An open feasibility trial of a digital treatment for adolescents with eating disorders was conducted in routine clinical care. The trial was guided by the UK Medical Research Council’s framework for developing and evaluating complex interventions and followed CONSORT (Consolidated Standards of Reporting Trials) standards for feasibility trials (Checklist 1) [47,48]. In line with these recommendations, 4 key progression criteria were assessed to evaluate feasibility across multiple domains: sample representativeness, adherence, acceptability, and preliminary efficacy. In this study, we specifically report findings related to sample representativeness as an indicator of feasibility. Sample representativeness was examined in relation to overall eating disorder psychopathology to better understand which adolescents sought digital treatment and the extent to which the participants reflected the intended target population. This was assessed by analyzing baseline characteristics from the feasibility trial and evaluating how well the key components of the intervention corresponded to the clinical profile of the sample. The study used baseline data from participants included in the feasibility trial between March 2024 and May 2025. In accordance with feasibility trial design, no formal sample size calculation was conducted.

Adolescents receiving the digital treatment were patients at child and adolescent psychiatric outpatient clinics in Norway, specifically within the catchment area of Haukeland University Hospital and Helse Fonna. These clinics are part of the state-funded Norwegian specialist health care system and are located within the Western Norway Regional Health Authority, serving a population of approximately 122,000 youth under 18 years [49].

Participants were recruited from 8 child and adolescent psychiatric outpatient clinics. Information about the study was disseminated through meetings with clinic managers, emails to staff, and flyers placed in waiting rooms. Interested participants were provided with a link to the study website for information about the study and a brief preliminary online screening. The online screening captured information on disordered eating, eating-related behaviors, functional impairment in daily life, as well as key inclusion criteria such as age, internet access, and language. Eligible participants were given the opportunity to leave their contact information and were contacted by telephone to confirm inclusion and exclusion criteria. A diagnostic assessment was conducted by trained clinicians within the research group using the Mini-International Neuropsychiatric Interview (MINI) [52] and Eating Disorder Assessment for DSM-5 (EDA-5) [53]. As the International Statistical Classification of Diseases, Tenth Revision (ICD-10) is the standard diagnostic framework in Norway, final diagnoses and all inclusion and exclusion criteria were determined according to ICD-10 [51]. While MINI was used as a diagnostic instrument, the EDA-5 served to systematically assess eating disorder symptoms and support clinical judgment, rather than as an independent diagnostic tool. In response to slow recruitment, additional promotion of the study was carried out via upper secondary schools and advertisements on social media. As the study was conducted in routine clinical care, all participants had to be eligible for specialized care in accordance with national priority guidelines [54]. Participants who were not already patients at an outpatient clinic were required to visit their general practitioner to confirm medical stability and obtain a referral. Eligible participants were given access to the native treatment application, in which they signed consent forms, with those aged 15 years required to provide additional parental consent.

The intervention was developed in line with the person-based approach (PBA) [55]. PBA is a methodological framework for understanding the most effective way to apply appropriate behavior change techniques in the specific context of the intervention and its intended users [55]. In line with PBA, adolescents with lived experience of eating disorders [27], mental health professionals, and the existing evidence base helped us identify key issues, needs, and challenges the intervention should address. Drawing on this in-depth understanding, key components were identified and systematically embedded into the design of the intervention. A logic model was created to describe how the key components of the intervention should lead to positive behavior change and better health (Figure 1). During development, the key components were incorporated and user-tested to refine the digital treatment to meet user requirements. This model serves as background for the second aim of this study: exploring whether the following key components of the novel digital treatment align with the adolescents it is designed to support: eating disorder symptomatology, psychosocial impairment, emotional difficulties (emotion dysregulation, anxiety, and depression), negative self-evaluation, and motivational factors.

The development process resulted in a therapist-guided digital treatment delivered through a smartphone app. The intervention is based on a transdiagnostic cognitive behavioral framework targeting mechanisms that maintain eating disorders, alongside elements from emotion regulation approaches to enhance distress tolerance [56,57]. The treatment consists of eight modules delivered in a fixed sequence: (1) What is an eating disorder? (2) Motivation and change, (3) Balancing food and activity, (4) Involving others, (5) Management of emotions, (6) Managing social situations, (7) Improving self-esteem, and (8) Preventing relapse. Modules were released weekly, although 2 were designed to be completed over a 2-week period (modules 3 and 5). Adolescents received weekly clinician support through scheduled telephone consultations lasting approximately 15 to 30 minutes. In addition, they could send asynchronous messages to clinicians to address technical or practical issues. During the feasibility trial, the intervention was incorporated into the existing workflows of participating Child and Adolescent Mental Health Services (CAMHS) outpatient clinics. The clinicians delivering the intervention were employed within CAMHS and were assigned approximately 10% of a full-time position, corresponding to a caseload of around 3 patients each. Standardized guidelines were developed to support clinicians in delivering the intervention. Clinicians also received brief introductory training on the digital platform and treatment procedures. In addition, they participated in fortnightly 1-hour group supervision sessions led by an experienced clinical psychologist.

