Back to Annual Meeting

Back to Annual Meeting

APHA Scientific Session and Event Listing

Chiehwen Ed Hsu, PhD, MPH¹, Yuwen Chiu, MD, MSPH¹, Jerry A. Miller, MS, PhD², and Francisco Soto Mas, MD, PhD, MPH³. (1) Department of Public and Community Health, University of Maryland, 2387 HHP Building, Valley Drive, College Park, MD 20742, 301-405-8161, edhsu@umd.edu, (2) National Center for Birth Defects and Developmental Disabilities, Centers for Disease Control and Prevention, 1600 Clifton Road, MS E-86, Atlanta, GA 30333, (3) Health & Literacy Program, University of Texas at El Paso, College of Education, El Paso, TX 79968-0574

Purposes: To determine the excess human infection clusters of Avian Flu in the world as of May 2006, we examined whether spatiotemporal clusters are still alive, and determined what geographic regions, time, and level of risk the clusters present. Methods: Data were collected from the WHO website where weekly reported human H5N1 infection cases between January 1st, 2004 and May 23rd, 2006 were tallied. Two hundred eighteen cases were analyzed for the 2.5-year study period. The space-time permutation method (both retrospective and prospective) of Spatial Scan Statistic pioneered by Kulldorff et al. was employed to detect potential clusters of the disease in 10 countries. Clusters were detected at a circle of maximum 400 km radius, and a temporal period of 1 week and 1 month, respectively. Results: Eleven clusters in 5 countries were identified. In retrospective analysis, Turkey, Egypt, and Azerbijan were detected as likely clusters, followed by Thailand and Indonesia. Prospective analysis found that Egypt and Indonesia are the most likely clusters contingent upon the temporal settings of either a week or month (see Table 1 and 2). Discussion: Detected clusters varied by the temporal settings of analysis defined by investigators. Egypt has the most likely cluster in prospective analysis when a cluster is defined as 1 month temporal period, but Indonesia has the most likely cluster if a cluster is defined as 1 week period. Recently outbreaks in Egypt should warrant equal, if not more attention than Indonesian one.

Learning Objectives:

• Recognize the application of public health informatics, such as using the space-time permutation model could help identify disease clusters.
• Assess the importance of disease clusters using statistical significance, including retrospective and prospective methods.
• Evaluate the methodology from the perspective of disease outbreak and persistence, periodicity of case surveillance data, and parameters of the model.

Keywords: Health Information, Risk Communication

Presenting author's disclosure statement:

Not Answered