Early Detection of Disease Outbreaks with Applications in New York City

• Activity Number: 167
• Type: Topic Contributed
• Date/Time: Tuesday, August 13, 2002 : 8:30 AM to 10:20 AM
• Sponsor: ENAR
• Author(s): Martin Kulldorff*+ and Farzad Mostashari and Jim Miller
• Affiliation(s): University of Connecticut and New York City Department of Health and New York City Department of Health
• Keywords: disease surveillance ; spatial statistics ; scan statistic ; GIS ; geography ; New York

Abstract

The early detection of disease outbreaks is important so that disease control and preventive measures can be implemented as early as possible when new health hazards emerge, thereby minimizing the public health consequences of the outbreaks. Many disease outbreaks are local in nature. In this talk we describe new statistical methods for prospective time-periodic surveillance systems that are designed to detect localized outbreaks without a priori assumptions about their geographical location or size, or their temporal duration. The statistical inference is adjusted for the multiple testing inherent in the many locations, sizes, and time durations considered, as well as the repeated analyses over time. The methods are illustrated with data from New York City, including West Nile Virus surveillance from the summer of 2001, as well as syndromic disease surveillance using daily hospital emergency admissions data during 2001/02.