A smartphone application to investigate the relationship between digital media use and mental health using ecological momentary assessment in a clinical sample of youth: a feasibility study
MetadataShow full item record
BACKGROUND: There is significant evidence supporting a link between excessive digital media use and adverse mental health outcomes. Since the average American teenager spends approximately 7 hours a day using digital media, this relationship has become of considerable interest in the field of public health. Studies have shown that anxiety and depression may both be associated with increased screen time as well as with Problematic Internet Use (PIU). PIU refers to a set of symptoms related to an individual’s inability to control their use of the Internet and appears to have characteristics of both substance use disorders and impulse control disorders. However, research on PIU has many limitations, including the lack of formal diagnostic criteria, unequal gender representation in study samples, and the relative dearth of studies conducted in adolescents, especially those with pre-existing mental health issues due to their status as a protected population. This study aims to address these limitations in existing literature through use of ecological momentary assessment (EMA), a research method that samples participants in their own environment repeatedly using one of various data collection methods (e.g. paper and pencil diary, text message surveys, app-based surveys). The specific aims of this study were: 1) to explore the feasibility and acceptability of a six-week daily EMA protocol in a clinical population of adolescents and young adults using a smartphone application, 2) to describe the relationship between PIU and depression/anxiety in a clinical sample of adolescents and young adults ages 12-23, and 3) to identify possible relationships on which to focus future studies of problematic digital media use and psychiatric symptomatology in this vulnerable population. METHODS: The study enrolled 25 adolescents and young adults ages 12-22 years who owned a smartphone and received mental health services at a community hospital in the greater Boston area. Participants were surveyed once a day for six weeks using a smartphone application, mindLAMP, to record self-report data. The daily survey included validated depression, anxiety, and PIU scales (the PHQ-8, GAD-7, and PIU-SF-6, respectively) as well as two sub-surveys inquiring about risky online behaviors and screen time use, and were collected via mindLAMP surveys. Participants also completed a short exit survey once their study period concluded. Feasibility was assessed by evaluating the sample-wide mean rate of response on daily surveys for the six-week protocol, rates of daily survey initiation and completion, and study completion rate. Acceptability was determined based on whether this data collection method provided more data than the current standard of care, which typically involves one meeting per week for psychotherapy. RESULTS: 96% of participants completed the study, and results demonstrated a 39.8% mean response rate, 43.2% mean survey initiation rate, and 38.5% mean survey completion rate. Type of phone use (overall rate p=0.029, initiation rate p=0.023, completion rate p=0.037), presence of an anxiety disorder (overall rate p=0.006, initiation rate p=0.038, completion rate p=0.004), and presence of co-morbid diagnoses (overall rate p=0.042, completion rate p=0.047) were significantly related to response rate, whereas age, gender, symptom severity, presence of an affective disorder, gender dysphoria, or ADHD were not. Regarding specific aim two, significant associations were detected between symptom scale scores and PIU and risky online behavior assessment (QUAL) scores such that: PIU-SF-6 and GAD-7 scores were positively correlated (p=0.032), PIU-SF-6 and PHQ-8 scores were positively correlated (p=0.050), GAD-7 and QUAL scores were positively correlated (p=0.004), and PHQ-8 and QUAL scores were positively correlated (p=0.0002). These results replicate findings in the literature of a significant relationship between anxiety and depression and both PIU and risky online behaviors. The lack of observed significant relationship between PIU and QUAL contradicts findings within the literature, and is possibly due to our small sample size. Significant associations were not detected between between categorical diagnoses of ADHD, Anxiety Disorders, Affective Disorders, or Gender Dysphoria and PIU-SF-6 scores. CONCLUSIONS: Results indicate that this EMA protocol is feasible and acceptable in this population, suggesting that this methodology may represent a new avenue to conduct research in adolescents and young adults with mental health concerns and, in the future, deliver treatment to these individuals. Limitations of this study include small sample size and participant bias in self-report data. Recommendations for future research include replication of this EMA protocol with a larger number of participants, incorporating methodologies that address the self-report bias.