Predicting Panic Disorder with Socioeconomic, Physiological, and Behavioral Markers: A Machine Learning Study

Panic disorder (PD) is a debilitating psychiatric condition characterized by recurrent, unexpected panic attacks that significantly impair an individual's functioning across multiple domains. Despite its prevalence and impact, early detection remains challenging. This study utilized data from the National Health and Nutrition Examination Survey (NHANES) from 1999 to 2004 to develop a machine learning (ML)-based predictive model for PD. Our integrated approach revealed that PD risk is influenced by an association of physiological, socioeconomic, and behavioural factors. Key predictive variables included cardiovascular measures (blood pressure patterns, cardiovascular fitness status), weight-related factors (weight at age 25, weight history), pain indicators (low back pain), and socioeconomic determinants (poverty-to-income ratio, education level). The model achieved a precision of 0.96 and recall of 0.81, demonstrating robust performance in distinguishing individuals at risk for PD. Our findings revealed associations between basic cardiovascular measures (blood pressure patterns, cardiovascular fitness) and PD risk, alongside socioeconomic and behavioural factors. These results suggest that integrating multidimensional data into predictive models can enhance clinical screening protocols and early intervention strategies for PD.