence of other psychopathology in youth [11] and several adverse outcomes in adulthood [12-14]. Even though there is an overlap between NSSI and suicide attempts [4,15-17], there are some key differences between the two, concerning intention, repetition, and medical lethality [1,17-19]. In addition, NSSI is one of the strongest risk factors for future suicide attempts [20], and cessation of self-injury in adolescence can potentially decrease the risk of future suicidal behavior [21].

Therefore, NSSI is serious and requires urgent treatment. Treatments that target NSSI specifically are needed because NSSI is a transdiagnostic phenomenon [2-5], and targeting NSSI can potentially prevent suicides [21,22]. Focusing on the maintenance factors of NSSI in treatment could be useful; the emotion regulation function (ie, that one engages in NSSI to decrease or escape aversive emotions) has been identified as the most common function of NSSI [23], and treatment for NSSI that targets emotion regulation difficulties seems to have beneficial effects on other health outcomes as well [24]. At present, no well-established treatment is available for treating NSSI [25-29], but dialectical behavior therapy (DBT) is probably efficacious for adolescents [25]. Nevertheless, challenges with accessibility and costs connected to DBT have been highlighted and briefer interventions have been called for [22].

Our research group has previously developed a brief emotion regulation individual therapy for adolescents (ERITA) [30] derived from emotion regulation group therapy (ERGT) [24,31] drawing from the principles of DBT and acceptance and commitment therapy. The aim of ERITA and ERGT is to decrease NSSI through learning and using new skills to regulate emotions. In contrast to ERGT, ERITA is a 12-week individual therapy that is specifically for adolescents and includes a parallel online course for caregivers. As ERITA seems promising [30], and as online treatment has several advantages, such as (1) the possibility of frequent contact with the therapist [32], (2) seems effective for several conditions among adolescents [33], and (3) can facilitate help-seeking as fear of stigmatization can be a barrier [34] and individuals with perceived stigmatized problems may prefer online treatment [35], our research group has further adapted ERITA [30] to an online version (ie, online ERITA). The quantitative results from the pilot trial indicate that online ERITA seems acceptable, feasible, and useful [36]. Online ERITA was, in a recent systematic review, the only online treatment tested for adolescents with NSSI [37].

Consequently, as research on online treatment for adolescents with NSSI is scarce, our knowledge of the experience of online treatment is limited [38]. Collaboration with patients has been suggested to be important when developing novel interventions for self-injury [29]. One way to engage patients in the development of interventions is through qualitative research methods. Qualitative evaluations of interventions allow for individual experiences, positive and negative feedback, and could together with quantitative evaluations contribute to a richer understanding of an intervention [38]. Regarding adolescents with self-injury, there is only 1 previous qualitative study investigating the experience of online interventions, specifically the experience of a mobile app as an adjunct to face-to-face treatment [39]. To our knowledge, no study has investigated the experiences of online treatment for adolescents with NSSI nor have the experiences of caregivers been explored. Involving caregivers in treatment for self-injury has been suggested [40], and increased knowledge of the experience of such involvement, also from the caregivers’ perspective, could help to further develop successful treatment for adolescents with NSSI. Overall, exploring individual experiences of novel treatments can potentially help improve existing treatments as well as increase our understanding of factors that are important in the treatment of adolescents with NSSI.

This study aims to explore the experiences of online ERITA for those with NSSI and their caregivers.

This qualitative study was part of a pilot trial of online ERITA [36]. This study was approved by the regional ethical review board of Sweden in November 2015 (reference number: 2015/1895-31/5). Furthermore, it is presented according to the Consolidated Criteria for Reporting Qualitative Research standards [41] for reporting qualitative research.

