Guidelines for Constructing a Statewide Hospital Syndromic
Surveillance Network

Zygmunt F. Dembek, K. Carley, J. Hadler
Connecticut Department of Public Health, Hartford, Connecticut

Corresponding author: Zygmunt F. Dembek, Epidemiology Program, Connecticut Department of Public Health, 410 Capitol Ave., MS 11EPI, P.O. Box 340308, Hartford, CT 06134-0308. Telephone: 860-509-7994; Fax: 860-509-7910; E-mail: zygmunt.dembek@po.state.ct.us.

Disclosure of relationship: The contributors of this report have disclosed that they have no financial interest, relationship, affiliation, or other association with any organization that might represent a conflict of interest. In addition, this report does not contain any discussion of unlabeled use of commercial products or products for investigational use.

Abstract

Introduction: The process to initiate a comprehensive and inexpensive statewide hospital emergency department-based syndromic surveillance system (HEDSS) in Connecticut can serve as a template for others.

Objectives: With limited financial resources, the Connecticut Department of Public Health (CDPH) determined the requirements necessary to establish and routinely conduct hospital emergency department (HED)-based syndromic surveillance.

Methods: A statewide survey assessed ability and willingness of Connecticut hospitals to participate in HED syndromic surveillance. The New York City HED-based system protocol and analysis programs, available without financial charge, were modified for use in Connecticut. This system is based on hospitals sending daily standardized files of chief complaint data through encrypted e-mail or an FTP protocol to CDPH with subsequent categorization into syndromes using a SAS program. Anticipating regional surveillance needs during the Republican National Convention in New York City (RNC), CDPH initiated HEDSS in August 2004.

Results: Most Connecticut HEDs were willing and able to participate on a voluntary basis. Beginning in July 2004, hospital recruitment began. By the time of the RNC, 11 of 32 Connecticut hospitals participated in HEDSS. Since then, an additional six HEDs have joined.

Conclusion: Establishing a voluntary statewide HEDSS was possible using an existing, readily available protocol with minimum financial resources and consensus from a statewide workgroup over a several-month time period.

Introduction

Many public health jurisdictions have implemented emergency department (ED) chiefcomplaint-based syndromic surveillance systems as part of their effort at public health preparedness (1). In Connecticut, a hospital admissions syndromic surveillance system (HASS) had been implemented (2) but not an outpatient emergency department system. After observing the experience of others to determine the utility and appropriate level of investment, during 2004, the decision was made to develop a prospective hospital ED chief complaint-based syndromic surveillance system. The goal was to have readily available data to monitor and prospectively detect unusual disease activity at times of heightened public health alert (e.g., during the Republican National Convention [RNC] in New York City), and to investigate and monitor possible outbreaks, including influenza, as detected through other surveillance initiatives. The criteria for the system included voluntary participation, potential for automated transmission from participating EDs (no sustained labor involved), financial sustainability (if possible, no ongoing software maintenance or licensing costs), simplicity of operation (no sustained outside consultant help needed), and minimum daily personnel costs (5 day per week operation under most circumstances).

Methods

Consensus-Building and Readiness Assessment

A sequential consensus-building process was used to determine interest and readiness of hospital EDs to participate and to select a system with the desired characteristics. To determine interest and readiness, the Connecticut Department of Public Health (CDPH) met bimonthly for 8 months (November 2003--June 2004) with representatives from the state's hospitals to assess ED data collection status and determine the best method of instituting statewide hospital ED-based syndromic surveillance. For biologic terrorism planning purposes, Connecticut is divided into two hospital regions, from which representatives at the central hospitals in each region (i.e., Centers of Excellence) regularly meet with CDPH to conduct centralized planning for all hospitals. The workgroup consisted of administrative, medical, and information technology experts from the Centers of Excellence and representatives of CDPH and the Connecticut Hospital Association. An interest and technologic readiness survey was developed and a survey of all hospital EDs conducted during February 2004.

During workgroup meetings, it was decided to use an existing ED-based syndromic surveillance system rather than to design one de novo. Speakers from four existing systems were invited to present their systems to the workgroup. They included the New York City (NYC) syndromic surveillance system (3), the Harvard University National Bioterrorism Syndromic Surveillance Demonstration Program system (4), the Department of Defense system (ESSENCE) (5) and the University of Pittsburgh system (RODS) (6). Subsequently, a group decision was made as to which system to implement in Connecticut.

HEDSS Implementation and Operation

In July 2004, anticipating regional surveillance needs during the RNC, CDPH resurveyed the state's hospitals to determine their ability to participate in a voluntary ED syndromic surveillance effort by the end of August in time for the RNC. Initially, hospitals that began participation before the end of August 2004 were also promised $2,000 to cover relevant IT expenses to develop the ED data file for export and to write the electronic program for its creation and transmission on a daily basis. Later, the compensation period was extended to the end of September 2004. HEDSS was then initiated during late August 2004.

