Effectiveness of a transdiagnostic individually tailored Internet-based and mobile-supported intervention for the indicated prevention of depression and anxiety (ICare Prevent) in Dutch college students: study protocol for a randomised controlled trial

We designed a randomized controlled trial to evaluate the (cost-)effectiveness of a transdiagnostic individually tailored online intervention compared with care as usual (CAU) in reducing symptoms and preventing the onset of a full episode of depression and anxiety in college students. We hypothesize that participants in the two active intervention groups will experience larger symptom reduction of depression and anxiety than those in the control group. Moreover, we hypothesize that the lesser use of personnel (i.e. eCoaches) in the unguided condition will be reflected in a more favorable cost-effectiveness ratio for this group compared to the guided condition.

This study is part of a large European Horizon 2020 research project (ICare) and will be conducted with similar study designs, procedures, inclusion and exclusion criteria, and primary outcome measures among the general population in Germany, Switzerland, and Spain. The study protocol of the other ICare Prevent trials will be published elsewhere (Weisel, Zarski, Berger, Krieger, Schaub, Moser, et al.: Efficacy and cost-effectiveness of guided and unguided internet-based mobile-supported indicated prevention of depression and anxiety (ICare Prevent): A multi-country three-armed randomized controlled trial., unpublished). Within the large ICare project, several Internet-based interventions will be used for different mental health conditions. The overall aim is to increase access and uptake of such interventions by overcoming current barriers to implementation and establishing a comprehensive model of mental health promotion in Europe. With the goal of improving quality of healthcare, effectiveness, and acceptance of this kind of intervention, the ICare project addresses core goals of EU health policy.

Students will be recruited and provided with contact details through e.g. a website (https://​icare-online.​eu/​nl/​), social media (e.g. Facebook), and information material distributed on campus at Dutch universities. They are eligible if they: (1) are aged ≥ 16 years; (2) experience at least mild self-reported symptoms of depression, defined as a score ≥ 16 on the Center for Epidemiological Studies Depression scale (CES-D; [41]), and/or anxiety, defined as a score ≥ 5 on the seven-item version of the Generalized Anxiety Disorder scale (GAD-7; [42]); (3) do not meet diagnostic criteria for a mood or anxiety disorder based on the Mini International Neuropsychiatric Interview (M.I.N.I.; [43]) at screening. Exclusion criteria are: (1) being on a waitlist for, currently receiving, or having received psychotherapy over the past six months for any mental health condition; (2) meeting diagnostic criteria for a lifetime bipolar disorder (M.I.N.I.) or having received a diagnosis of a psychiatric disorder; (3) being at moderate to severe risk for suicide (M.I.N.I.); (4) being in remission of a major depressive disorder (MDD) episode, defined as an MDD diagnosis in the previous six months and experiencing at least one cardinal symptom (e.g. persistent and pervasive low mood, loss of interest or pleasure in usual activities) during the previous three weeks (M.I.N.I.); (5) self-reported inability to read or write Dutch; (6) no informed consent; (7) no access to a computer or the Internet; or (8) participating in similar studies at time of inclusion.

Upon contact, potential participants will receive an information letter and informed consent form by email, the latter to be signed and returned. In addition, the email contains a link to the ICare Prevent platform on which participants have to register. After the informed consent form has been returned, the screening questionnaires for all self-reported inclusion and exclusion criteria will be made available on the platform. The screening consists of: demographic questions; the CES-D; and the GAD-7. Next, eligible participants will be asked for their telephone number in order to conduct the M.I.N.I. interview, the clinician rated version of the Quick Inventory for Depression Scale (QIDS-CR; [44]) and the Structured Interview Guide for the Hamilton Anxiety Rating Scale (SIGH-A; [45]). The interviews will be performed by trained clinical psychology master students under the supervision of a member of the research team; 10% will be recorded and rated by a second rater to assess inter-rater reliability. Consent for the recordings will be asked for at the beginning of the call. Disagreement between raters will be solved by discussion or by asking an experienced psychotherapist. Eligible participants will receive a confirmation by email and access to the baseline self-report questionnaires on the online platform. Once completed, they will be randomized to one of the two intervention conditions or the control group.

Non-eligible participants will be advised to consult their general practitioner or a (student) psychologist if they think they need help for their complaints. Those with moderate to severe suicide risk will also be advised to contact their general practitioner. In addition, they will be informed about the national suicide helpline (113online; [46]). The research team will contact those participants again one week later and ask whether they sought help.

The allocation scheme will be derived by computer using a random number generator at a 1:1:1 ratio. Participants will be randomized at an individual level by an independent institute (University of Münster, Germany) and randomization will be stratified by type of subclinical symptoms (depression or anxiety). Due to the nature of the study it is not possible to blind either participants or coaches, as they will be notified to which of the three conditions they have been assigned to. However, raters who perform the clinical interviews to assess outcomes will be blinded. In order to ensure that blinding was successful, participants will be informed about the importance not to indicate their allocation status at the beginning of the telephone interviews. Moreover, raters will be asked to guess the allocation status of each participant after the second interview at post intervention. The results will then be compared to what would have been expected by chance. The flowchart of the trial is shown in Fig. 1.

