Authors: Brittany Krzyzanowski*, University of Minnesota - Minneapolis
Topics: Geographic Information Science and Systems
Keywords: Regionalization, Privacy, Neighborhood Health
Session Type: Paper
Presentation File: No File Uploaded
There is a general lack of understanding of the HIPAA privacy standards that regulate how aggregated spatial data can be shared. Misinterpretation of these standards impacts the way researchers share mapped protected health information (PHI). Maps are not present in the majority of neighborhood health publications, and the few studies that do share maps do so at the overly conservative county-level. To address these issues, we need to improve spatial health researcher's understanding of HIPAA privacy regulations, so they can be confident to share data at finer scales while also staying in compliance with HIPAA standards. The proposed project explores the challenge of sharing finer-scale PHI while maintaining patient privacy by using regionalization to create higher resolution HIPAA-compliant geographical aggregations. Four different regionalization approaches (max-p-regions, REDCAP, AMOEBA, and SOM) will be used to each create a configuration of regions that align with census boundaries, optimize intra-unit homogeneity, and maximize the number of spatial units while meeting the minimum population required under HIPAA guidelines (20,000 persons per unit). The relative utility of each configuration will be assessed according to model-fit, homogeneity, compactness, suppression, and geographic information loss. The primary goal of this research is to make health data more useful and accessible by enabling researchers to share higher-resolution HIPAA-compliant data at a much finer scale than what the research community is currently relying on.