The HEDSS system is based on the New York City ED syndromic surveillance system protocol (3). To implement the system, a standardized file format was requested from participating hospitals for daily data transmission of ED visit data (Table). The file contains the hospital of origin; the time and date that each patient was seen in the ED; the age, sex, and home ZIP code of the patients; and chief-complaint information. Each day, participating hospitals send a file containing this information for each patient ED visit for the preceding 24-hour period. All participants use a secure method of data transmission including encrypted password-protected e-mail (two hospitals), daily e-mail notification allowing CDPH access to the hospital's password-protected secure FTP website (one hospital), and posting the file directly to the Connecticut secure FTP website (14 hospitals). Participating hospitals send CDPH unique daily data files and do not access each other's files.

The files are downloaded daily from the state website to an epidemiologist's desktop for merging and analysis. FileZilla shareware, available free from SourceForge.net (http://filezilla. sourceforge.net) is used as a viewer for this purpose. FileZilla is an FTP/SFTP client for Windows and can act as a site manager with the potential to perform multiple simultaneous transfers, Secure Sockets Layer (SSL) and Kerberos Generic Security Services (GSS) authentication/encryption with speed and a simple interface.

Chief-complaint data are transformed into syndromes for analysis using SAS code components developed by the New York City Department of Health and Mental Hygiene (NYCDOHMH). These were available free from NYCDOHMH and were modified for use by available staff. Daily analysis (using SAS macro codes developed by NYCDOHMH) generates tables for syndrome totals; syndrome by hospital; chief-complaint text; a line list; and syndromes indicative of anthrax, botulism, plague, and smallpox.

Results

The initial readiness survey indicated that 97% of the state's 32 acute-care hospitals collected ED chief complaint data, and 90% entered this data into an electronic database, most (97%) within 24 hours. In comparison, 53% entered a patient's discharge diagnosis into a database within 72 hours. A total of 77% of all hospitals could produce electronic extracts of ED chief-complaint data, 86% were willing to share deidentified data, and 76% would participate in an ED syndromic surveillance system. Ultimately, the system that was selected for adaptation in Connecticut was the NYC system. It was chosen for several reasons. First, using it would enable Connecticut to be fully compatible with New York in terms of definition of syndromes and exchange of information. Connecticut is part of the NYC metropolitan area with at least 100,000 commuters to the city daily. In addition, the cost for implementation and maintenance was very low. The system is simple for hospitals to participate in, the SAS program used for analysis was free and readily able to be adapted to Connecticut needs, and NYCDOHMH staff were available to share their experiences with implementation. All implementation work was done with available staff at CDPH.

Despite short notice during the summer, by the time of the RNC, 11 of 32 Connecticut hospitals were participating in HEDSS. A daily surveillance summary was sent to NYCDOHMH and in turn, a NYC metro-area syndromic surveillance report, including status reports from Connecticut, New York, New Jersey, NYC, and New Jersey counties, was provided daily by NYCDOHMH to all participating NYC metropolitan health agencies. Currently, 17 hospitals participate in the HEDSS.

Historically, if nonresidents have a reportable illness and are identified in a Connecticut hospital, this is reported to CDPH and in turn, the case is reported to the neighboring jurisdiction. If a syndromic disease alert occurred in hospitals within adjoining jurisdictions, communication of this information would be made to the pertinent health authorities. CDPH does not report Connecticutspecific patient data to out-of-state health authorities, nor do out-of-state health authorities report their residents' patient-identifiable data to CDPH.

The full utility of the HEDSS is being evaluated. The methodology used to evaluate the HEDSS efficacy includes:

• Comparisons of historic disease syndrome events with HASS. Those syndromes mutually reported (e.g., respiratory illness, fever, and rash illness) from daily HEDSS analysis are compared with biweekly HASS analysis. HEDSS analysis is completed within 48--72 hours after the event occurrence versus weekly for HASS. Therefore, HEDSS presents more recent data trends and, if accurate, widespread disease occurrences should later be reflected in HASS and through community-reported disease.
• Comparisons with other area-wide hospital syndromic disease events. For example, during the RNC, daily syndromic disease trends and significant events were shared by the respective health agencies between Connecticut, NYC, New York state and neighboring New York counties, and New Jersey. Large disease events (e.g., an influenza outbreak) should be reflected in syndromic surveillance reporting of shared illness within neighboring health jurisdictions.
• The state's total burden of ED visits and disease syndromes, to include those hospitals not participating in HEDSS, is unknown. However, the 17 hospitals in HEDDS are located in seven of the eight Connecticut counties and comprise approximately 65% of the number of staffed hospital beds statewide. Six of these hospitals are located in the two counties (Litchfield and Fairfield) adjoining New York and comprise approximately 70% of the staffed beds in those counties. As more hospitals participate in HEDSS, data will reflect whether the State's ED visits and disease syndromes are reflected in similar proportions.
• Disease trend analysis will continue to be conducted for counties or other geographic areas by examining ZIP code information (spatial trends), and for ED visit time (hourly) or successive days (temporal trends).