Families who had participated in the quantitative part of the pilot trial of online ERITA [36] were the population of interest. Inclusion criteria for adolescents in the pilot trial [36] were as follows: (1) being 13-17 years old; (2) fulfilling criteria for the proposed NSSI disorder [42]; (3) having engaged in ≥1 NSSI episode during the past month at the preassessment; (4) having stability of psychotropic medications (if any) for at least 2 months; and (5) having at least one caregiver committed to participate in the caregiver course. Exclusion criteria for adolescents in the pilot trial [36] were as follows: (1) severe suicidal ideation; (2) diagnosis of psychotic or bipolar I disorder, or substance dependence; (3) ongoing dialectical behavioral or mentalization-based therapy; and (4) insufficient understanding of the Swedish language. To participate in the qualitative part of the pilot trial, participants had to accept participation in interviews and terminate treatment before May 14, 2017.

Potentially eligible families were informed and prompted to participate in the interviews, either when meeting their therapist at posttreatment follow-up or by telephone. From the 17 eligible families, families were continuously selected for interviews as the data collection proceeded based on maximum variation sampling [43]; the target was to reach variability in the sample regarding age, gender, and past month NSSI frequency before and after treatment. One family was selected for interviews but later declined. 

Information power [44] was assessed continuously by analyzing the amount of new information from the interviews and the quality of data. From previous research, 6 to 12 interviews have shown to capture all themes and concepts needed to answer the research question [44-47]. We expected that more than 6 interviews per group (ie, separating adolescents and caregivers) were needed based on the variability within the adolescent group and the scope of the study. Conversely, sample specificity was good in relation to the research question. After interviews with 9 families (ie, 9 adolescents and 11 caregivers), there was redundancy in interview data; therefore, it was deemed that a data saturation point was reached [48]. The participant flow is shown in Figure 1.

In total, 9 adolescents and 11 caregivers participated (ie, 20 interviews in total); 7 families completed the interview in conjunction with the posttreatment follow-up, and 2 families completed the interview in conjunction with the 6-month follow-up. Participating adolescents were 14-17 years old (median 16; IQR 16-17), and 6 adolescents identified as female and 3 adolescents as nonbinary. Before the initiation of the treatment, past month NSSI frequency varied between 3 and 22 (median 10; IQR 9-15), and after the treatment, the past month NSSI frequency varied between 0 and 14 (median 2; IQR 1-8). All adolescents had completed all the 11 modules in the treatment, and all caregivers had completed all 6 modules in the caregiver course. Of the participating caregivers, 9 were mothers and 2 were fathers. Their ages ranged from 43-55 years (median 50; IQR 46-53).

Interviews were conducted between February and September 2017. The interviewers were 2 female psychology master students (OS and JS) educated in the theory and method of online ERITA, but they had no prior relation to the participants before the interviews. The interviews were conducted either at a child and adolescent mental health clinic (9 interviews), in the home of the families (6 interviews), or over telephone (5 interviews).

The interviews were conducted with the adolescents and caregivers in separate rooms, without nonparticipants. Before starting the interviews, the participants were informed about the purpose of the study; the occupation and experience of the interviewer; their right to terminate at any time; and how the material would be processed, stored, anonymized, and presented. Written informed consent (ie, caregivers and adolescents aged ≥15 years provided own written consent, and caregivers provided written consent on behalf of adolescents aged ≤14 years, according to Swedish law) was obtained before the interviews started. The duration of the interviews ranged from 13-41 minutes. To decrease the potential risk of the negative impact from talking about mental health issues and experiences of treatment, participants could choose the location of the interview, and all participants were asked before and after the interview if the situation induced distress. Interviewers were prepared to assist in case of reported or observed adverse events. However, no adverse events were reported before, during, or after the interviews.

Semistructured interviews were chosen based on the research questions and allowed for diverse experiences [49]. A total of 2 interview guides, 1 for the adolescents and 1 for the caregivers (Multimedia Appendices 1 and 2), were developed based on previous research (ie, on structure [49] and content [25,26,37,40]) and clinical experience and reviewed by the authors. The interview guide was pilot-tested on the first participant and evaluated with the supervisor (HE); no changes were deemed necessary. All interviews were audio recorded and later transcribed verbatim by the interviewers. The audio recording was erased immediately after the completion of the transcription.

