A panic attack (PA) is an intense form of anxiety accompanied by multiple somatic presentations, leading to frequent emergency department visits and impairing quality of life. A prediction model for PAs could help clinicians and patients monitor, control, and do early intervention for recurrent panic attacks, enabling more personalized treatment for panic disorder. This study aimed to provide a seven-day PA prediction model and determine the relationship between a future PA and various features, including physiological factors, anxiety and depressive factors, and air quality index. We will enroll 200 participants (150 participants join case management with wearables study, 50 participants join TAU group) with PD (DMS-5 and MINI interview). Participants used smartwatches (Garmin vivosmart 4) and mobile applications to collect their sleep, heart rate, activity level, anxiety, and depression scores (BDI, BAI, STAI-S, STAI-T, and PDSS-SR) in their real-life for a duration of one year. We also included air quality indexes from open data. To analyze these data, our team used six machine learning methods: random forests, decision trees, LDA, AdaBoost, XGBoost, and regularized greedy forests, or other deep learning methods.

Participants enrollment Through primary psychiatrists' invitations, we will recruit 150 En Chu Kong Hospital psychiatric clinic PD participants and 50 patients from Mackay Memorial Hospital.

Inclusion criteria The inclusion criteria were 1) a primary diagnosis of PD by DSM-5, 2) more than 20 years old, and 3) an essential ability to navigate smartwatches and mobile phone applications. If participants did not fulfill criteria 3), we will invite them as the control group (i.e. TAU)

Exclusion criteria The exclusion criteria were 1) current substance abuse, 2) cardiopulmonary incapacity, 3) incapacity or limited mental capacity, and 4) acute suicidal ideation. This study required sufficient mental capacity on the part of participants to cooperate by wearing smartwatches continuously, properly maintaining the smartwatches, and completing regular, valid online questionnaires. Limited mental capacity implies that the participants have difficulties understanding, remembering, or using the information to make or communicate a decision. Our team evaluated participants' mental capacity during the diagnostic interview (DI), Mini International Neuropsychiatric Interview (MINI), and the informed consent process by certified psychiatrists and nurse practitioners. The information on acute suicidal ideation was from diagnostic interviews and questions from MINI part A and pre-assessment BDI.

Inclusion Criteria:

1. Diagnosis of panic disorder by DSM5
2. More than 18 years old
3. An essential ability to navigate smartwatch and mobile phone applications

Exclusion Criteria:

1. Current substance abuse
2. Cardiopulmonary incapacity
3. Incapacity or limited mental capacity
4. Acute suicidal ideation

Patients from outpatient clinics with a main diagnosis of Panic disorder