Improving Poisoning Surveillance in North Carolina Using Emergency Department Data: A Comparison of Three Poisoning Case Definitions

BACKGROUND:  The North Carolina mortality rate for poisoning nearly tripled between 1999 and 2010.  This increase is largely due to an increase of opioid analgesic fatalities.  In 2011, the number of unintentional poisonings (1,140) was second only to motor vehicle collisions (1,210) as the leading cause of injury death. Not all poisonings result in death, however. Far greater numbers of poisonings are nonfatal and result in hospitalization, emergency department or outpatient clinic visits, or resolve without medical attention.  Assessing morbidity using hospital discharge and emergency department visit data, provides a better representation of the burden of poisonings in NC.  The quality of surveillance data depends on the definition used to identify cases of injury.  NC has used the definition of poisoning suggested by CDC’s Injury Indicator Report based on External Cause of Injury codes (E-codes); it is unclear, however, if this definition accurately captures the true number of emergency department visits due to poisoning in NC.  For comparative purposes, we applied the poisoning definition suggested by the Safe States Injury Surveillance Workgroup-7 (ISW-7), as well as a drug overdose definition developed by injury prevention partners in NC.

METHODS:  We examined all 2011 emergency department visits captured through the NC Disease Event Tracking and Epidemiologic Collection Tool (NC DETECT) which met the coding criteria of a poisoning based on at least one of three definitions.  Summary statistics were generated for emergency department visits that received an E-code; visits were then classified by injury intent and mechanism.

RESULTS:  Using the standard CDC definition based on the first-listed E-code, there were 17,576 poisonings of which 8,430 were unintentional drug overdoses.   When this definition was expanded to include poisoning diagnosis codes in addition to E-codes, there were 35,193 poisonings, of which 13,720 were drug-related. The definition suggested by injury prevention partners produced 14,566 emergency department visits based on the first-listed diagnosis code or E-code.  Employing the code created by ISW-7, the number of total poisonings increased to 178,706, of which 87,666 emergency department visits were drug-related. Much of the difference observed in the number of poisonings was due to the inclusion of substance abuse codes in the ISW-7 definition.

CONCLUSIONS:  The number of cases identified through surveillance is highly dependent on the case definition used.  It is important to determine a priori as an investigator whether the case definition’s sensitivity or specificity is of greater importance.