Data were collected using self-report questionnaires administered online. Sociodemographic data (age, gender, living situation, engagement at school and with friends) were gathered using a questionnaire developed specifically for this study. The selected measures have been applied in adolescent and eating disorder populations, used in evaluations of digital interventions for this group, and are consistent with international consensus recommendations on patient-centered outcome measures in eating disorders [58,59].

Data were analyzed using SPSS (version 29). Given the descriptive nature of this study, data are presented in terms of frequencies, percentages, means, and standard deviations.

This study was approved by the Regional Committees for Medical and Health Research Ethics in Norway (REC-639031). In addition, the study was conducted in line with the Haukeland University Hospital’s guidelines for ethics and privacy considerations. Informed consent was obtained via the smartphone app.

Following 14 months of recruitment, baseline data were used for this study as the project neared completion. A total of 641 individuals accessed the online screening portal (Figure 2) between March 2024 and May 2025. Of these, 577 individuals were ineligible due to incomplete screening, being outside the age range, not meeting diagnostic criteria for an eating disorder (low symptom severity), or living outside the catchment area. In addition, 129 individuals met the criteria for the online screening but could not be contacted as they did not leave their contact information. Of the 64 adolescents eligible for approach, 11 were excluded because they were unable to be reached or declined participation. Of the 53 adolescents contacted by telephone, 20 were excluded from the study due to age (n=10), a primary diagnosis of anorexia nervosa (n=4), receiving face-to-face psychological treatment (n=3), low symptom severity (n=2), or ARFID (n=1). After inclusion, 5 adolescents changed their minds regarding participation, 2 were unable to be reached, and 1 was excluded due to a medical condition. Twenty-five adolescents completed the baseline screening and enrolled in the study. These were recruited through CAMHS (n=15), upper secondary schools (n=5), and social media (n=5).

All participants were women (Table 1). The mean age of the study sample was 16 (SD 0.87, range 15-18) years. Over half of the sample were living with both parents (60%). All participants were engaged in school (100%), and almost all (96%) were engaged with friends on a regular basis.

Most of the adolescents were diagnosed at inclusion with atypical anorexia nervosa (40%), with the second largest group being eating disorder, unspecified (28%) (Table 2). The rest of the sample was equally distributed between atypical bulimia nervosa (16%) and binge-eating disorder (16%). The mean age at eating disorder onset was 13 (SD 1.68, range 9‐16) years. Half of the adolescents had previously received face-to-face treatment for an eating disorder (52%, 13/25).

The adolescents’ motivational factors for participating in a novel digital treatment are summarized in Table 3. Most of the adolescents’ responses to items 1 and 3 were in the higher categories, 4 (“agree”) and 5 (“strongly agree”), indicating high internal treatment motivation. When asked to rank (0%‐100%) how much effort they would put into carrying out the treatment, the mean score was 83% (SD 14.34%), ranging from 50% to 100%. Regarding the involvement of others, the mean score was somewhat lower (52%, SD 33.18%), with a range from 0% to 100%.

The distribution of eating disorder symptoms was in the severe range, with a mean EDE-QS score of 21.56 (SD 6.80) [61]. Table 4 provides an overview of the distribution of symptoms, range, cutoff, and mean.

The global impairment score was high at 28.08 (CIA) [62] (Table 4). The impairment subscales show difficulties across the personal, social, and cognitive domains [63].

To explore the adolescents’ emotional difficulties, emotion dysregulation and symptoms of anxiety and depression were assessed (Table 4). The sample’s abilities to regulate emotions indicate moderate to severe difficulties within this domain, with a total mean DERS-18 score of 53.72 (SD 13.92) [66]. The adolescents had difficulties in all the 6 DERS-18 subscales. In addition, they reported anxiety symptoms above the cutoff (GAD-2 score 3.44, SD 1.69) [68]. The mean score for depression was below the cutoff (PHQ-2 score 2.60, SD 1.56) [70]; however, using an optimal cutoff score for adolescents of 2 or above, the score exceeded the threshold [71].

