ISDS Conference

Since July 2004 the BioSense program at the Centers for Disease Control and Prevention (CDC) has received data from DoD military and VA outpatient clinics (not in real time). In January 2006 real-time hospital data (e.g. chief complaints and diagnoses) was added. Various diagnoses from all sources are binned into one or more of 11 syndrome categories.

Objective

This paper'­s objective is to compare syndromic categorization of newly acquired real-time civilian hospital data with existing BioSense data sources.

Many disease-outbreak detection algorithms, such as control chart methods, use frequentist statistical techniques. We describe a Bayesian algorithm that uses data D consisting of current day counts of some event (e.g., emergency department (ED) chief complaints of respiratory disease) that are tallied according to demographic area (e.g., zip codes).

Objective

We introduce a disease-outbreak detection algorithm that performs complete Bayesian Model Averaging (BMA) over all possible spatial distributions of disease, yet runs in polynomial time.

As major disease outbreaks are rare, empirical evaluation of statistical methods for outbreak detection requires the use of modified or completely simulated health event data in addition to real data. Comparisons of different techniques will be more reliable when they are evaluated on the same sets of artificial and real data. To this end, we are developing a toolkit for implementing and evaluating outbreak detection methods and exposing this framework via a web services interface.

In order to be best prepared to identify health events using electronic disease surveillance systems, it is vital for users to participate in regular exercises that realistically simulate how events may present in their system following disease manifestation in the community. Furthermore, it is necessary that users exercise methods of communicating unusual occurrences to other intra and extra-jurisdictional investigators quickly and efficiently to determine first, if an event actually exits and if one does its characteristics. A simulation exercise held in the National Capital Region (NCR) in the spring of this year exercised a novel format for engaging users while testing the utility of an embedded event communication tool.

Objective

This is a description of an innovative design and format used to exercise public health preparedness in a tri-jurisdictional disease surveillance system in the spring of 2006.

Since we donít know when such a disaster may occur, we have to perform this syndromic surveillance routinely, and thus the system should be automatic. Namely, information is drawn from electronic medical records (EMR), and is statistical analyzed, aberrations are detected and then Results are reported by e-mail or HP. It is preferable that this system be fully automatic. Though many systems of this type have been developed in the US, they have not been well developed in Japan. So as to develop such a system, we made a prototype system and have been performing prospectively and evaluating the system.

This paper describes a new class of space-time scan statistics designed for rapid detection of emerging disease clusters. We evaluate these methods on the task of prospective disease surveillance, and show that our methods consistently outperform the standard space-time scan statistic approach.

So as to be able to detect nosocomial outbreaks at an early stage, we examine the construction of a syndromic surveillance system for inpatients who have fever, respiratory symptoms, diarrhea, vomiting or rash and evaluate its statistical properties.

While several authors have advocated wavelets for biosurveillance, there are few published wavelet method evaluations using real syndromic data. Goldenberg et al. performed an analysis using wavelet predictions as a way of detecting a simulated anthrax outbreak. The commercial RODS application uses averaged wavelet levels to normalize for longterm trends and negative singularities. In line with the implementation in and in contrast to, we introduce two preconditioning steps to account for the strong day-of-week effect and holidays, and then use all levels of the wavelets to predict or alarm.

Objective

Syndromic data are created by processes that operate on different time scales (daily, weekly, or even yearly) and can include events of different durations from a 1-2 day outbreak of foodborne illness to a more gradual, protracted flu season. The duration of an outbreak caused by a new pathogenic strain or a bioterrorist attack is indeterminate. Wavelets are well suited for detecting signals of uncertain duration because they decompose data at multiple time and frequency scales. This study evaluates the use of several wavelet-based algorithms for both time series forecasting and anomaly detection using real-world syndromic data from multiple data sources and geographic locations.

A comprehensive definition of a syndrome is composed of direct (911 calls, emergency departments, primary care providers, sensor, veterinary, agricultural and animal data) and indirect evidence (data from schools, drug stores, weather etc.). Syndromic surveillance will benefit from quickly integrating such data. There are three critical areas to address to build an effective syndromic surveillance system that is dynamic, organic and alert, capable of continuous growth, adaptability and vigilance: (1) timely collection of high quality data (2) timely integration and analysis of information (data in context) (3) applying innovative thinking and deriving deep insights from information analysis. In our view there is excessive emphasis on algorithms and applications to work on the collected data and insufficient emphasis on solving the integration challenges. Therefore, this paper is focused on information integration.

Objective

EII is the virtual consolidation of data from multiple systems into a unified, consistent and accurate representation. An analyst working in an EII environment can simultaneously view and analyze data from multiple data sources as if it were coming from one large local data warehouse. This paper posits that EII is a viable solution to implement a system covering large areas and disparate data sources for syndromic surveillance and discusses case studies from environments external to health.

Our goal was to assess the impact of sentinel sample size and criteria for a signal on performance of daily prospective space-time permutation detection by comparing results in varying size random samples from a large health plan to results found in the full membership.