Evaluation of a Therapeutic Education Application in the Treatment of Young People with Moderate or Problematic Screen Use

Overview

This single-center, controlled, and randomized study evaluates the effectiveness of the Phonix Care app in regulating screen use among young people aged 11 to 25. Faced with high and often concerning levels of screen consumption among youth, this research aims to provide an innovative intervention method beyond current psychotherapeutic and pharmacological approaches, which are often limited by the risk of relapse and the difficulty in delaying the short-term rewards offered by screen activities \[1, 2, 3\]. Phonix Care is designed to encourage awareness and self-regulation of screen use, thus promoting more responsible and autonomous behavior. The primary outcome measure is based on a problematic screen use score derived from the Digital Addiction Scale. Secondary objectives include examining the effects of the app on screen consumption, physical health, mental health, and motivation towards studies, measured through a series of questionnaires and objective evaluations. The study is conducted on 138 subjects, divided into two groups: an experimental group and a control group, over a participation period of six months. Statistical analyses will include descriptive analyses, multiple linear regression, and mediation models to assess the impact of Phonix Care. The expected outcomes of this research include significant contributions to the scientific literature regarding screen use among youth, as well as advances in adolescent and young adult health and psychology. In practice, the evaluation of Phonix Care could lead to the development of an effective medical device to quantify and treat problematic screen use, offering a complementary therapy to existing methods to prevent or remedy this issue. 1. Winkler A, Dörsing B, Rief W, Shen Y, Glombiewski JA. Treatment of Internet addiction: A meta-analysis. Clinical Psychology Review. 2013;33(2):317-29. https://doi.org/10.1016/j.cpr.2012.12.005 2. Xu LX, Wu LL, Geng XM, Wang ZL, Guo XY, Song KR, Liu GQ, Deng LY, Zhang JT, Potenza MN. A review of psychological interventions for Internet addiction. Psychiatry Research. 2021;302: 114016. https://doi.org/10.1016/j.psychres.2021.114016 3. Zajac K, Ginley MK, Chang R, Petry NM. Treatments for Internet gaming disorder and Internet addiction: A systematic review. Psychology of Addictive Behaviors. 2017;31(8):979-94. https://doi.org/10.1037/adb0000315

