P0518 - Replacing emergency department head computed tomography with “rapid” magnetic resonance imaging for children with suspected ventricular shunt malfunction

Background: Children with ventricular shunts undergo frequent neuroimaging in the emergency department (ED) to evaluate for possible shunt malfunctions. Historically, the primary imaging modality has been computed tomography (CT), which exposes patients to ionizing radiation and an increase in the lifetime risk of malignancy. “Rapid” magnetic resonance imaging (rMRI) is an accurate and feasible alternative to CT for shunt evaluations. Therefore, we sought to reduce use of CT and increase use of rMRI as the primary neuroimaging study for the evaluation of suspected shunt malfunction. Specifically, we aimed to reduce the proportion of head CTs among patients warranting neuroimaging for the evaluation of possible shunt malfunction in the ED by 40% within 12 months.

Methods: This was a single center quality improvement project in a tertiary care pediatric ED. Interventions included 1) multidisciplinary team development of a standardized guideline for shunt evaluation in the ED which promoted use of rMRI over CT, 2) educational outreach across relevant hospital divisions, and 3) creation of a standardized electronic health record orderset to compliment the guideline. We evaluated patients <18 years for 12 months pre- and 24 months postimplementation of a clinical effectiveness guideline for the imaging evaluation of patients with possible shunt malfunction. Our primary outcome measure was the rate of CT. Balancing measures included: time to neuroimaging, ED length of stay, total neuroimaging performed within 24 hours for admitted patients, time to operative intervention, follow-up neuroimaging within 7 days for discharged patients, and 3-day ED revisits. We present rates adjusting for repeat visits by the same patient.

Results: There were 283 encounters by 178 patients pre- and 471 encounters by 331 patients post- implementation of the guideline, and the patient populations were similar between the time periods. 74.4% of patients received any neuroimaging pre- and 79.5% post-guideline implementation (p=0.09). The CT rate decreased from 84.1% to 39.5% and the rMRI rate increased from 15.9% to 60.5% (p<0.001) pre- and post-guideline, respectively. Analyzing the CT utilization reduction with a statistical process control p-chart revealed a significant and sustained reduction over a 16-month time period.(Figure) We observed increases in the mean time to neuroimaging (+52 min; [95%CI:40,64], p<0.001), time to operative intervention (+107 min; [95%CI:29.4, 185, p=0.01]), and ED length of stay (+42 min; [95%CI:25, 59], p<0.001), without changes in follow-up neuroimaging (p=0.8), 3-day revisits (p=0.29) or follow-up imaging rates (p=0.35).

Conclusion: Multidisciplinary implementation of a standardized guideline, with complimentary educational outreach and supportive orderset availability, reduced use of CT and increased use of rMRI in an ED setting. Reduction in radiation exposure with widespread ED rMRI use for patients with ventricular shunts should be balanced against the increased time added for obtaining imaging, in LOS, and for operative intervention.