Understanding the Relationship Between Mood Symptoms and Mobile App Engagement Among Patients With Breast Cancer...

Area of Science:

• Digital mental health
• Oncology
• Psychometrics

Background:

• Mobile health interventions are crucial for managing cancer treatment side effects.
• Understanding the link between patient mood and app engagement is vital for treatment success.
• A data-driven approach is needed to analyze mood and app engagement in cancer patients.

Purpose of the Study:

• To outline a step-by-step process for predicting mood in breast cancer patients using app engagement data.
• To demonstrate the application of this process using data from a mobile mental health app.

Main Methods:

• Data preprocessing, feature extraction, and predictive modeling were employed.
• Engagement patterns were compared between patients with high/low anxiety and depression.
• Random forest and XGBoost classifiers were evaluated for mood prediction accuracy.

Main Results:

• Engagement patterns differed between high and low anxiety/depression groups.
• Predictive models showed varying accuracy; the best model achieved 84.6% accuracy.
• A combination of survey and app engagement features with a random forest classifier yielded the highest performance.

Conclusions:

• The developed analytic process is feasible for understanding mood and app engagement in breast cancer patients.
• Predicting mood using self-report and engagement features can improve clinical decision-making.
• This approach supports the development of personalized digital mental health interventions for cancer patients.