Effects of fully accessible magnetic resonance imaging in the emergency department

Background: The Joint Commission Comprehensive Stroke Center certification requires that magnetic resonance imaging (MRI) be available on site, 24 hours a day, 7 days a week for evaluation of stroke in emergency department (ED) patients. Increased access to advanced diagnostic imaging has been shown to increase utilization, ED length of stay (LOS), and health care costs. EDs nationwide face decisions to pursue certification and increase MRI access. Understanding changes in utilization and the downstream effects may inform these decisions.

Objectives: The objective was to determine changes in emergency MRI utilization following placement of a 24/7 accessible MRI in the ED and its effects on resource utilization for rule-out stroke and neurology consult patients.

Methods: This was a retrospective cohort study comparing MRI use during the 32 months before and 26 months after MRI acquisition period in the ED of a Level I trauma and stroke center. An interrupted time-series design was used to account for changes in clinical practice patterns following MRI acquisition. Time-series plots and segmented regression analyses are presented to compare utilization patterns pre- and post-MRI and to understand potential confounding due to secular trends. Statistical hypothesis testing was used to determine differences in utilization, demographics, and clinical characteristics for cohorts pre- and post-MRI.

Results: MRI utilization in the ED increased 38.4% for rule-out stroke and 51.4% for neurology consult patients after MRI acquisition. The proportion of rule-out stroke patients receiving MRI increased from 32.5% pre-MRI to 45.0% post-MRI (p < 0.001). The proportion of neurology consult patients increased from 32.6% pre-MRI to 49.4% post-MRI (p < 0.001). Considering baseline increases in MRI utilization rates for both cohorts over time, segmented regression models detected more substantial and significant changes in utilization after MRI acquisition for the larger neurology cohort (p < 0.001) compared to the rule-out stroke cohort (p = 0.095). However, hospital admission rates declined 16.7% for rule-out stroke patients (68.2% pre, 56.8% post; p < 0.001) and remained constant for neurology patients (56.5% pre, 57.5% post; p = 0.414). Patients who obtained MRI in the ED had increased ED LOS, but decreased hospital LOS (admitted patients), compared to those with no MRI for pre and post cohorts.

Conclusions: Emergency MRI utilization increased substantially after placement of a fully accessible MRI in the ED. Patients receiving emergency MRI had increased ED LOS, decreased admission rates for some patients (rule-out stroke), and reduced hospital LOS for those admitted. Potential changes in ED patient resource utilization should be considered when determining whether to acquire an MRI for Comprehensive Stroke Center certification.