The technology platform used to deliver the ICare Prevent intervention is provided by Minddistrict. This company is full partner of the ICare project consortium and responsible for the provision and maintenance of the platform. Its content management system is used to upload the interventions, add new participants and eCoaches, as well as questionnaires. Access to the platform is provided by means of email and personalized password combination. It is currently used both in clinical research and in routine practice to provide guided and unguided self-help interventions for a variety of mental health conditions. The platform conforms to all required quality standards and operates according to the ISO 27000 and NEN 7510 norms. All data are securely stored on ISO 27000 certified servers and transmitted using HTTPS with SSL certificates (AES-256 and SHA-1, 2048-bit RSA). Unauthorized access to the platform is therefore not possible. A data management and safety plan has been developed as part of the larger ICare project.

ICare Prevent is an Internet-based intervention with mobile-support by means of an application (app). It uses both transdiagnostic and individually tailored components that are based on previously developed Internet-based modules, all using evidence-based CBT principles [47‐51]. ICare Prevent consists of seven main sessions and one booster session. Participants are advised to complete one main session per week. The intervention contains text, exercises, images, explanatory videos, as well as audio files and downloadable information. Testimonials are used to illustrate (homework) exercises. The main sessions follow the same basic structure. First, the content of the current session is explained, followed by a review of the homework exercises from the previous session. In accordance with the individually tailored approach of the intervention, participants can choose content depending on their predominant complaints, i.e. depression or anxiety. For example, in the fifth session, participants can either practice problem-solving or exposure techniques. The content per session is summarized in Table 1. Throughout sessions 2–7, participants can choose to complete one of eight choice modules following the main intervention content. In accordance with the transdiagnostic approach of the intervention, these modules aim at increasing resilience and tackle problems common to both depression and anxiety. In addition, five diaries with different content, such as positive activities and sleep, are accessible through the online platform and the app. An overview of all choice modules and diaries is given in Table 2.

The intervention can be used on (laptop) computers, mobile phones, and tablets. The mobile app can be used optionally. It provides access to the diaries and allows participants to activate push notifications. These short messages contain small exercises, such as short relaxation techniques, as well as motivational texts. The intervention uses a responsive web design in order to tailor content to participants’ needs, based on the choices they make. An optional read-aloud function is included. The expected time to complete each session is approximately 60 min. The two intervention conditions only differ in the support participants receive. In the guided condition, trained eCoaches (clinical psychology master students) spend approximately 20 min, but no more than 30 min per session on sending individual manualized feedback on the exercises through the intervention platform. In addition, they will send reminders for module completion in order to increase adherence. The eCoaches will be supervised by the members of the research team. Participants in the unguided condition receive only automatic motivational messages that aim to increase adherence by positively reinforcing participants for completing intervention modules and reminding them in case of a delay in completion.

Participants in all three conditions may use CAU services. We monitor healthcare services used, as well as other means to alleviate symptoms (e.g. as talking with relatives and friends), by means of a self-report questionnaire, the Client Service Receipt Inventory (CSRI; [52]). In the Netherlands, CAU for subclinical symptoms of depression and/or anxiety can include visits to the general practitioner, making use of support staff trained in the provision of help for mild psychological symptoms, or (student) psychologists according to Dutch guidelines [53, 54]. Alternatives include homeopaths, alternative medicine providers, or self-medication. Participants in the control group will receive access to ICare Prevent one year after trial inclusion.

Following screening and baseline (all groups), assessment points for the intervention groups include five weeks after randomization or directly after session 5 (whichever happens first) and eight weeks after randomization or post intervention (whichever happens first). Assessment points for the control group include five and eight weeks after randomization. Follow-up measures for all three groups consist of six and 12 months after randomization. Assessments include self-report questionnaires on the intervention platform and diagnostic interviews administered over the telephone. An overview of all measures taken by assessment point is given in Table 3.

The primary outcome is change in disorder-specific symptom severity from baseline to post intervention. For depression, this will be assessed using the QIDS-CR [44]. This interview has been used in different settings and has good psychometric properties [44, 61]. The SIGH-A [45] will be used for the assessment of anxiety. Its reliability and validity has been found to be good [45].