Thematic analysis was used for data analysis, according to the steps recommended by Braun and Clarke [50]. Data analysis included a constant movement back and forth between the steps described below. In the first step, all transcripts were reread to gain familiarity with the data. In the second step, initial codes were generated; OS and JS coded 5 interviews together, and HE reviewed the coding as a verification step. After the verification step, OS coded the rest of the interviews and consulted HE throughout the process. All codes were reviewed several times by HE and OS. Examples of codes are listed in Table 1. In the third step, possible themes were investigated. Codes were clustered into categories and possible themes, and OS and HE reviewed different clustering opportunities and looked for negative cases. Data were split between adolescents and caregivers and compared in terms of both content and structure. In the fourth step, the themes were reviewed until they were deemed exclusive and fit the material. In the final step, the themes were labeled and the materials belonging to each theme were summarized. Discussions were ongoing in the research group throughout the process, and the final results were approved by all authors. Original quotes from the transcripts were chosen to show the relationship between the data and themes. For the qualitative analysis, the data program Open Code version 4.03 was used [51].

This study conveys the experiences of online ERITA for those with NSSI and their caregivers. The main findings were that online ERITA is experienced as offering several means of support (eg therapist, the mobile app) and that the caregiver involvement in this format is acceptable and beneficial. The perceived self-responsibility in online ERITA had both positive (ie, sense of empowerment) and negative aspects (ie, increased distress) connected to it. Several different skills seem to facilitate emotion regulation ability, and having access to such skills could hinder NSSI. Finally, decreased emotion regulation difficulties seemed to be an important treatment outcome for adolescents.

Interestingly, although several treatment effects were described, the most important changes identified by adolescents were improvements in emotion regulation and how that affected functions in daily life. This may not be surprising, given that NSSI per se is seldom the main concern for the individual but rather the emotional distress triggering NSSI [52]. Access to emotion regulation skills was mentioned as a reason for decreased NSSI, illustrating the potential mediating role that emotion regulation has in decreasing NSSI [30,36,53]. Furthermore, the results are in line with those of previous studies [24,30,36], indicating that interventions that are designed specifically for NSSI and target emotion regulation difficulties can affect other symptoms and the overall function. This is expected, as emotion regulation deficits seem to be important to a wide range of psychopathologies [54]. Altogether, these results highlight the importance of measuring several outcomes when evaluating treatment for NSSI to capture the processes through which the treatment works and the changes that are connected to decreased distress and increased function.

Our results indicate that adolescents used several different skills to regulate their emotions. Given that several different subscales of difficulties in emotion regulation have been strongly connected to NSSI (eg, nonacceptance of emotional responses, impulse control difficulties, and difficulties engaging goal-directed behavior) [23], several different emotion regulation skills could be useful to target these different difficulties. Therefore, it does not need to be concerning that no single emotion regulation skill was identified as of certain importance from the interviews with adolecents. Investigating whether it is possible to predict what skills the individual would benefit the most from, based on reports of difficulties in different emotion regulation subscales, could be an important next step to identify potent skills and components and further tailor the treatment to the individual. Online treatment could provide a useful framework for dismantling studies (eg, testing the effectiveness of a specific treatment component) [55] as treatment content could be controlled and modules can be easily added or removed.

In this study, therapist support was perceived as essential and therapeutic alliance was continually developed through the online format. The question has been raised if the online format can facilitate therapeutic alliance when treating self-injury [37]. Our results indicate that the online format facilitated self-disclosure, in line with the experiences of other pediatric clinical groups who have received online treatment [38]. In general, our results validate the importance of therapeutic alliance for this target group [26], which seems to be facilitated by both online, telephonic, and face-to-face contact in this context. The availability of frequent and flexible online and telephonic contact was important for the participants, in line with previous findings [25,27].

The importance of family involvement for this population [25,40] was also underscored from the results of this study. Involving caregivers in a parallel online course was highly appreciated by the caregivers and experienced as positive, or at least not negative, by the adolescents. Involving family members in this online format could be a valuable addition to traditional face-to-face family sessions. The importance of providing a sufficient dose of caregiver or family component has been emphasized [25], and the online format could be a time-efficient method for the therapist to increase such dose. Conversely, face-to-face meetings, as a complement to the online treatment, were requested by caregivers in this study, implying the importance of flexibility in format. Moreover, caregivers’ own skills training in emotion regulation was appreciated and helpful, and this is important as youth NSSI could affect the well-being of caregivers negatively [56].

