254503 Improving the scale of study for spatio-temporal modeling of arthropod-borne zoonotic diseases

Monday, October 29, 2012

Stephen Jones, PhD , Medical Informatics, BlueCross BlueShield of Tennessee, Chattanooga, TN
William Conner, PhD , Dept of Forest Resources, Clemson University, Georgetown, SC
Background: Arthropod-borne zoonotic diseases vary geographically and occur in significant clusters. Spatio-temporal modeling of these diseases is often conducted at the county level which can mask smaller isolated high risk areas. Administrative medical claims data aggregated to the ZIP code level could improve disease surveillance activities while simultaneously protecting the identity of infected patients.

Research Objective: To compare spatio-temporal clustering outcomes using ZIP code versus county level administrative data for tick- and mosquito-borne diseases known to occur in Tennessee

Methods: A cross-sectional sampling was performed for 10 consecutive years (2000-2009) across 615 ZIP population-weighted code centroids and 95 counties in Tennessee. Disease incidence data were extracted from administrative medical claims data from a managed care organization (MCO). SaTScan™ software was used to detect significant clusters using a retrospective space-time permutation analysis.

Principal Findings: Cluster outcomes were mixed when comparing ZIP code to county level results. Significant Rocky Mountain spotted fever clusters were spatially and temporally identical across scales, the ZIP code Lyme disease cluster was nearly half the size of the county level cluster, and the ZIP code West Nile virus cluster was spatially and temporally identical to a non-significant county level cluster.

Conclusions & Implications: We demonstrated spatio-temporal modeling at the ZIP code scale using medical claims data from a MCO is possible, and may provide enhanced information compared to county-level assessments. Findings suggest these significant cluster areas have underlying geographic/habitat features that explain their existence. Focused disease/vector prevention efforts in non-endemic areas are warranted.

Learning Areas:
Biostatistics, economics
Environmental health sciences
Epidemiology
Public health biology

Learning Objectives:
1.Discuss the benefits of using administrative medical claims data in the tracking and reporting of zoonotic diseases 2.Compare significant disease clusters at two spatial scales (ZIP vs. county)

Keywords: Zoonoses, Geographic Information Systems

Presenting author's disclosure statement:

Qualified on the content I am responsible for because: Dr. Jones currently serves as a biostatistical research scientist and technical team lead for the predictive and spatial analytics group at the corporate headquarters of BlueCross BlueShield of Tennessee. His primary responsibilities include leading proactive disease identification and predictive modeling in order to improve health outcomes. Dr. Jones has extensive experience in spatial analytics, including utilizing space-time models for both retrospective and prospective disease surveillance activities.
Any relevant financial relationships? No

I agree to comply with the American Public Health Association Conflict of Interest and Commercial Support Guidelines, and to disclose to the participants any off-label or experimental uses of a commercial product or service discussed in my presentation.