Expecting the Unexpected: Predicting Panic Attacks From Mood, Twitter, and Apple Watch Data.

Area of Science:

• Psychiatry
• Mental Health
• Behavioral Science

Background:

• Panic attacks are a significant mental health concern affecting 11% of adults annually.
• Current diagnostic criteria suggest panic attacks occur without warning.
• Emerging evidence indicates individuals can often identify triggers preceding panic attacks.

Purpose of the Study:

• To prospectively investigate qualitative and quantitative factors linked to the onset of panic attacks.
• To explore the relationship between environmental and physiological data and panic attack occurrence.
• To identify potential predictors for anticipating panic attacks.

Main Methods:

• Prospective study involving 87 participants experiencing panic attacks.
• Retrospective trigger identification and daily mood reporting.
• Analysis of wearable sensor data (ambient noise, heart rate) and Twitter mood ratings in a subsample (n=32).

Main Results:

• 95% of participants retrospectively identified a trigger for their panic attacks.
• Worsened individual mood and lower Twitter-based mood ratings correlated with increased next-day panic attack likelihood.
• Higher ambient noise and elevated resting heart rate were associated with a greater probability of experiencing a panic attack the following day.

Conclusions:

• Panic attack onset may be predictable, challenging the notion of attacks occurring without warning.
• Identified factors (mood, environmental, physiological) offer potential for developing predictive models.
• Findings support the development of novel prevention and intervention strategies for panic disorder.