Skip to main content

CUSUM

Current biosurveillance systems run multiple univariate statistical process control (SPC) charts to detect increases in multiple data streams. The method of using multiple univariate SPC charts is easy to implement and easy to interpret. By examining alarms from each control chart, it is easy to identify which data stream is causing the alarm. However, testing multiple data streams simultaneously can lead to multiple testing problems that inflate the combined false alarm probability. Although methods such as the Bonferroni correction can be applied to address the multiple testing problem by lowering the false alarm probability in each control chart, these approaches can be extremely conservative. Biosurveillance systems often make use of variations of popular univariate SPC charts such as the Shewart Chart, the cumulative sum chart (CUSUM), and the exponentially weighted moving average chart (EWMA). In these control charts an alarm is signaled when the charting statistic exceeds a pre-defined control limit. With the standard SPC charts, the false alarm rate is specified using the in-control average run length (ARL0). If multiple charts are used, the resulting multiple testing problem is often addressed using family-wise error rate (FWER) based methods that are known to be conservative - for error control. A new temporal method is proposed for early event detection in multiple data streams. The proposed method uses p-values instead of the control limits that are commonly used with standard SPC charts. In addition, the proposed method uses false discovery rate (FDR) for error control over the standard ARL0 used with conventional SPC charts. With the use of FDR for error control, the proposed method makes use of more powerful and up-to-date procedures for handling the multiple testing problem than FWER-based methods.

Objective: To propose a computationally simple, fast, and reliable temporal method for early event detection in multiple data streams.

Submitted by elamb on

Recently published studies evaluate statistical alerting methods for disease surveillance based on detection of modeled signals in a data background of either authentic historical data or randomized samples. Differences in regional and jurisdictional data, collection and filtering methods, investigation resources, monitoring objectives, and systemrequirements have hindered acceptance of standard monitoring methodology. The signature of a disease outbreak and the baseline data behavior depend on various factors, including population coverage, quality and timeliness of data, symptomatology, and the careseeking behavior of the monitored population. For this reason, statistical process control methods based on standard data distributions or stylized signals may not alert as desired. Practical algorithm evaluation and adjustment may be possible by judging algorithmperformance according to the preferences of experienced human monitors.

 

Objective

This presentation gives a method of monitoring surveillance time series on the basis of the human expert preference. The method does not require detailed history for the current series, modeling expertize, or a well-defined data signal. It is designed for application to many data types and without need for a sophisticated environment or historical data analysis. 

Referenced File
Submitted by hparton on

Pandemic 2009 H1N1 influenza and recent H7N9 influenza outbreaks made the public aware of the threat of influenza infection. In fact, annual influenza epidemic caused heavy disease burden and high economic loss around the world [1, 2]. Although the virological surveillance provided the high sensitivity and specificity for testing results, the timeliness and the cost of the test were not feasible for extensive public health surveillance. In addition, traditional sentinel physician surveillance also encountered many challenges such as the representativeness and reporting bias. The seamless surveillance system without extra labor reporting would be the ideal approach. Taiwan had as high as 99% of health insurance coverage. The real-time monitoring of the ILI clinical visits in the communities could reflect the severity of influenza epidemics. In this study, we used an innovative two-stage approach for detecting aberrations during 2009 pandemic influenza in Taiwan.

Objective

This study proposed a two-stage approach for early detection of aberrations of influenza-like illness (ILI) using the small-area based claim data of outpatient and emergency room visit.

Submitted by elamb on

Absenteeism has great advantages in promoting the early detection of epidemics. School absenteeism surveillance could timely detect the aggregations of absentees in time and space, so as to provide effective early warning and prevention and control of infectious diseases outbreaks in schools. Since April 1, 2012, an integrated syndromic surveillance system (ISSC) has been implemented in rural Hubei Province, China. With school absence data, finding the optimal model and related appropriate parameters for early warning of epidemics is necessary and practical.

Objective

To explore the optimal model and its related parameters via EWMA and CUSUM (C1, C2, C3) models in school absenteeism surveillance for early detection of infectious disease outbreaks in rural China.

