Purpose: This study assessed the association between homicide rates in northern Mexico and potentially avoidable use of emergency departments (ED) in the US-Mexico border region. The border region is largely rural and underserved, making the identification and correction of potential barriers to access crucial.

Methods: We used secondary data from state inpatient and ED discharge databases for California and Arizona for 2005-2010. A retrospective observational analysis using generalized linear models was used to determine whether the probability that an ED encounter was potentially avoidable was associated with homicide rates in the nearest Mexican municipality. To conduct the analysis, the location of ED encounters were identified and matched with homicide rates in the nearest Mexican municipality and regional characteristics. The probability that an ED encounter was potentially avoidable was calculated using the Billings ED algorithm.

Findings: We found that 77% of ED encounters were potentially avoidable, with a higher percentage in border counties. There was no statistically significant relationship between homicide rates and the probability that an ED encounter was for a potentially avoidable condition for the full analytic sample (n = 24,859,273) and the uninsured and underinsured in the sample (n = 11,700,123).

Conclusion: A substantial majority of ED encounters in the US-Mexico border region were potentially avoidable. However, there was not a strong relationship between homicide rates in northern Mexico and the distribution of ED discharges in Arizona and California. Given the large percentage of potentially avoidable ED encounters and the ongoing violence in Mexico, continuing to monitor this relationship is important.