Improving the breast MRI process: A longitudinal QI project at the Houston VA.

Abstract

371 Background: The number of female veterans is increasing by more than 18,000 per year, and 700 veterans enrolled in VA healthcare are diagnosed with breast cancer each year. The American Cancer Society recommends screening MRI in women with 20-25% or greater lifetime risk of breast cancer, and primary care physicians and oncologists order breast MRIs regularly. The Michael E. DeBakey Veteran’s Affairs Medical Center (MEDVAMC) performs mammograms and ultrasounds on site, but breast MRI is not currently available. Care in the Community arranges for veterans to obtain these studies at outside facilities. This is a complex process, requiring coordination among multiple departments at the MEDVAMC, outside MRI facilities, and patients. Methods: We created a comprehensive process map for the current breast MRI ordering process to determine areas for improvement. We reviewed all breast MRI orders from 1/2019 to 5/2021 at the MEDVAMC and determined the cancellation rate and the rate of benign (BIRADS 1-2) vs non benign (BIRADS 0, 3-5) screening outcomes. We created a Pareto analysis of cancellation reasons. Our process measures were percentage of placed breast MRI orders that were cancelled and delay from MRI order to MRI performed. Our outcome measures were percentage of order cancellations due to incorrect or incomplete orders and delay from MRI order placement to results uploaded. Results: Of the 434 orders that were placed for breast MRIs, 64% were cancelled. Only 117 of the 167 patients that had MRIs ordered ultimately received one. Of the 104 patients without known malignancy who competed an MRI, 45% had non-benign findings requiring follow-up. Our pareto analysis showed that the top cancellation reason was incorrect orders, usually contrast or side verbiage. In June 2021, the CITC order set for MRI was changed from a free text box to discrete, clickable options for side and contrast. In December 2021, we held educational interventions on how to avoid delay or cancellation. Our p-chart of percentage of cancellations due to incorrect ordering shows an extremely promising decrease, with the last seven consecutive subgroups of 15 patients below the mean. The order cancellation percentage is also decreasing. The average delay pre-intervention from MRI order placement to results uploaded was 67 days. In January 2022, we implemented a breast MRI nurse navigator to help acquire results and expedite scheduling. Post-intervention, the average delay has decreased to 38.5 days. Using XmR and S charts, we have also shown an increase in precision. Conclusions: Providing discrete, clickable options within the MRI order set reduced the number of cancellations due to incorrect verbiage. Early data suggests that the educational intervention improved the cancellation rate. Instituting a nurse navigator decreased both the delay and variability for the time from order placement to results received, which will reduce the delay to follow-up of non-benign findings.