Queries

Time series analysis is very common in syndromic surveillance. Large scale biosurveillance systems typically perform thousands of time series queries per day: for example, monitoring of nationwide over-thecounter (OTC) sales data may require separate time series analyses on tens of thousands of zip codes. More complex query types (e.g. queries over various combinations of patient age, gender, and other characteristics, or spatial scans performed over all potential disease clusters) may require millions of distinct queries. Commercial OLAP databases provide data cubes to handle such ad hoc queries, but these methods typically suffer from long build times (typically hours), huge memory requirements (requiring the purchase of high-end database servers), and high maintenance costs. Additionally, data cubes typically require 1 second or more to respond to each complex query. This delay is an inconvenience to users who want to perform multiple queries in an online fashion; additionally, data cubes are far too slow for statistical analyses requiring millions of complex queries, which would require days of processing time.

Objective

We present T-Cube, a new tool for very fast retrieval and analysis of time series data. Using a novel method of data caching, T-Cube performs time series queries approximately 1,000 times faster than standard state-of-the-art data cube technologies. This speedup has two main benefits: it enables fast anomaly detection by simultaneous statistical analysis of many thousands of time series, and it allows public health users to perform many complex, ad hoc time series queries on the fly without inconvenient delays.

The Indiana Public Health Emergency Surveillance System (PHESS) currently receives approximately 5,000 near real-time chief complaint messages from 55 hospital emergency departments daily.  The ISDH partners with the Regenstrief Institute to process, batch, and transmit data every three hours.  The Electronic Surveillance System for the Early Notification of Community-based Epidemics (ESSENCE) tool is utilized to analyze these chief complaint data and visualize generated alerts.1   

 

The ISDH syndromic surveillance team discovered that certain chief complaints of interest were coded into the “other” syndrome and not visible in typical daily alert data.  Staff determined that even a single chief complaint containing keywords related to specific reportable diseases could be of significant public health value and should be made available to investigating epidemiologists.2 

 

In addition, data quality is critical to the success of the program and must be evaluated to ensure optimal system performance.  Metrics related to data flow and completeness were identified to serve as indicators of hospital connectivity or coding problems.  These measures included the percent change in daily admits and the proportion of chief complaints missing the patient address.

Objective

This paper describes the development of targeted query tools and processes designed to maximize the extraction of information from, and improve the quality of, the hospital emergency department chief complaint data stream utilized by the Indiana State Department of Health (ISDH) for syndromic surveillance.

Automated disease surveillance systems that analyze data by syndrome categories have been used to look for outbreaks of disease for about 10 years. Most of these systems notify users of increases in the prevalence of reports in syndrome categories and allow users to view patient level data related to the increase. For most situations this level of investigation is sufficient, but occasionally a more dynamic level of control is required to properly understand an emerging illness in a community. During the SARS outbreak, for example, the respiratory syndrome was defined too broadly to allow users to track SARS. However, some systems, allowed users to build dynamic queries that allowed them to search their data by using the SARS case definition [1]. Users could perform free-text queries that identified records containing specific keywords in the chief complaint or specific combinations of ICD9 codes. This advanced querying capability has proven to be one of the most used features used by monitors of disease surveillance systems. Objective: The objective of this project is to build a new, more flexible query interface that allows users to define and build their query as if they were writing a logical expression for a mathematical computation. The interface is designed so that it can be easily adapted to fit into nearly any syndromic surveillance system.The interface will be evaluated in future versions of the ESSENCE and BioSense Systems.

Oregon’s statewide syndromic surveillance system (Oregon ESSENCE) has been operational since 2012. Non-federal emergency department data (and several of their associated urgent care centers) are the primary source for the system, although other data sources have been added, including de-identified call data from OPC in 2016. OPHD epidemiologists have experience monitoring mass gatherings and have a strong relationship with OPC, collaborating on a regular basis for routine and heightened public health surveillance. Nevertheless, surveillance for the Great American Solar Eclipse (August 2017) presented a challenge due to the 107 reported simultaneous statewide eclipse-watching events planned for the day of the eclipse (some with estimated attendance of greater than 30,000 people and most in rural or frontier regions of the state). Scientific literature is limited on mass gathering surveillance in the developed world, particularly in rural settings, so OPC and OPHD worked together to develop a list of health conditions of interest, including some that would warrant both an ED visit and a call to OPC (e.g., snake bites). Monitoring visits in both data sources in would allow for assessment of total burden on the healthcare system, especially in the case of snake bites where only specific bites require administration of anti-venom.

Objective:

Identify surveillance priorities for emergency department (ED) and Oregon Poison Center (OPC) data ahead of the 2017 Great American Solar Eclipse gatherings in Oregon and create a suite of queries for use in the Health Intelligence Section of the Oregon Public Health Division (OPHD) Incident Management Team (IMT).