Submitted by knowledge_repo… on

One criterion for evaluating the effectiveness of a surveillance system is the system’s positive predictive value. To our knowledge few studies have described the positive predictive value of syndromic surveillance signals for naturally occurring conditions of public health importance.

 

Objective

We evaluated the positive predictive value of signals detected by our syndromic surveillance system.

Referenced File
Submitted by elamb on

Although the majority of work in syndromic surveillance has been its application to bioterrorism and infectious diseases, one of the emerging priorities for its use is for the monitoring of environmental health conditions. Heat-related illness (HRI) is of growing public health importance, especially with global warming concerns and increased frequency of heat waves. Ambient temperatures are responsible for significant morbidity and mortality, as was demonstrated during the 1995 heat wave in Chicago that resulted in over 700 excess deaths and 33,000 emergency room visits due to HRI. A syndromic surveillance system that is able to detect early indications of excess HRI may start the public health response earlier, and thus reduce associated morbidity and mortality. The utility of 911 ambulance dispatch data for the early detection of heat-related illness was explored.

 

Objective

This paper describes the use of 911 ambulance dispatch data for the early detection of HRI in Toronto, Ontario, Canada.

Referenced File
Submitted by elamb on

The South Carolina (SC) Department of Health and Environmental Control uses multiple surveillance systems to monitor influenza activity from October to May of each year, including participating in the U.S. Influenza Sentinel Providers Surveillance Network. A percentage of influenza-like-illness surpassing the national 2.5% baseline is considered evidence of increased influenza activity by the CDC; this baseline is historical and does not change throughout the influenza season. Though not a part of the national influenza surveillance, SC also requires health care providers in the state to report positive rapid influenza tests, by number, on a weekly basis. Currently, only a trend analysis is used on weekly reports of positive rapid influenza test data for SC. A more robust method for determining statistically significant increases in activity for these two influenza surveillance systems is needed, and would provide a more accurate assessment of the status of seasonal influenza activity in SC.

 

Objective

Use the Early Aberration Reporting System (EARS) to analyze influenza sentinel provider surveillance data and positive rapid influenza test reports to identify weeks where influenza activity was significantly increased in South Carolina. Demonstrate the utility of using EARS to detect increases in influenza activity using existing surveillance systems.

Referenced File
Submitted by elamb on

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.

Referenced File
Submitted by elamb on

San Diego County Public Health has been conducting syndromic surveillance for the past few years. Currently, the system has become largely automated and processes and analyzes data from a variety of disparate sources including hospital emergency departments, 911 call centers, prehospital transports, and over-the-counter drug sales. What has remained constant since the system’s initial conceptualization is the local opinion that the data should be analyzed and interpreted in a variety of ways, in anticipation for the variety of contexts in which events that are of public health interest may unfold. Relatively small increases in volume that are sustained over time will likely be detected by methods designed to detect “small process shifts”, and include the CUSUM and EWMA methods. Larger increases in volume that are not sustained over time will likely be detected by other employed methods (P-Chart in the event of a non-proportional increase in volume, U-Chart in the event of a proportional increase in volume). A retrospective analysis was conducted on historical data from various data sources to determine the frequency of signals and detected events as well as the context within which the alert occurred (i.e., the “shape” of the data). Findings regarding several actual public health events will also be discussed.

 

Objective

This paper describes the frequency, various “shapes” and magnitudes of data anomalies, and varying ways actual public health events may present themselves in syndromic data.

Referenced File
Submitted by elamb on

Analysis of time series data requires accurate calculation of a predicted value. Non-regression methods such as the Early Aberration Reporting System CuSum are computationally simple, but most do not adjust for day of week or holiday. Alternately, regression methods require larger counts, more computer resources, and possibly longer baseline periods of data. As increasing volumes of data are reported and analyzed, the predictive accuracy of simpler methods should be assessed and optimized.

 

Objective

To compare the predictive accuracy of three non-regression methods in analysis of time series count data.

Referenced File
Submitted by elamb on
Subscribe to CUSUM