Timely Local Surveillance of Opioid-Related Deaths Using Medical Examiner Data - Multnomah County, Oregon

BACKGROUND:  Over the past two decades, opioid-related deaths (ORDs) have become an increasingly prominent public health issue in the United States. Local public health agencies continue to struggle with response and community-based interventions, partially due to a lack of reliable and up-to-date surveillance data for ORDs. National and state agencies can rely on delayed information to produce broader descriptive reports, but for effective and immediate response, local agencies require more timely surveillance tools like those used for syndromic surveillance of infectious disease. Multnomah County Health Department (MCHD) in Portland, Oregon, has historically relied on analysts to manually sort through ORD Medical Examiner (ME) cases in Excel, to produce ad hoc reports and respond to media requests. However, the acceptability of the system is low among MCHD staff, prompting an evaluation of the system and development of a more effective surveillance technique. 

METHODS:  An evaluation of this surveillance system was conducted using the 2001 CDC Updated Guidelines for Evaluating Public Health Surveillance. With no gold standard for comparison of system sensitivity, an automated text-search method was developed in SAS 9.3 and applied to the ME data, allowing for comparison with the historic method. 

RESULTS:  The evaluation indicates the ORD surveillance system at MCHD is sensitive compared to the SAS method, but it is also time-consuming, complicated, and poorly accepted by system users. Inconsistent coding practices for “cause of death” variables in ME records further diminish the quality of the data for epidemiologic use and complicate case definitions. In comparison, the automated system is timelier, more sensitive, has improved simplicity, and uses text-search methods to more efficiently abstract ORDs from ME data. 

CONCLUSIONS:  Timely surveillance of ORDs should be used to inform public health response and evaluate effectiveness of interventions, policies, and programs. Informatics and data sharing play an increasingly important role in surveillance methodology. Collaboration between local health departments, police, medical examiners, and emergency medical response agencies can yield more comprehensive surveillance tools. For immediate use at the local level, opiate-death surveillance can utilize these rapidly updated data sources through efficient abstraction, matching, and analysis techniques. Updated surveillance systems rooted in ME data should employ case definitions and measures that can capture true ORDs even when “cause of death” codes may not be entirely conclusive.