Influenza Surveillance

Influenza causes significant morbidity and mortality, with attendant costs of roughly $10 billion for treatment and up to $77 billion in indirect costs annually. The Centers for Disease Control and Prevention conducts annual influenza surveillance, and includes measures of inpatient and outpatient influenza-related activity, disease severity, and geographic spread. However, inherent lags in the current methods used for data collection and transmission result in a one to two weeks delay in notification of an outbreak via the Centers for Disease Control and Prevention’s website. Early notification might facilitate clinical decision-making when patients present with acute respiratory infection during the early stages of the influenza outbreak.

In the United States, the influenza surveillance season typically begins in October and continues through May. The Utah Health Research Network has participated in Centers for Disease Control and Prevention’s influenza surveillance since 2002, collecting data on outpatient visits for influenza-like illness (ILI, defined as fever of 100F or higher with either cough or sore throat). Over time, Utah Health Research Network has moved from data collection by hand to automated data collection that extracts information from discrete fields in patients’ electronic health records.

We used statistical process control to generate surveillance graphs of ILI and positive rapid influenza tests, using data available electronically from the electronic health records.

Objective

The objective of this study is to describe the use of point-ofcare rapid influenza testing in an outpatient, setting for the identification of the onset of influenza in the community

The American Health Information Community Harmonized Use Case for the Biosurveillance minimum data set (MDS) was implemented to establish data exchange between regional health information organizations (RHIOs) and the New York State Department of Health (NYSDOH) for accelerating situational awareness through the Health Information Exchange (HIE) Project. However, the completeness, timeliness of the reporting and quality of data elements in the MDS through RHIOs are still unknown and need further validation before we can utilize them for NYSDOH public health surveillance.

Objective

Evaluate the availability, timeliness, and accuracy of MDS data elements received from one RHIO for emergency department (ED), in-patient, and outpatient visits. Compare the characteristics of patients meeting the HIE influenza-like illness definition who were admitted to the hospital or expired versus those discharged home.

Current influenza-like illness monitoring in Idaho is on the basis of syndromic surveillance using laboratory data, combined with periodic person-to-person reports collected by Idaho state workers. This system relies on voluntary reporting. Electronic medical records with relational databases offer a method of obtaining data in an automated fashion. Clinical data entered in CPRS includes real-time visit information, vital signs, ICD-9, pharmacy, and labs. ICD-9 and vital signs have been used to predict influenza-like illness in automated systems. We sought to combine these with lab and pharmacy data as part of an automated syndromic surveillance system.

Objective

The objective of this paper is to study whether syndromic surveillance using data from the Veterans Administration electronic medical record computerized patient record system (CPRS) correlates to officially reported influenza activity levels in the State of Idaho.

The HEDSS system was implemented in 2004 to monitor disease activity.1 In all, 18 of 32 emergency departments (ED) and urgent care clinic provide data. Chief complaints are routinely categorized into eight syndromes. The fever/flu syndrome is used for early detection and monitoring of influenza in the community. Area-based measures, such as zip code, enable linkage to area-based socioeconomic census data. Neighborhood poverty, defined as the percentage of persons living below the federal poverty level in a geographic area, predicts a wide range of disease outcomes.

Objective

To describe the relationship between neighborhood poverty and emergency department visits for fever/flu syndrome illnesses reported through the Connecticut Hospital Emergency Department Syndromic Surveillance (HEDSS) system.

Unfortunately, confirmation and notification of all A/H1N1 (2009) patients in Japan was ceased on 24 July when the cumulative number of patients was about 5000. After that, as all suspected patients are not necessarily confirmed or reported, the only official surveillance was the sentinel surveillance for influenza-like-illness (ILI) patients from 5000 clinics accounting for almost 10% of all clinics and hospitals in Japan. However, because the surveillance results are reported weekly, it tends to lack timeliness. To collect and analyze the information in more timely manner, we, Infectious Disease Surveillance Center, National Institute of Infectious Diseases, developed a full automatic daily reporting system of ILI patients. Using this information, we had estimated Rv and predict its course in every week.

Objective

This paper summarized our effort for real-time estimation of pandemic influenza A/H1N1pdm in Japan.

Real-time emergency department (ED) data from the BioSense surveillance program for ILI visits and ILI admissions provide valuable insight into disease severity that bridges gaps in traditional influenza surveillance systems that monitor ILI in outpatient settings and laboratory-confirmed hospitalization, but do not quantify the relationship between ILI visits and hospital admissions.

Objective

The purpose of this analysis is to gain understanding of the burden of influenza in recent years through analysis of clinically rich hospital data. Patterns of visits and severity measures such as the ratio of admissions related to influenzalike illness (ILI) by age group from 2007 to 2010 are described.

The data elements required for the proper functionality of VA’s ESSENCE system are all currently available within VA’s 128 VistA systems. These data are made available to VA’s ESSENCE system via a series of complicated MUMPS extraction routines, multiple data transformations crossing multiple servers, networks, operating systems and HL7-parsing routines on a daily interval. With recent changes emerging in VA’s information technology infrastructure, a new data architecture supporting ESSENCE’s surveillance capabilities is becoming possible.

Objective

To describe the new data warehouse, HAIISS Data Warehouse (HDW) architecture whereby VA’s Electronic Surveillance System for Early Notification of Community-based Epidemics (ESSENCE) will receive its required data elements from VA’s 128 VistA systems in a more accurate, robust and time sensitive manner.

During the 2009 H1N1 influenza pandemic, the Washington State Department of Health (DOH) temporarily made lab-confirmed influenza hospitalizations and deaths reportable. As reporting influenza hospitalizations is resource intensive for hospitals, electronic sources of inpatient influenza surveillance data are being explored. A large Health Information Exchange (WA-HIE) currently sends DOH the following data elements on patients admitted to 14 hospitals throughout eastern Washington: hospital, admission date, age, gender, patient zip code, chief complaint, final diagnoses, discharge disposition, and unique identifiers. WA-HIE inpatient data may be valuable for monitoring influenza activity, influenza morbidity, and the basic epidemiology of hospitalized influenza cases in Washington.

Objective

To evaluate the timeliness, completeness, and representativeness of influenza hospitalization data from an inpatient health information exchange.

Management policies for influenza outbreaks balance the expected morbidity and mortality costs versus the cost of intervention policies under outbreak parameter uncertainty. Previous approaches have not updated parameter estimates as data arrives or have had a limited set of possible intervention policies. We present a methodology for dynamic determination of optimal policies in a stochastic compartmental model with sequentially updated parameter uncertainty that searches the full set of sequential control strategies.

Objective

This abstract highlights a methodology to build optimal management policy maps for use in influenza outbreaks in small populations.