Secondary outcomes include self-reported reduction in depressive and anxiety symptoms from baseline to post intervention. In addition to screening, the GAD-7 will be used to measure anxiety symptoms throughout the trial. The nine-item Patient Health Questionnaire (PHQ-9; [62]) will be used to assess depressive symptoms. Research on the PHQ-9 has generally shown good psychometric properties among all age groups [62‐64]. Moreover, the time to onset of a mood or anxiety disorder within the 12-month follow-up period will be assessed by the M.I.N.I. Finally, academic performance will be measured by the Presenteeism Scale for Students (PSS; [65]). It is a valid and reliable measure that will be used to assess presenteeism in the student population [65]. In addition, students will be asked to provide the number of points they acquired on the European Credit Transfer System (ECTS) during the semester.

The CSRI [52] will be used to track societal costs. It has been adapted to the specifics of the Dutch college student context, measuring costs associated with healthcare use. In addition, it assesses absenteeism and presenteeism in college as well as in student jobs. For the latter, the Dutch minimum wage will be taken as an estimate of students’ income.

The primary endpoint is change in disorder specific symptom severity from baseline to post intervention. Based on the evidence from a meta-analysis by Cuijpers et al. [103] on the effectiveness of psychotherapies for subclinical symptoms of depression, as well as a randomized controlled trial on the effectiveness of an Internet-based intervention for subclinical depression ([24‐26], the expected effect size is d = 0.35. Due to a lack of information on effect sizes for subclinical symptoms of anxiety, d = 0.25 is used as a conservative estimation for both conditions. A study comparing the effectiveness of Internet-based psychotherapy with different levels of support found effect sizes in a similar range [39]. The sample size calculation is based on a repeated measure ANOVA and follows the recommendations given in Muller [104] and Muller et al. [105]. Thus, based on a global significance level α = 0.05 and power β = 0.95, 252 participants (84 per study arm) will be needed for an effect of 0.25 to be significant.

All analyses for the primary and secondary outcomes will be based on multilevel mixed model regression analyses for continuous outcomes in order to assess change over time. Time will be used as a predictor variable and baseline measures will be included as covariates. Linear models will be used for normally distributed data and negative binomial models for left-skewed data. All analyses will be performed on an intention-to-treat (ITT) sample. Per-protocol analyses including only participants who conformed to their group allocation will be performed in addition. Missing data will be handled using either multiple regression imputation techniques if appropriate or full information maximum likelihood estimation [106]. The secondary outcome, time to mental health disorder onset during 12 months follow-up in all conditions, measured by the M.I.N.I., will also be compared using survival curve analysis. The curves will be compared using cox proportional hazard regression analysis using baseline symptom severity as a covariate. Hazard ratios will be calculated as measure of the effect size of group differences. Within secondary analyses, interactions with predictors will be added to the model in order to identify risk and protective factors.

Cost-effectiveness will be assessed from a societal perspective for both disorder-free days and QALYs. The analysis will be performed according to the ITT principle. Multiple imputation according to the MICE algorithm developed by Van Buuren et al. [107] will be used to impute missing cost and effect data. Incremental cost-effectiveness ratios (ICERs) will be calculated by dividing the pairwise differences in the mean societal costs between the groups by the pairwise differences in mean effects between the groups. Bivariate regression models will be used to estimate cost and effect differences while adjusting for confounding if necessary. Statistical uncertainty will be estimated using bias-corrected accelerated bootstrapping with 5000 replications. Uncertainty surrounding the ICERs will be presented in cost-effectiveness planes and acceptability curves [108].

To our knowledge, this is the first trial that investigates the (cost-)effectiveness of a transdiagnostic individually tailored Internet-based intervention for the prevention of depression and anxiety in a college student population. Data on both symptom reduction as well as the time to onset of a mood or anxiety disorder will be measured. Although some research suggests that Internet-based interventions have the potential to be cost-effective compared to no treatment, no data are available for preventive transdiagnostic individually tailored Internet-based interventions for college students. Specifically, the added (economic) value of support by means of an eCoach needs to be closely monitored in the context of this low-intensity intervention.

Some possible limitations of the present trial should be noted. First, the sample size of this trial conducted among Dutch college students is insufficient for directly comparing the guided and unguided treatment arm as well as analyzing the time to a mental health disorder onset. Similarly, analyses of moderators and mediators will be underpowered and may therefore provide only limited insights. To counter this issue, data from all participating trials in the ICare project will be pooled in order to achieve enough statistical power. Second, an evaluation of societal costs might be difficult due to the fact that most students are seldom in full-time employment. The detrimental effect of depression and anxiety on absenteeism and presenteeism in the context of employment is therefore difficult to quantify. In order to increase the validity of the results, we have adapted the cost-effectiveness questionnaire (CSRI; [52]) to fit the specifics of a student population by focusing on absenteeism and presenteeism in student jobs, as well as in lectures. Finally, previous research has shown that adherence to Internet interventions is a point of concern, especially in unguided formats [111, 112]. A series of automatic motivational messages and reminders will be used in order to increase adherence in the unguided study arm and we will closely observe factors related to dropout. This will be beneficial in designing future interventions and improving uptake of intervention content.