Can Caregiver Smartphone use Predict Child and Caregiver Mental Health? An Application of Machine Learning

• In Event: Poster Session 09
  In Poster Session: PS 09 Topic - Parenting & Parent-Child Relationships

Abstract

Significant mental health problems can and commonly do occur throughout childhood. These challenges can lead to impairment across a variety of domains, including academic, social, emotional, and behavioral functioning (Sheehan, 2017). Given that 10-20% of children are affected by mental health symptoms throughout childhood (WHO, 2019), identifying key targets for intervention is crucial. Caregiver-child relationships have been heavily studied in the context of child mental health, as caregiving behaviors, and caregiver mental health symptoms have both been associated with psychological wellbeing among children (Morris et al., 2017). However, understanding which factors influence caregiver mental health is an understudied area, and can help to inform prevention programming. One factor that may influence caregiver mental health symptoms is smartphone use. The use of smartphones has been found to co-occur with depression, anxiety, stress (Elhai et al., 2017), and low self-esteem (Grant et al., 2019). However, much of this work has been conducted among adolescents and college students. Additionally, understanding how caregiver smartphone use impacts child mental health symptoms has not been examined. While data on smartphone use is often collected through subjective self-report measures (e.g., retrospective report of time spent using a smartphone), the current study will use a machine learning approach to assess associations between an objective measure of caregiver smartphone use and caregiver-child mental health symptoms. Participants (N = 50) will be caregivers of a child aged 6-9 enrolled into a larger NIH study designed to develop and test a just-in-time-adaptive intervention delivered via mobile technology to address caregiver-child conflict. Given the challenges in obtaining IRB approval of a remote protocol during COVID-19 pandemic, data collection has been delayed. However, we anticipate that the majority of the final sample will include two-parent households, and predominately White, and Latinx participants. All participants will download a Behavioral Tracking app onto their smartphone or a lender smartphone, which will track phone metrics and input features such as SMS and social media transcripts, app use metrics, screen time, browser history, call frequency and duration, phone pick-ups, and typed text. Self-report measures, to be completed remotely, include demographics, the Child Behavior Checklist (child mental health), the Symptom Checklist-27 plus and the Distress Tolerance Scale (caregiver mental health). An artificial neural network will use smartphone data along with the input features, in both a training and test set, to determine if the data can successfully predict our output, caregiver and child mental health symptoms. Results will inform our understanding of common phone use features associated with caregiver-child mental health symptoms. While prior work has found that smartphone use can have deleterious effects on youth, these findings will shed light on the implications of caregiver smartphone use behaviors. Providing interventions on smartphone devices or educational content on how caregiver smartphone use can impact caregiver or child mental health may be one way to address symptoms and ultimately improve caregiving behaviors and caregiver-child mental health.

Authors

• Julie V Cristello, Florida International University
  Presenting Author

• Jacqueline Duong, University of Texas at Austin
  Non-Presenting Author

• Natalia Simo Fiallo, Florida International University
  Non-Presenting Author

• Jonathan S. Comer, Florida International University
  Non-Presenting Author

• Adela C Timmons, Florida International University
  Non-Presenting Author