Influenza

1) Describe Arizonaís integrated influenza surveillance for school children with a retrospective analysis of data from multiple sources including School-based Syndromic Surveillance Program (SSSP), laboratory-confirmed influenza case reports, sentinel influenza-like illness (ILI) reports, and hospital discharge data. 2) Demonstrate how ILI data collected from SSSP can be integrated into evidence from other data sources to prospectively monitor and detect early increases in influenza among school children.

Influenza surveillance provides public health officials and healthcare providers with data on the onset, duration, geographic location, and level of influenza activity in order to guide the local use of interventions. The Influenza Sentinel Provider Surveillance Network tracks influenza-like illness (% ILI) across the U.S. population. Objective: This presentation describes the use of influenza antiviral data from retail pharmacies to supplement influenza surveillance.

Hospital syndromic surveillance data may be a useful tool in detecting increases in influenza-like-illness (ILI) and for monitoring seasonal trends or pandemic activity on a local level. A previous comparison of hospital syndromic surveillance data with ILI surveillance data manually abstracted from emergency department notes revealed that the general respiratory category performed better than symptomspecific subcategories. However, only about half of all patients hospitalized for influenza meet the ILI criteria defined as fever and either cough or sore throat. Hospital discharge data are used retrospectively to determine disease burden, but is not of use for acute monitoring due to the substantial lag time. Knowing how accurately admission data reflect discharge data can assist with interpretation of real or near-real time data streams commonly used in syndromic surveillance systems.

Objective

Timely unplanned hospital admissions data in a general respiratory syndrome category and/or with a pneumonia or influenza admission diagnosis are compared with hospital discharge data to determine accuracy for prediction of influenza disease burden.

Accurate and precise estimation of disease rates for a given population during a specified time frame is a major concern for public health practitioners and researchers in biosurveillance. Many diseases follow distinct patterns; incidence and prevalence of many diseases increase approximately exponentially with age, including many cancers, respiratory infections, and gastroenteritis. With increasing demographic information available in biosurveillance systems leading to more complex and comprehensive disease databases, seeking concise and informative summary measures of disease burden over space and time is becoming more critical for public health surveillance. In this paper we present two summary measures of disease burden in the elderly that simultaneously reflect disease dynamics and population characteristics.

Objective

To better estimate disease burden in the elderly population we illustrate an approach—the Slope Intercept Modeling for Population Linear Estimation (SIMPLE) method—that summarizes age-specific disease rates in the 65+ population using the observed exponential increase in disease rates with age in this dynamic and rapidly growing population subgroup.

The H5N1 avian influenza virus is now considered endemic in poultry in some parts of the world and the continued exposure in humans suggests that the risk of the virus evolving into a more transmissible agent in humans − a step towards worldwide pandemic – remains high. Universities, with large assembly of students and student movements determined by the class schedules and travel routes between classes, in addition to the faculty and staff located in close proximity, are extremely susceptible environments to the spread of pandemic events. Moreover, large universities in the U.S. often have a good proportion of international students, who commute to/from their home country within their study period. Therefore, a good surveillance system to detect disease outbreaks is essential to support a system that is robust to this high impact low probability disruptive event.

Objective

This paper describes a framework for an aberration detection method − change-point analysis for mean and variance − adapted for Poisson-distributed data, for syndromic surveillance in an academic environment.