Demographic factors influencing nonurgent emergency department utilization among a Medicaid population.

Abstract

To use administrative medical encounter data to examine nonurgent emergency department (ED) utilization as it relates to member characteristics (i.e., age, gender, race/ethnicity, urbanicity and federal poverty level (FPL)). This 1year cross-sectional study used medical claims from a managed care organization for Medicaid members enrolled from October 1, 2010 - September 30, 2011. ED encounters occurring during the study period were classified as either urgent or nonurgent using ICD-9 diagnosis codes obtained from medical claims. Examples of urgent diagnoses include head traumas, burns, allergic reactions, poisonings, preterm labor or maternal/fetal distress. A total of 187,263 members aged 2 to 65years were retained for study. A zero-inflated Poisson regression model examined the influence of member-level characteristics on nonurgent ED utilization, while simultaneously adjusting for all factors. Females were 41% more likely to have a nonurgent ED visit (p≤0.0001). Members ages 50-65 were least likely to have a nonurgent ED visit (p≤0.0001). White members had higher odds of having at least one nonurgent ED visit (p≤0.0002). Rural members were 7.7% less likely to have a nonurgent ED visit. Members in the 400%+FPL category were less likely to seek nonurgent care from an ED (p≤0.0001). A nonurgent ED visit occurs when care is sought at an ED that could have been handled in a primary care setting. Approximately 30-50% of all ED visits in the United States are considered nonurgent. This study supports the need to determine factors associated with misuse of ED services for nonurgent care. Demographic factors significantly impacting nonurgent ED utilization include gender, age, race/ethnicity, urbanicity and percent of the FPL. Results may be useful in ED utilization management efforts.