2 Effectiveness of Provider-Focused Interventions to Eliminate Care Disparities: Results from the Equity in Diagnostic Imaging Trial (EDIT)

Abstract

African Americans have almost twice the incidence of subarachnoid hemorrhage (SAH) compared to Whites, yet studies indicate they are less likely to receive neuroimaging when presenting to the emergency department (ED) with headache, even when adjusting for age, sex, co-morbidity and acuity. Regulatory efforts and quality initiatives to reduce CT imaging for ED patients with headache may exacerbate this disparity and increase the likelihood that Black patients, who are at higher risk for SAH, do not receive an appropriate ED workup for severe headache. The Equity in Diagnostic Imaging Trial (EDIT) was designed to develop and test the effect of two provider-focused interventions on this racial disparity in diagnostic imaging rates for ED patients with headache. The EDIT team designed two evidence-based, provider-focused interventions: 1) an audit-feedback (AU-FB) intervention that provided periodic information on the higher incidence of SAH among Blacks and their lower rates of neuroimaging; and 2) clinical decision support (CDS) embedded in the electronic health record with similar content. These interventions were tested in sequential blinded, randomized, controlled trials; providers were unaware of the study. Attending physicians, emergency medicine residents, and physician assistants working in the adult ED of a large urban academic medical center were randomized to either AU-FB or control. After a 12-month intervention period and a 9-month washout period, providers were re-randomized to either CDS or control for another 12-month period. All adult patient visits (age ≥ 18), identified through an automated report of patients for whom an adult headache template was used by the ED clinician during the study period, were reviewed. Patients with history of trauma were excluded. Data elements, abstracted by automated reports and structured manual review by trained research staff, included: ED provider names; patient age, sex, race, ethnicity, insurance type, emergency severity index, chief complaint, vital signs, pain score, neuro imaging order, ED disposition, ED diagnosis, and elements of the prior medical history, history of present illness, and physical examination significantly associated with the decision to order diagnostic imaging in a previously reported observational study, which confirmed the presence of a Black-White imaging disparity at the study site. Comorbidities were categorized using the Elixhauser Comorbidity Index. 8,185 ED visits were reviewed; 130 (1.6%) of these were ED revisits within 2 days of a previous ED visit for headache. These were dropped resulting in a final sample of 8,055 visits. The primary outcome for both trials is likelihood of a neuroimaging order for Black vs White patients. Secondary analyses will examine neuroimaging for all racial/ethnic/sex patient groups; comparative effectiveness of the two interventions; provider subgroups (attending, resident, PA); and concordance vs discordance of provider-patient race/ethnicity. The EDIT Trial tested the effectiveness of two provider-focused interventions addressing a racial disparity in diagnostic imaging rates for ED patients with headache. These interventions are particularly timely given current efforts to decrease CT utilization for ED patients with headache.