Deciphering data anomalies in BioSense

INTRODUCTION

Since June 2004, CDC's BioIntelligence Center has monitored daily nationwide syndromic data by using the BioSense surveillance application.

OBJECTIVES

The BioSense application has been monitored by a team of full-time CDC analysts. This report examines their role in identifying and deciphering data anomalies. It also discusses the limitations of the current surveillance application, lessons learned, and potential next steps to improve national syndromic surveillance methodology.

METHODS

Data on clinical diagnoses (International Classification of Diseases, Ninth Revision, Clinical Modifications [ICD-9-CM]) and medical procedures (CPT codes) are provided by Department of Veterans Affairs and Department of Defense ambulatory-care clinics; data on select sales of over-the-counter health-care products are provided by participating retail pharmacies; and data on laboratory tests ordered are provided by Laboratory Corporation of America, Inc. All data are filtered to exclude information irrelevant to syndromic surveillance.

RESULTS

During June-November 2004, of the approximately 160 data anomalies examined, no events involving disease outbreaks or deliberate exposure to a pathogen were detected. Data anomalies were detected by using a combination of statistical algorithms and analytical visualization features. The anomalies primarily reflected unusual changes in either daily data volume or in types of clinical diagnoses and procedures. This report describes steps taken in routine monitoring, including 1) detecting data anomalies, 2) estimating geographic and temporal scope of the anomalies, 3) gathering supplemental facts, 4) comparing data from multiple data sources, 5) developing hypotheses, and 6) ruling out or validating the existence of an actual event. To be useful for early detection, these steps must be completed quickly (i.e., in hours or days). Anomalies described are attributable to multiple causes, including miscoded data, effects of retail sales promotions, and smaller but explainable signals.

CONCLUSION

BioSense requires an empirical learning curve to make the best use of the public health data it contains. This process can be made more effective by continued improvements to the user interface and collective input from local public health partners.

BioSense: implementation of a National Early Event Detection and Situational Awareness System

BioSense is a CDC initiative to support enhanced early detection, quantification, and localization of possible biologic terrorism attacks and other events of public health concern on a national level. The goals of the BioSense initiative are to advance early detection by providing the standards, infrastructure, and data acquisition for near real-time reporting, analytic evaluation and implementation, and early event detection support for state and local public health officials. BioSense collects and analyzes Department of Defense and Department of Veterans Affairs ambulatory clinical diagnoses and procedures and Laboratory Corporation of America laboratory-test orders. The application summarizes and presents analytical results and data visualizations by source, day, and syndrome for each ZIP code, state, and metropolitan area through maps, graphs, and tables. An initial proof of a concept evaluation project was conducted before the system was made available to state and local users in April 2004. User recruitment involved identifying and training BioSense administrators and users from state and local health departments. User support has been an essential component of the implementation and enhancement process. CDC initiated the BioIntelligence Center (BIC) in June 2004 to conduct internal monitoring of BioSense national data daily. BIC staff have supported state and local system monitoring, conducted data anomaly inquiries, and communicated with state and local public health officials. Substantial investments will be made in providing regional, state, and local data for early event detection and situational awareness, test beds for data and algorithm evaluation, detection algorithm development, and data management technologies, while maintaining the focus on state and local public health needs.

Public Health Information Network--improving early detection by using a standards-based approach to connecting public health and clinical medicine

Public health departments and their clinical partners are moving ahead rapidly to implement systems for early detection of disease outbreaks. In the urgency to develop useful early detection systems, information systems must adhere to certain standards to facilitate sustainable, real-time delivery of important data and to make data available to the public health partners who verify, investigate, and respond to outbreaks. To ensure this crucial interoperability, all information systems supported by federal funding for state and local preparedness capacity are required to adhere to the Public Health Information Network standards.

Effect of telemedicine on health outcomes in 87 infants requiring neonatal intensive care

OBJECTIVE

This is an evaluation of a telemedicine system for the rapid interpretation of neonatal echocardiograms from a regional, level III neonatal intensive care unit (NICU). The use of telemedicine to support the cardiology needs of NICUs is increasing. However, there is very little published objective information regarding health outcomes or costs resulting from such telemedicine systems. The primary hypothesis tested was that the utilization of a telemedicine system for the interpretation of neonatal echocardiograms reduces the intensive care length of stay of low birthweight (LBW) infants.

STUDY DESIGN

All infants who were admitted to neonatal intensive care at New Hanover Regional Medical Center during the first six months of the system were studied by the use of echocardiograms. They were compared with infants who were born in the same period of the previous year. The outcome measures were the intensive care length of stay, rate of transfer to academic medical centers, and mortality rate.

RESULTS

A statistically non-significant reduction of 5.4 days in the intensive care length of stay (LOS) of low birthweight infants was observed (p = 0.37). The cost per echocardiogram transmitted was calculated at $33 compared to previous method of sending videotapes via overnight courier.

CONCLUSIONS

While the sample size was inadequate to demonstrate improvements in health outcomes, the magnitude of the change and the low costs of the system suggest that this intervention is practical for obtaining rapid diagnostic and treatment support. Larger studies are warranted to confirm these findings and determine whether faster diagnosis and earlier initiation of treatment improve health outcomes of newborn infants.

The Western New York Health Resources Project: developing access to local health information

The Western New York Health Resources Project was created to fill a gap in online access to local health information resources describing the health of a defined geographic area. The project sought to identify and describe information scattered among many institutions, agencies, and individuals, and to create a database that would be widely accessible. The project proceeded in three phases with initial phases supported by grant funding. This paper describes the database development and selection of content, and concludes that a national online network of local health data representing the various geographic regions of the United States would contribute to the quality of health care in general.