Predicting Depression with GPS Data in Germany During COVID-19
Author Information
Author(s): Taiane de Azevedo Cardoso, Matilda Hamlin, Brent Davis, Johanna-Sophie Edler, Michael Winter, Holger Steinmetz, Caroline Cohrdes, Harald Baumeister, RĂ¼diger Pryss
Primary Institution: Robert Koch Institute
Hypothesis
Can GPS-based environmental factors predict self-reported depressive symptoms during the COVID-19 pandemic?
Conclusion
The study found that passive data from GPS can help predict depressive symptoms and identify regional risk factors.
Supporting Evidence
- The final elastic net regression model explained 6% of the variance in depressive symptoms.
- Three main predictors were identified: current COVID-19 infections, places in nursing homes, and proportion of fathers receiving parental benefits.
- The study utilized a data-driven statistical approach to analyze a large number of predictors.
Takeaway
This study used smartphone GPS data to see if it could help understand depression during the pandemic, and it found some useful patterns.
Methodology
Participants used the CORONA HEALTH app to provide GPS data and complete the Patient Health Questionnaire-9 to assess depressive symptoms.
Potential Biases
Selection bias may exist due to the requirement of Android smartphones and consent for passive data collection.
Limitations
The study's cross-sectional design limits causal conclusions, and the sample size may weaken significance.
Participant Demographics
Participants were German-speaking adults aged 18 and older, with a final sample of 249 after exclusions.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website