Evaluation of the District of Columbia's National Electronic Disease Surveillance System

BACKGROUND:  The National Electronic Disease Surveillance System (NEDSS) Base System (NBS) software was developed by the Centers for Disease Control and Prevention (CDC) in 2001 to afford health departments the capability to interface, monitor disease rates, and electronically report findings using NEDSS guidelines. The District of Columba (DC) implemented the NBS in 2011, replacing the Washington DC Automated Disease Surveillance System (WADSS). Since the implementation of the software to conduct surveillance activities there has not been a complete evaluation of its use. The goal of this analysis was to evaluate the efficiency and effectiveness of current DC surveillance activities utilizing the NBS and highlight areas for future improvement. 

METHODS: To evaluate NBS use in DC, one food and one vector-borne disease (Salmonellosis and Lyme disease, respectively) with different mandated reporting times were selected. During 2012 and 2013 in DC, Salmonellosis was the second most frequently reported foodborne illness, and Lyme disease was the most frequently reported vector-borne disease. The evaluation was completed using CDC’s 2001 “Updated Guidelines for Evaluating Public Health Surveillance Systems”, paying special attention to data quality, timeliness, and positive predictive value (PPV). Final analyses were completed with Microsoft Excel (2010) and SAS (Version 9.3).

RESULTS: Among DC residents, 168 reports of Salmonellosis and 886 reports of Lyme disease were submitted in 2012 and 2013. State of residence was incorrectly recorded for 8.33% of Salmonellosis and 1.69% of Lyme disease reports electronically filed in the NBS as DC residents. Two (1.19%) Salmonellosis reports omitted gender and four (2.38%) omitted age. Over 14% of all confirmed and probable Salmonellosis cases were reported in the required 24 hours and 78.75% were reported to CDC. PPV for Salmonellosis case definition accuracy was 76.19%. All Lyme disease reports recorded gender and three (0.33%) omitted age. Over 25% of confirmed and probable cases were reported within the required 48 hours. All required Lyme disease cases were uploaded to CDC. PPV for Lyme disease case definition accuracy was 94.24%.

CONCLUSIONS: Overall data quality and PPV for Salmonellosis was lower than Lyme disease despite the large difference in total submissions. Additionally, the majority of reports for each illness were submitted after the required timeframe in DC. Findings from this analysis will help guide future protocols for disease investigations conducted in DC and can also be used to encourage timelier reporting among healthcare providers.