Individualization of mHealth Interventions for children and adolescents

Abstract

Introduction: Insufficient physical activity has been established as a significant risk factor for non-communicable diseases, increasing the risk of conditions such as cardiovascular disease, hypertension, diabetes, dementia, obesity, and breast and colon cancer. Meanwhile regular physical activity is associated with positive effects on stress management and related health risks. The benefits of physical activ-ity are particularly impactful for children and adolescents, as behavioral changes during adolescence can extend into adulthood. However, the prevalence of insufficient physical activity and sedentary behavior among youth worldwide is steadily increasing, potentially due to the rise of digitalization and increased screen time. Surprisingly, longitudinal and representative data from the KIGGS study demonstrates no correlation between screen time and reduced physical activity in children. Addressing the digital realities of modern childhood and adolescents, digital health interventions (e.g., mHealth) may provide a life-relevant and motivating entry point for changing physical activity related behaviors. Numerous meta-analyses have demonstrated the effectiveness of smartphone interventions among youth, although effect sizes remain low. The lack of scientific foundations of content, non-specific approaches, inadequate age-appropriateness, low individualization, and poor usability are cited as possible reasons. Thus, innovative approaches are needed to increase the effectiveness and adherence of digital health interventions among children and adolescents, involving evidence-based techniques for behavior change (e.g., gamification, goal setting), age-appropriate developmental theories, motivational aspects, and multi-level individualization. As such, this dissertation focuses on the following research questions: (1) How does individualization and age as moderators impact the effectiveness of mHealth interventions for reducing sedentary behavior in children and adolescents? (2) What are the feasible mHealth-based physical activity and health objectives that can be achieved within the family context involving early adolescents? (3) How can digital health literacy be promoted in the school setting to encourage reflective and responsible use of mHealth applications among mid-adolescents? (4) How does individualization affect the effectiveness of physical activity-based mHealth interventions?

Methods: A cumulative dissertation consisting of seven pre-registered publications in national and international peer-reviewed journals was developed to address these research questions. To answer the first research question, a systematic review followed by a meta-analysis (1) was conducted to assess the effectiveness of digital health interventions for preventing insufficient physical activity and seden-tary behavior in children and adolescents across different developmental stages, as well as to compare individualized interventions with non-individualized interventions. To answer the next two research questions, which were based on the results of the aforementioned literature review, two mixed-methods cross-sectional studies were conducted to examine the prerequisites for digital health promotion for children and adolescents in (2) family and (3) school settings. Qualitative sub-studies were analyzed using qualitative content analysis with MAXQDA, while quantitative sub-studies were analyzed using SPSS, rStudio, and JASP. The family-focused study (2) was an explanatory sequential mixed-methods study that aimed to identify family health goals through interviews with N=60 parents and focus groups with N=120 adolescents. The subsequent quantitative sub-study surveyed N=1008 families nationwide on their interest in the identified family health goals and their health behavior. On the other hand, the school-focused study (3) was an exploratory sequential mixed-methods study that integrated an online survey of N=118 biology and physical education teachers and six focus group interviews with teachers and students (N=34). The surveys covered questions about the equipment and use of digital media, digital health literacy, and potential barriers of mHealth intervention in the context of physical education. The insights gained from the meta-analysis (1) and cross-sectional mixed-methods studies (2 & 3) were integrated into a multi-arm randomized controlled trial (4), including a study protocol, to answer the fourth research question and to evaluate individualized, sensory mHealth interventions. The experimental study includes N=995 participants, randomized to multiple study-arms with different levels of individualization including sensor- and app-based biofeedback, health needs of each individual, vital signs, and behavioral patterns. The study includes three measurement points at intervals of 8 weeks, with primary outcomes defined as heart rate variability, behavioral change (HAPA), and physical activity.

Results: The systematic literature search yielded 1101 studies, of which 12 were included in the qualitative synthesis and 10 in the meta-analysis (1). Findings indicated that digital health applications can effectively address insufficient physical activity, but their effectiveness in mitigating sedentary behavior remains uncertain. Additionally, our analysis suggested that highly individualized digital interventions may produce larger effect sizes in the context of insufficient physical activity, and that age-related differences may exist with respect to the degree of individualization required to achieve optimal outcomes. Addressing early adolescent target group, the subsequent family focused mixed methods study (2) found differences in health goals among families. Qualitatively identified mHealth related goals in the areas nutrition, mindfulness, abstinence, organized activities, resilience, nature as well as physical activity and combined the health behavior index of participants in a multiple regression model. The results revealed resilience, physical activity, and nature to be significant predictors of health behavior. Additional multiple logistic regression models identified healthier eating habits, communal cooking, outdoor activities, learning exercises for on-the-go, spending time in nature, stress management, and dietary changes as primary goals in the field of mHealth that children and adolescents would undertake with their parents. Addressing mid adolescents, both studies in the school focused project (3) identified a lack of knowledge and media infrastructure. The target groups showed a high interest in and need for the enhancement of digital health literacy. Compared to teachers of other subjects, physical education teachers showed lower digital health literacy and less interest. The results highlight the need for an improved infrastructure (e.g., access to Wi-Fi) and the exacerbated need for digital health literacy promotion in the school setting. In the randomized controlled trial focusing on late adolescents and adults (4), 170 of 995 eligible participants (26%) completed the post-measurement. MANOVA indicated small to moderate time*group interaction effects with physical activity-related outcomes of moderate to vigorous physical activity and inactivity-disruption counts in the app focused study-arms, but not for step counts and inactivity. Stress-related HRV parameters did not change over time. Despite high drop-out rates and a complex study design, individualized interventions revealed initial effects on physical activity but not the expected effects on stress-related outcomes.

Discussion: The aim of this dissertation was to investigate the impact of individualization on the effectiveness of mHealth interventions for children and adolescents at different developmental stages. The results revealed that each developmental stage of children has unique requirements. For instance, in early childhood and adolescence, the involvement of the social environment of the family was shown to be beneficial, whereas in middle adolescence, the development of health literacy for independent use of mHealth interventions obtains amplified relevance. In late adolescence, individualization of interventions through biofeedback or more complex methods such as machine learning becomes significant. Despite several limitations, the individualized mHealth interventions were found to affect the physical activity and health behaviors of children and adolescents more than non-individualized interventions, provided that they adequately address the digital health literacy according to the child's developmental stage, involve social systems, are based on central theories of health behavior change, and have an educational approach. Future approaches should focus on the appropriate use of health data to develop context-specific and relevant interventions that are adjusted according to gender, culture, and competence. Therefore, individualization alone appears to be a partial aspect of the effective application of mHealth interventions, but tackles many obstacles related to digital solutions for the reduction of insufficient physical activity and sedentary behavior as well as other health behaviors. These aspects are combined in the proposed Youth mHealth Behavior Change Model, which combines the HAPA model with the Self-Efficacy Model and the presented study findings of this dissertation, providing a framework for physical activity related health behavior change for children and adolescents via mHealth interventions.