Quality assurance: A risk analysis of our application was conducted by an external organization, Surgiqual Institute. Their audit validated that our cybersecurity systems and risk management procedures were state-of-the-art in compliance with medical legislation applicable to our application. They produced a document to state that, based on their audit, they affirm the responsibility for ensuring the technical and legislative compliance of our application. \------------------------------------------------------------------------------------------------------------------------------------ Data checks: each data type had to match with a user profile template (JSON FORMAT) : * unique\_encrypted\_code\_name (characters AND number) * usage\_data (characters) * date (characters : DD-MM-YY) * screen\_type (characters) * app\_name : (characters) * usage\_data\_per\_app : usage\_value (number) * questionnaire\_data (characters) * date (characters : DD-MM-YY) * questionnaire\_name (characters) * questionnaire\_item : questionnaire\_response (number) * additional\_personal\_data (characters) Source data verification: a preliminary technical study (with 15 participants) was conducted to: * Validate that the self-reported hourly screen usage in the technical study corresponded to the data passively collected by our telemetric measurement applications over one month. * Confirm that the self-reported responses to online questionnaires matched the actual answers provided by the participants. * Ensure that each participant's key could be used only once. * Verify that participants assigned to the experimental group couldn't bypass the application's restrictions. Data dictionary: * Daily application usage data (Source : application Phonix Care) : * First opening schedule * Last opening schedule * Opening Frequency * Usage duration * Questionnaires responses (Source : the participant through the application Phonix Care) o Digital Addiction Scale * Regulation of Screen Time Consumption * International Physical Activity Questionnaire * Sleep Schedules diary * Revised Screen for Child Anxiety Related Emotional Disorders (SCARED-R) * University of Laval Loneliness Scale (ULS) * Rosenberg Self-Esteem Scale * Education Motivation Scale (EMS) * Experimental arm only : specific screen rules during the 5-months intervention period and the number of challenges that were completed Standard Operating Procedures (SOPs) were split into 10 steps : 1. Patient Recruitment: * Patient recruitment procedures will be conducted in accordance with the study protocol. * Recruitment efforts will be documented and tracked using electronic records maintained within the clinical trial management system (CTMS). 2. Data Collection: • Data collection will be performed using the Phonix Care application for daily application usage data. * Participants will input responses to questionnaires directly into the Phonix Care application. * For participants in the experimental arm, specific screen rules adherence and challenge completion will be recorded within the Phonix Care application. 3. Data Management: * Data collected from the Phonix Care application will be securely transmitted and stored on a dedicated health server hosted by a certified data management provider (AZNETWORK). * Access to the data will be restricted to authorized personnel only, with appropriate user permissions assigned based on roles and responsibilities. * Pseudo-anonymization procedures will be implemented to protect participant confidentiality. * Regular data backups will be performed to ensure data integrity and availability. 4. Data Analysis: • Data analysis will be conducted using statistical software approved by the study investigators (notably R, SPSS and Python). • Analysis will include aggregating daily application usage data, questionnaire responses, and experimental arm-specific data to assess intervention efficacy and participant outcomes. 5. Reporting for Adverse Events: * Any adverse events reported by participants will be documented in a dedicated electronic data capture system. * Adverse events will be promptly reviewed by the study investigators and reported to the appropriate regulatory authorities as per regulatory requirements. 6. Change Management: * Any modifications to the study protocol or data management procedures will be documented and approved by the study sponsor and ethics committee. * Changes will be communicated to relevant study personnel, and updated procedures will be implemented accordingly. 7. Quality Assurance: * Regular quality checks will be conducted to ensure data accuracy and consistency. * Data validation checks will be performed to identify any discrepancies or anomalies in the collected data. * Internal audits will be conducted periodically to review data management procedures and compliance with SOPs. 8. Training and Compliance: * Study personnel involved in data collection and management will receive training on SOPs and data handling procedures. * Compliance with SOPs will be monitored and enforced throughout the duration of the study. 9. Record Keeping: • All study-related documentation, including SOPs, data management logs, and training records, will be maintained in a secure electronic repository and duplicated to a secured space into a specific room of the AGEIS laboratory. • Records will be retained in accordance with regulatory requirements and study protocol specifications. 10. Documentation and Archiving: * Upon study completion, all study documentation will be archived for future reference and audit purposes. * Archiving procedures will adhere to regulatory guidelines and institutional policies. Sample size assessment: To evaluate the effectiveness of Phonix Care using the overall score from the Digital Addiction Scale by Hawi et al. (2019), with an average Cohen's effect size d= 0.30 to 0.40 and a standard deviation of 19.25 (mean= 56.3), here are the necessary sample sizes for different statistical powers (1-β), with a significance level of α= 0.05: 80% power: from 96 to 174 participants required. 85% power: from 110 to 200 participants required. 90% power: from 129 to 233 participants required. \------------------------------------------------------------------------------------------------------------------------------------ Plan for missing data: We conduct an analysis of the missing data mechanism according to the rules set by Little and Rubin. Although it is very rare, if we validate the hypothesis that the missing data are completely random (Missing Completely At Random), we conduct the analyses using the incomplete data set. This data set will not bias the estimates. The most likely case is the validation of the Missing At Random hypothesis, which suggests that the missing data are due to one or more factors in our possession (e.g., experimental condition, threshold of problematic use), we proceed with multiple imputations before conducting our analyses. To determine if certain factors can explain whether the data are missing or not, we use logistic regression analyses via the GLM package on R Studio version 4.0.2. In the case of multiple imputations, we use the MICE package on R Studio version 4.0.2. \------------------------------------------------------------------------------------------------------------------------------------ Statistical analysis plan: We first proceed with the descriptive analysis of screen usage profiles and the number of profiles observed in our sample. We expect to observe at least three usage profiles: moderate, intensive, and problematic. To do this, we use the K-means clustering method. Next, the variables measured by questionnaire undergo longitudinal confirmatory factor analyses to ensure that, despite experimentation, we observe some longitudinal invariance of the measurement constructs (i.e., weak invariance). For our primary research objective, we conduct analyses using multiple linear regression. By controlling for certain factors that may have an effect on problematic screen usage (e.g., gender, age), we evaluate the simple effects of digital addiction scores before the study and the assignment group, and then the interaction effect between this addiction score and the assignment group on digital addiction scores at the end of the study. To address our secondary objectives, we conduct multiple linear regression and mediation analyses for each of the secondary objective variables as dependent variables in linear regressions and as mediation variables.