Enhancing Diabetes Mellitus Surveillance with Emergency Department Data in the District of Columbia

BACKGROUND: In the District of Columbia (DC), diabetes mellitus (DM) prevalence is 8 percent and accounts for 40,000 Emergency Department (ED) visits per year, approximately. Monitoring hyperglycemia in ED patients presents opportunities to identify uncontrolled DM, previously undiagnosed DM, and high risk for DM. The DC Electronic Surveillance System for the Early Notification of Community-based Epidemics (ESSENCE), a web-based syndromic surveillance system, collects ED chief complaints and limited discharge diagnosis data to detect unusual patterns of illness based on pre-diagnostic data. Its applications for monitoring non-communicable diseases are not well studied. The objectives of this study are to 1) determine common ED chief complaints of DM-related hyperglycemic patients by linking diagnostic hospital data with ESSENCE data; and 2) evaluate the usefulness of ESSENCE for DM surveillance.  

METHODS: Data will be collected from ESSENCE, reported by DC acute care hospitals during calendar year 2012. The following variables will be included for ED patients aged ≥20 years: age, sex, zip code, medical record number (MRN), hospital location, visit date/time, chief complaint, and discharge diagnosis. One DC hospital will provide ED diagnostic data for measures of blood glucose (BG), Hemoglobin A1c, and blood pressure. We will link ESSENCE and diagnostic data using MRN, and assess hyperglycemic (BG ≥200mg/dl) patient chief complaints by age, sex, and DM status. We will determine common chief complaints of DM-related hyperglycemic patients and design a query for ESSENCE. The query will be performed with all District hospital data. We will evaluate the validity and usefulness of tracking diabetes with ESSENCE using percent agreement between chief complaints and discharge diagnoses from ESSENCE. Descriptive statistical analyses will be performed using SAS v9.3.  

RESULTS: We expect a large proportion of 2012 ESSENCE data to be linked with diagnostic data. Based on previous reports, we expect hyperglycemia will be found in 6 per 10,000 adult ED visits and one-third of hyperglycemic patients will not have a previous DM diagnosis. We expect this study will identify more accurate keywords that determine DM diagnoses, which can be used to query ESSENCE chief complaints and monitor DM in DC. 

CONCLUSIONS: Using diagnostic data for DM-related hyperglycemic patients will increase our understanding of precursory chief complaints for related ED visits. These methods will inform DM program planning and health systems interventions to improve the identification of DM and prevent acute DM complications that result in ED visits.