To evaluate the sample’s level of negative self-evaluation, we measured the adolescents’ self-esteem (RSES mean score 11.76, SD 6.33), which was in the range indicating severely low self-esteem [72].

The total burden of symptoms was calculated (Figure 3). Median response scores ranged from 0 to 6, with the median score on 5 measures above the cutoff.

The second aim of this study was to explore whether the key components of the novel intervention (for more details, see Figure 1) align with the characteristics of the adolescents recruited for this study. Table 5 shows an overview of the key components, the results from measuring these characteristics, and an evaluation of the alignment.

As part of examining the feasibility of a novel digital treatment, this study assessed sample representativeness. Specifically, the study examined the characteristics of adolescents seeking digital treatment for eating disorders within routine clinical care and explored whether the key components of the novel digital treatment aligned with the characteristics of the adolescents it is designed to support. Although recruitment was not specified as a predefined feasibility criterion, the findings revealed substantial challenges in this area. Given the central role of recruitment in feasibility research, particularly for understanding whether the intended population can be reached, these challenges are addressed first in the discussion to provide important context for interpreting the remaining findings.

The results revealed considerable challenges in enrolling participants. Despite substantial traffic to the online screening portal, only a small proportion of adolescents progressed to study participation, suggesting a gap between initial interest and actual engagement. This pattern may reflect well-established features of eating disorders, such as ambivalence toward treatment and fluctuating motivation, which can reduce readiness to seek or accept help [3,9]. Additional challenges may be related to adolescence as a developmental stage, where factors such as stigma, reliance on caregivers, and competing demands can influence help-seeking behavior [75]. Intervention-specific and contextual factors may also have contributed. Some adolescents may have been hesitant to engage in a digital treatment format or may have preferred traditional face-to-face care [34,37]. At the same time, recruitment via social media appeared to reach individuals who had not previously accessed clinical services, indicating an unmet need for more accessible entry points to care. It should also be noted that recruitment challenges are common in clinical research [76]. In this study, such challenges may have been further exacerbated by practical constraints within clinical settings, including limited time and resources among clinicians. Together, these findings suggest that while digital interventions may lower initial barriers to help-seeking, they do not fully overcome challenges related to participation in studies and treatment.

Atypical anorexia nervosa was the most frequently assigned diagnosis, which aligns with data from a Norwegian community-based prevalence study [10]. Furthermore, the adolescents’ distributions of eating disorder symptoms and psychosocial impairment are consistent with the diagnostic thresholds for inclusion in this study. Moreover, their elevated levels of emotional difficulties and negative self-evaluation are in accordance with other studies examining these mechanisms among adolescents with eating disorders [77-79]. Low self-esteem and mood intolerance are particularly pertinent in the adolescent age range [78], and anxiety and depression are the most prevalent comorbid disorders [79]. The adolescents’ total burden of symptoms across domains emphasizes the sample’s severity. The adolescents reported a high level of internal pretreatment motivation. The strong internal motivation may be linked to the adolescents’ severe eating disorder symptoms [9] or to their being more likely to engage in the intervention due to recruitment via a self-selection process [80]. Providing adolescents with a sense of control and agency in treatment decision-making may contribute to improved adherence to therapeutic interventions [37].

The fact that none of the adolescents included in this study were men reflects the well-established gender discrepancies in the field of eating disorders [5]. Stigma associated with the disclosure of mental health issues and cultural perceptions that eating disorders are typically female disorders might influence help-seeking behaviors among adolescent boys [81]. Moreover, although individuals that identify as transgender or nonbinary are at particularly high risk for developing eating disorders [82,83], no participants in this study identified with a gender other than women. The results of this study highlight the presence of underrepresented groups within the adolescent population seeking treatment for eating disorders [17,81,83].

The distribution of eating disorder symptoms within the sample indicates that the treatment targets core symptoms relevant to the population. In addition, the intervention’s focus on functioning is well-matched to adolescents’ needs. Moreover, the inclusion of a transdiagnostic sample aligns with the goal of the novel intervention to address the multifaceted and overlapping nature of these conditions. Drawing on perspectives from adolescents with lived experience of an eating disorder, the design of the digital treatment embedded emotional difficulties and negative self-evaluation as key components [27]. The results of this study show that adolescents’ high levels of difficulties within these domains support the relevance of the treatment in addressing emotional dysregulation, anxiety, depression, and self-worth. This alignment may indicate that the intervention is designed for the users and contexts in which it will be implemented. Furthermore, adolescents’ levels of pretreatment motivation were higher than anticipated when designing the novel intervention. Variability across adolescents highlights that tailored engagement strategies may be an important component for optimizing uptake and adherence to the digital treatment. Although the intervention was developed through a co-design process to reflect the needs of the intended target population, a one-size-fits-all approach may not fully address the heterogeneity among adolescents with eating disorders. Additional flexibility or personalization within the intervention may be needed to better accommodate individual differences and enhance engagement.

