Prediction of Positive Patient Health Questionnaire-2 Screening Using Area Deprivation Index in Primary Care

Area of Science:

• Public Health
• Machine Learning in Healthcare
• Social Determinants of Health Research

Background:

• Major depressive disorder affects millions, posing a significant public health challenge in the US.
• Primary care settings are crucial for identifying individuals at risk for depression.
• Understanding the impact of social determinants of health (SDoH) on depression screening is vital for targeted interventions.

Purpose of the Study:

• To predict Patient Health Questionnaire-2 (PHQ-2) positivity in primary care patients.
• To analyze the relationship between SDoH, including the Area Deprivation Index (ADI), and PHQ-2 positivity.
• To identify key demographic, behavioral, and socioeconomic factors associated with increased depression risk.

Main Methods:

• Retrospective analysis of 74,636 electronic health records from 15 South Florida primary care clinics.
• Utilized machine learning algorithms (over 40 explored) to predict PHQ-2 positivity (cutoff >2).
• Incorporated SDoH via ADI (using zip+4), demographics, health behaviors, and community involvement.

Main Results:

• Random Forest algorithm demonstrated outstanding performance in predicting PHQ-2 positivity.
• Key predictors identified include Area Deprivation Index (ADI), age, education, visit type, coffee intake, and marital status.
• Model accuracy varied across clinics, highlighting localized predictive power.

Conclusions:

• The Area Deprivation Index (ADI), serving as a proxy for SDoH, is a significant predictor of PHQ-2 positivity.
• Individual factors combined with SDoH effectively predict depression screening outcomes.
• Findings can inform health organizations for proactive health needs assessment and resource allocation.