Skip to main content

Spatio - Temporal Scan

Description

Scan statistics are highly successful for the evaluation of space-time clusters. Recently, concepts from the graph theory were applied to evaluate the set of potential clusters. Wieland et al. introduced a graph theoretical method for detecting arbitrarily shaped clusters on the basis of the Euclidean minimum spanning tree of cartogram transformed case locations, which is quite effective, but the cartogram construction step of this algorithm is computationally expensive and complicated.

 

Objective

We describe a method for prospective space-time cluster detection of point event data based on the scan statistic. Our aim is to detect as early as possible the appearance of an emerging cluster of syndromic individuals because of a real outbreak of disease amidst the heterogeneous population at risk.

Submitted by hparton on
Description

Syndromic surveillance uses syndrome (a specific collection of clinical symptoms) data that are monitored as indicators of a potential disease outbreak. Advanced surveillance systems have been implemented globally for early detection of infectious disease outbreaks and bioterrorist attacks. However, such systems are often confronted with the challenges such as (i) incorporate situation specific characteristics such as covariate information for certain diseases; (ii) accommodate the spatial and temporal dynamics of the disease; and (iii) provide analysis and visualization tools to help detect unexpected patterns. New methods that improve the overall detection capabilities of these systems while also minimizing the number of false positives can have a broad social impact.

Submitted by elamb on
Description

Modern public health surveillance systems have great potential for improving public health. However, evaluating the performance of surveillance systems is challenging because examples of baseline disease distribution in the population are limited to a few years of data collection. Agent-based simulations of infectious disease transmission in highly detailed synthetic populations can provide unlimited realistic baseline data.

Objective

To create, implement, and test a flexible methodology to generate detailed synthetic surveillance data providing realistic geo-spatial and temporal clustering of baseline cases.

Submitted by elamb on
Description

The New York City Department of Health and Mental Hygiene (NYC DOHMH) collects data daily from 50 of 61 (82%) emergency departments (EDs) in NYC representing 94% of all ED visits (avg daily visits ~10,000). The information collected includes the date and time of visit, age, sex, home zip code and chief complaint of each patient. Observations are assigned to syndromes based on the chief complaint field and are analyzed using SaTScan to identify statistically significant clusters of syndromes at the zip code and hospital level. SaTScan employs a circular spatial scan statistic and clusters that are not circular in nature may be more difficult to detect. FlexScan employs a flexible scan statistic using an adjacency matrix design.

 

Objective

To use the NYC DOHMH's ED syndromic surveillance data to evaluate FleXScan’s flexible scan statistic and compare it to results from the SaTScan circular scan. A second objective is to improve cluster detection in by improving geographic characteristics of the input files.

Submitted by elamb on
Description

The space-time scan statistic is a powerful statistical tool for prospective disease surveillance. It searches over a set of spatio-temporal regions (each representing some spatial area S for the last k days), finding the most significant regions (S, k) by maximizing a likelihood ratio statistic, and computing p-values of these potential clusters by randomization.

The standard, "population-based" method assumes that, for each spatial location si on each day t, we have a population pti and a count (observed number of cases) cti. Then, under the null hypothesis of no clusters, we expect each count cti to be proportional to its population pti. We then search for regions (S, k) with disease rate (cases per unit population) significantly higher inside the region than outside. In the original space-time scan statistic, the populations are assumed to be given, and in, populations are estimated assuming independence of space and time.

Here we propose an alternative, "expectation-based" method, in which we infer the expected number of cases bti in each spatial location, based on the time series of previous counts. In this case, under the null hypothesis of no clusters, we expect each count cti to be equal to bti, rather than proportional to population. We then search for regions (S, k) with counts that are significantly higher than expected.

 

Objective

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.

Submitted by Sandra.Gonzale… on
Description

BioSense data includes Department of Defense and Veterans Affairs ambulatory care diagnoses and procedures, as well as Laboratory Corporation of America lab test orders. Data are mapped to eleven syndrome categories. SaTScan is a spatio-temporal technique that has previously been applied to surveillance at the metropolitan area level. Visualization of national results involves unique issues, including displaying cluster information that crosses jurisdictions, zip codes with highly variant data volume, and evaluating large multiple state clusters. SaTScan was first implemented in June 2005 in the BioSense application for daily monitoring at CDC’s BioIntelligence Center.

 

Objective

The objective is to describe the visualization and monitoring of the national spatio-temporal SaTScan results in the BioSense application. This is the first application of this algorithm to a national early event detection and situational awareness system.

Submitted by elamb on
Description

In this study, we compare two methods of generating grid points to enable efficient geographic cluster detection when the original geographical data are prohibitively numerous. One method generates uniform grid points, and the other employs quad trees to generate non-uniform grid points. We observe differences in the results of the spatial scan approach to cluster detection for both of these grid generation schemes. In both our simulated experiment, and our analysis of real data, the grid generation schemes produce different results. Generally speaking, the quad tree scheme is more sensitive to detecting high resolution spatial clusters than the uniform scheme. The quad tree grid point scheme may be a useful alternative to the uniform (and other) grid point generation schemes when it is important to set up a surveillance system sensitive to clusters at unspecified spatial resolutions. The quad tree grid scheme may also be useful in a number of other geographic surveillance applications.

Submitted by elamb on
Description

Outbreak detection algorithms for syndromic surveillance data are becoming increasingly complex. Initial algorithms focused on temporal data but newer methods incorporate geospatial dimensions. As methods evolve, it is important to understand the effects on detection of both algorithm parameters and population characteristics. Intensive, iterative data analyses are required to accomplish this. Even with leading-edge computer hardware, it can take weeks or months to complete analyses using advanced signal detection techniques such as the space-time scan statistic in the SaTScan program.

Given the strategic significance and national security implications of timely and accurate detection, proper tools for studying and thus improving increasingly complex surveillance algorithms are warranted.

 

Objective

We describe a method to perform computationally intensive analyses on large volumes of syndromic surveillance data using open-source grid computing technology.

Submitted by elamb on
Description

OBJECTIVE

A “whole-system facsimile” recreates a complex automated biosurveillance system running prospectively on real historical datasets. We systematized this approach to compare the performance of otherwise identical surveillance systems that used alternative statistical outbreak detection approaches, those used by CDC’s BioSense syndromic system or a popular scan statistics.

Submitted by elamb on