Cluster Detection Using Scan Statistics and Risk Factors for Encephalitis from West Nile Virus Infection

• Activity Number: 480
• Type: Contributed
• Date/Time: Wednesday, August 4, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Sarah Baraniuk*+ and Kristy Murray
• Companies: The University of Texas Health Science Center at Houston and The University of Texas Health Science Center at Houston
• Keywords: scan statistics ; West Nile Virus ; spatial statistics ; survival analysis

Abstract

Risk factors for developing encephalitis from West Nile virus (WNV) have not been clearly identified (with the exception of advanced age). An important consideration in any analysis relating to the risk of contracting this virus includes available spatial or temporal factors. A retrospective medical chart review was done on confirmed WNV cases that were hospitalized in Houston between 2002 and 2008. It was found that of the 229 cases with available location data 132 were found to have severe disease outcomes (defined as those with encephalitis, meningoencephalitis, or encephalomyelitis). In combination with the length of time from onset to disease progression status we will adopt an exponential model within the scan statistic framework to investigate potential spatial factors. Both adjusted and unadjusted analyses will be discussed.