Smartphone "Biomarkers" Identify Individuals Vulnerable to Social Anxiety Disorder
Key findings
- In a study of "digital biomarkers," 59 undergraduates had smartphone data collected for two weeks (phone call content, text message timestamps and movement patterns) and completed questionnaires about social anxiety and depression
- 36% of students exhibited elevated symptoms of social anxiety disorder (SAD) on the Social Interaction Anxiety Scale
- Passive data collected via smartphone usage was strongly correlated with self-reported SAD severity (r = 0.70; p < 0.001)
- Smartphone biomarkers discriminated SAD symptoms from depression, negative affect and positive affect
Subscribe to the latest updates from Psychiatry Advances in Motion
Social anxiety disorder (SAD) is the persistent and exaggerated fear of evaluation in social situations that leads to avoidance. It is estimated to affect 13% of U.S. residents; though symptoms often go unreported and untreated because they are mistaken for shyness or perceived as a character flaw.
In recent years, smartphones have shown promise for monitoring and predicting individuals' behavior and psychological state. Features such as microphones and GPS provide insights into momentary behaviors that may represent psychopathology.
In the Journal of Medical Internet Research, Nicholas Jacobson, PhD, of the Geisel School of Medicine at Dartmouth College, Berta Summers, PhD, of the University of North Carolina Wilmington and researcher in the Department of Psychiatry at Massachusetts General Hospital, and Sabine Wilhelm, PhD, chief of Psychology at Mass General and the program director of the Obsessive-Compulsive Disorder and Related Disorders Program, report preliminary data suggesting smartphones can help identify individuals who have SAD.
Study Details
The researchers studied 59 U.S. undergraduate students who used their own Android devices. For two weeks, data were collected from three channels: phone call content, text message timestamps and accelerometer data reflecting movement patterns. The data were used to form digital biomarkers that were paired with machine learning models to predict the severity of participants' SAD symptoms.
To capture self-reported SAD symptom severity, the participants were asked to complete the Social Interaction Anxiety Scale; 36% evidenced elevated symptoms.
Because SAD is highly comorbid with depressive disorders, participants also completed the Depression Anxiety and Stress Scale as well as the Positive Affect Negative Affect Schedule.
Results
SAD symptom severity predicted via digital biomarkers was significantly correlated with self-reported SAD severity (r = 0.70; 95% CI: 0.54–0.81; p < 0.001).
Smartphone biomarkers also accurately discriminated SAD symptoms from:
- Depression (r = 0.36; 95% CI: 0.11–0.56; p = 0.005)
- Negative affect (r = 0.38; 95% CI: 0.14–0.58; p = 0.003)
- Positive affect (r = −0.14; 95% CI: −0.380 to 0.122; p > 0.05)
Broad Potential Implications
Smartphone data is already passively collected for most individuals in the U.S. "Digital phenotyping" may be a low-cost, low-burden way to identify people who are vulnerable to problematic levels of social anxiety.
Since smartphone biomarkers are based on observable phenomena rather than subjective feelings, they might help clinicians better conceptualize maladaptive thoughts and behaviors that underlie a broad range of psychiatric conditions.
view original journal article Subscription may be required
Learn more about the Obsessive-compulsive Disorder & Related Disorders Program
Learn more about research in the Department of Psychiatry