Discussion

An important factor for success of HEDSS was the commitment for statewide hospital participation. The CDPH-Hospital workgroup helped establish the interest and requirements for syndromic surveillance in Connecticut by mutual agreement through regular meetings over an 8-month period.

Although HEDSS is simple conceptually, generation of compatible datasets from different entities for merging is difficult. The most difficult facet of obtaining successful data transmission by each participating hospital to CDPH was ensuring that comprehensive data specifications were met for individual hospital ED datasets. This effort was successful once the responsible hospital IT staff or hospital IT contractor fully understood HEDSS data requirements and agreed to participate. It was often necessary for the hospital IT staff/contractor to identify a responsible person within their organization to perform the initial data formatting for the hospital's chief-complaint data to ensure that the data were automatically sent daily to CDPH and to have an established contact person should lapses in formatting or data transmission occur.

The full utility of HEDDS is still to be determined. The program is being assessed to monitor the onset and level of activity of influenza, and results are being compared with those obtained from the HASS. No direct patient identification information is collected by HEDDS (i.e., patient name, social security number, or hospital identification number). However, follow-up on individual case reports can be done with a phone call to the hospital ED to obtain casespecific data. The time of the patient visit to the ED and the patient's age and sex make it easy for the hospital to identify the person. HEDSS might be useful to identify individual cases of disease syndromes of interest during investigation of possible outbreaks.

Every report of adult onset chickenpox is followed by a CDPH case investigation, including cases identified by HEDSS. Pediatric cases of chicken pox also are examined to determine if they have been reported to CDPH through other reporting mechanisms. Given the 48--72 hour time lag from case presentation in the ED to CDPH analysis, complete case-specific information is obtained within 49--96 hours after a patient was seen at the hospital ED.

Conclusion

Establishing a voluntary statewide hospital ED syndromic surveillance system is possible using an existing readily available protocol with minimum financial resources over a several-month time period. Connecticut has a statewide hospital ED surveillance network that includes approximately 50% of all acute care hospitals and 65% of the state's hospital capacity. The success in implementing HEDSS is due, in part, to the consensus building that occurred between CDPH and the state's hospitals. HEDSS can be established by any state, large city, or county that wants to have hospital ED-based syndromic surveillance with a minimum investment and IT support.

Acknowledgments

The authors wish to acknowledge the contributions of the Centers of Excellence ED syndromic surveillance workgroup including Mark Tepping of Bridgeport Hospital, Christopher Cannon, Elaine Chapman, Elaine Forte, and Cindy Pugliese of Yale New Haven Hospital and Sandra Brown, Richard Kubica, Vivian Lane, Patricia Montanaro and Dr. Alan J. Smally of Hartford Hospital.

References

1. Sosin DM. Syndromic surveillance: the case for skillful investment. Biosurety and Bioterrorism: Biodefense Strategy, Practice and Science 2003;1:247--53.
2. Dembek ZF, Carley K, Siniscalchi A, Hadler J. Hospital admissions syndromic surveillance---Connecticut, September 2001-November 2003. MMWR 2004;53:50--2.
3. Heffernan R, Mostashari F, Das D, Karpati A, Kuldorff M, Weiss D. Syndromic surveillance in public health practice, New York City. Emerging Infectious Diseases 2004;10:858--64.
4. Yih WK, Caldwell B, Harmon R, et al. National bioterrorism syndromic surveillance demonstration program. MMWR 2004;53:43--6.
5. Lombardo JS, Burkom HS, Elbert YA, et al. A systems overview of the Electronic Surveillance System for the Early Notification of Community-based Epidemics (ESSENCE II). J Urban Health 2003;80:32--42.
6. Wagner MM, Espino J, Tsui F-C, et al. Syndrome and outbreak detection using chief-complaint data: experience of the real-time outbreak and disease surveillance project. MMWR 2004;53:28--31.

Use of trade names and commercial sources is for identification only and does not imply endorsement by the U.S. Department of Health and Human Services. References to non-CDC sites on the Internet are provided as a service to MMWR readers and do not constitute or imply endorsement of these organizations or their programs by CDC or the U.S. Department of Health and Human Services. CDC is not responsible for the content of pages found at these sites. URL addresses listed in MMWR were current as of the date of publication.

Disclaimer   All MMWR HTML versions of articles are electronic conversions from ASCII text into HTML. This conversion may have resulted in character translation or format errors in the HTML version. Users should not rely on this HTML document, but are referred to the electronic PDF version and/or the original MMWR paper copy for the official text, figures, and tables. An original paper copy of this issue can be obtained from the Superintendent of Documents, U.S. Government Printing Office (GPO), Washington, DC 20402-9371; telephone: (202) 512-1800. Contact GPO for current prices.

Date last reviewed: 8/5/2005

HOME  |  ABOUT MMWR  |  MMWR SEARCH  |  DOWNLOADS  |  RSS |  CONTACT
POLICY  |  DISCLAIMER  |  ACCESSIBILITY