Does Health Information Exchange Reduce Redundant Imaging? Evidence from Emergency Departments

Research Objective: When patients are treated in multiple delivery settings, the inability to access their records at the point of care impedes coordination and introduces inefficiencies.  As electronic health record (EHR) adoption increases, the hope is that these problems can be remedied by achieving greater interoperability among these systems. Broad-based electronic health information exchange (HIE), an important extension of EHRs in which key clinical data flow among providers, is projected to result in large quality gains and cost savings. Provider incentives to engage in repeat testing are mixed. Fee for service payment promotes redundant services, but concern about harmful effects of unnecessary testing on patients and societal costs may motivate providers to avoid redundancy when HIE provides relevant prior test results. In this study we evaluate whether HIE decreases repeat imaging in emergency departments (EDs).

Study Design/Methods/Data:  We used changes over time in HIE participation by EDs in California and Florida during 2006 to 2009 to identify its effect on repeat imaging. Using ordinary least square regression with ED fixed effects and trends, we estimated the impact of HIE participation on the probability of repeat imaging among patients visiting multiple unaffiliated EDs during thirty day periods.  We examined repeat imaging for three procedures (computed tomography [CT], ultrasound, and chest x-ray). Our principle sources of data included ED discharge records from the State Emergency Department Databases for California and Florida from 2007-2010 merged with HIMSS Analytics survey data reporting hospital participation in HIE. We compared 37 hospital-based EDs that initiated HIE participation to 410 EDs that did not participate in HIE during the study period.

To help interpret the practical impact of our findings, we used our results to predict the annual reduction in repeat imaging and associated costs that would be avoided if all EDs in the United States participated in HIE. Cost savings were calculated by weighting the 2012 Medicare reimbursement rate of each reported CPT codes by its relative proportion among all patients who received an image in the given modality (CT, ultrasound, or chest x-ray) and multiplying this average cost per image by the predicted annual images avoided with full HIE uptake.

 Principal Findings:  HIE reduced the probability of repeat imaging by -8.7 percentage points for CT (p<0.01), -9.1 percentage points for ultrasound (p<0.05), and -13.0 percentage points for chest x-ray (p<0.01), reflecting relative reductions of 44% to 67% from the sample means.  Extrapolated to the nation, this could represent $19 million in avoided annual costs.

 Conclusions: Our study is among the first empirical tests of the impact of HIE on redundant imaging, and we find that HIE reduced repeat imaging in ED settings during this critical early stage of its development. Furthermore, our findings demonstrate that ED providers respond to the availability of prior test results by avoiding repeat testing. These findings suggest that provider concerns for patient and social welfare overcome fee for service incentives that can motivate redundant service provision when relevant clinical information from previous episodes is made readily available.