Application of Spatial Analysis on Electronic Health Records to Characterize Patient Phenotypes: Systematic Review

Area of Science:

• Health Informatics
• Geographic Information Systems (GIS) in Healthcare
• Spatial Epidemiology

Background:

• Electronic health records (EHRs) contain valuable patient address data for spatial analysis.
• No systematic review has previously assessed the use of spatial analysis for characterizing patient phenotypes from EHRs.

Approach:

• Systematic review of English-language, peer-reviewed studies from major databases (PubMed/MEDLINE, Scopus, Web of Science, Google Scholar) up to August 20, 2023.
• Evaluated studies using individual-level health data from EHRs within the United States.

Key Points:

• Over 85% of studies employed basic geocoding or mapping, with only 49 meeting eligibility for advanced spatial analysis.
• Clustering techniques were the predominant spatial method; spatiotemporal analysis and modeling were less common.
• Publication rates surged after 2017, focusing on diverse clinical areas and outcomes like asthma, hypertension, and diabetes, but rarely using genomics, imaging, or notes for phenotypes.

Conclusions:

• Growing interest exists in spatial analysis of EHR data for patient characterization.
• Significant knowledge gaps remain in clinical health, phenotype data domains, and spatial methodologies.
• Future research should address these gaps to enhance clinical decision support through spatial analysis.