Several advantages of the supplementary mobile app were described by the adolescents, both as an easy and flexible way to engage in treatment and a means of support. These results, together with the previous promising results on the utility of mobile apps for self-injury [39], indicate the usefulness of supplementary add-on formats for both online and face-to-face treatments. However, not all adolescents used the mobile app, and the reasons mentioned involved not needing it and technical issues. The technical issues are unfortunate but are difficult to avoid in complex apps. The report from some of the adolescents that the technical issues prevented them from using the app underline the need for thorough pretesting of the app and also conducting pilot trials, such as this study, before the app is used in a large-scale clinical trial or implemented within health care. Nevertheless, whenever an app is used, continuous efforts should be made to detect and handle any issues because technical problems could have negative consequences on treatment credibility, adherence, and effectiveness.

Furthermore, difficulties in fulfilling assignments and adhering to treatment [57] and feeling worse on occasions [38] have been stated as possible negative effects of online treatment before. Such distress can be challenging for individuals with NSSI, as difficulties in regulating emotions are especially prominent within this group [23]. Detecting participants experiencing distress because of the self-responsibility inherent in the online format and offer support is therefore important. The findings could be interpreted as a need for more individual tailoring in online treatments for NSSI to facilitate successful treatment outcomes and identify patients in need of adaptions or potentially more extensive treatment first hand. This means of treatment may not be suitable for the most severely affected; although online treatment offers frequent contact that could facilitate detection of deterioration in some cases, there may be dimensions of a face-to-face contact (eg, nonverbal communication) that are lost in online treatment, which could possibly make it more difficult to detect deterioration in other cases. Nevertheless, online treatment could serve important functions; the results from this study indicate that some prefer online contact, and a third of the adolescents in this study (and a fifth in the overall pilot trial [36]) identified as nonbinary, which could indicate that the online format may facilitate treatment-seeking among gender minority groups.

Although our study provides important information on how skills training, caregiver involvement, and the online format can be experienced, it is not without limitations. Irrespective of sample size, sampling from a single pilot study limits the transferability of the results [58]. Nevertheless, by including adolescents of different ages, genders, and NSSI frequencies, we could collect rich data, hopefully making the findings transferable in the sense that they speak for more than the individuals interviewed. However, although we strived for variability in the sample, we have no adolescents who identified themselves as male, and none of them were aged 13 years. Moreover, some families were interviewed months after ending the treatment, possibly introducing recall bias but also introducing the possibility of capturing late effects.

Our results indicate the potential development of online ERITA. Primarily, it is important to systematically detect those at risk of distress because of the self-responsibility inherent in the online format. As online treatment generally offers structured treatment content, questions regarding treatment engagement and feelings of increased distress can easily be incorporated into the treatment modules. Accordingly, modifying and evaluating the efficacy of online ERITA is warranted, and a randomized controlled trial is currently being conducted and is registered at ClinicalTrials.gov (trial number: NCT03353961).

If online ERITA is proven to be efficacious, steps have already been taken to prepare for a large-scale implementation trial within Swedish child and adolescent mental health services. In this phase, we will not only monitor treatment effectiveness and potential adverse events but also study patient as well as caregiver experiences and evaluate the intervention from an organizational perspective.

The findings from this qualitative study exploring the experiences of online ERITA for those with NSSI and their caregivers provide some important insights. First, decreased emotion regulation difficulties were an important outcome for adolescents, implying the importance of targeting emotion regulation and measuring several outcomes when evaluating treatment for NSSI. In addition, several different skills seem to facilitate emotion regulation ability and having access to such skills could hinder NSSI. Second, involving caregivers through a parallel online course seems to be an accepted and beneficial format to deliver family interventions for this patient group. Third, it is possible to learn and practice skills training in online format, with several means of support, from both the mobile app and the therapist. Finally, although online treatment could be empowering, there is a risk of distress because of the self-responsibility inherent in the online format that needs to be addressed.