Half of the adolescents seeking digital treatment had previously received face-to-face treatment for an eating disorder. This emphasizes the relevance of digital interventions to complement or extend traditional care, particularly for individuals who may not have experienced sufficient benefit from prior treatment. Moreover, the adolescents’ high level of internal pretreatment motivation may suggest that digital treatment can also serve as a viable alternative to face-to-face approaches for some individuals. Suitable options to traditional eating disorder treatment are urgently needed, not only because of existing treatment challenges but also to address future public health care concerns. A remarkable increase in eating disorders among adolescents since the beginning of the COVID-19 pandemic emphasizes a real challenge for health care providers [4,84,85].

Adolescents with anorexia nervosa and bulimia nervosa were excluded from participating. Given the high prevalence of adolescents who go untreated [11], future steps should be taken to offer the intervention to adolescents with these disorders within routine clinical care. A potential solution involves combining digital treatment with periodic in-person medical evaluations. This would also address the needs of those who prefer face-to-face interventions but who also endorse the advantages of digital interventions (eg, availability in times of need and the ability to address stigma associated with help-seeking) [34]. Additionally, digital platforms offer opportunities to integrate tools for monitoring symptoms that may signal medical risk, thereby enhancing patient safety and clinical responsiveness.

Although emotional difficulties are common in adolescents with eating disorders [77-79], these domains are not a primary focus of first-line eating disorder treatments [17]. In the design of mental health technologies, a shift from a top-down approach to a user-driven process necessitates new models of behavior change that move beyond traditional psychotherapy frameworks [76]. The results from this study are in concordance with other studies that address the importance of targeting emotional difficulties in the treatment of eating disorders [29].

The current findings highlight the need to increase diversity among adolescents participating in eating disorder treatment [17,82,83]. In particular, the absence of male participants and the lack of representation of gender-diverse individuals underscore well-documented disparities in help-seeking and access to care. Future research should prioritize recruitment strategies that actively engage these populations, including the use of culturally sensitive language and appropriate measurement tools [17,82,83].

This study has important limitations that need to be addressed. Several aspects of the study design may introduce sampling bias. Recruitment was limited to adolescents eligible for specialist outpatient care within a specific catchment area, which may not reflect the broader geographical population. In addition, this recruitment approach may have constrained the extent to which the sample reflects the broader sociodemographic characteristics and contextual factors considered during the co-design process. The exclusion of those with anorexia nervosa and bulimia nervosa further narrows the sample. In addition, recruitment through self-selection options may favor adolescents who are more motivated, less burdened by barriers to help-seeking, or who already knew that the content would align with their preferences. Moreover, the lack of a control or comparison group may introduce sampling bias, as it is unclear whether participants differ systematically from those who did not participate or from those who would have received other treatments. These factors may reduce the representativeness of the sample and limit generalizability. Another limitation is the lack of systematic reporting and evaluation of the co-design process. Even though it is reasonable to consider that user involvement in the design and development process from the start influences the extent to which adolescents’ characteristics are embedded in the design, we do not know whether it was the involvement of users at a particular phase or the continuous involvement across all phases of design that offers benefits to design research.

The findings from this study underscore the relevance of digital interventions to complement or extend traditional eating disorder care for adolescents. Furthermore, the findings highlight the importance of enhancing diversity among adolescents participating in digital eating disorder treatments to ensure equitable access and representation. The findings emphasize the importance of designing digital treatments that are sensitive to normative biases and highlight the value of involving a diverse group of adolescents with lived experience of eating disorders in the design and development process from the start. In addition, this study highlights the importance of designing digital treatments that address the multifaceted nature of eating disorders and are tailored to the needs and preferences of a transdiagnostic population. The present findings are important insofar as they may indicate that the key treatment components align with the characteristics of adolescents seeking digital treatment for eating disorders within routine clinical care. Results from the open feasibility trial will show whether adolescents experience the novel digital treatment as acceptable, and credible, and whether the treatment reduces symptoms and increases coping skills.