1046 — Electronic Health Record-Based Interventions for Improving Appropriate Diagnostic Imaging: A Systematic Review and Meta-Analysis

Goldzweig CL, West Los Angeles Veterans Affairs Medical Center, Los Angeles, California; Orshansky G, West Los Angeles Veterans Affairs Medical Center, Los Angeles, California; Paige NM, West Los Angeles Veterans Affairs Medical Center, Los Angeles, California; Miake-Lye IM, West Los Angeles Veterans Affairs Medical Center, and 2University of California, Los Angeles Fielding School of Public Health, Los Angeles, CA; Beroes JM, West Los Angeles Veterans Affairs Medical Center, Los Angeles, California; Ewing BA, RAND Corporation, Southern California Evidence-based Practice Center, Santa Monica, California; Shekelle PG, West Los Angeles Veterans Affairs Medical Center, Los Angeles, California, 3RAND Corporation, Southern California Evidence-based Practice Center, Sant;

Objectives:
One driver of increasing health care costs is the use of radiologic imaging procedures. More appropriate use could improve quality and reduce costs. We reviewed interventions that use the computerized clinical decision support (CCDS) capabilities of electronic health records to improve appropriate use of diagnostic radiologic test ordering.

Methods:
This report was developed based on a nomination from the VHA Choosing Wisely Workgroup, including members from Health Services Research and Development, Diagnostic Services, and the Office of Patient Care Services. English-language articles were identified in PubMed from 1995 to September 2014 and from searches in Web of Science and PubMed of citations related to key articles. We included studies that assessed the effect of CCDS on diagnostic radiologic test ordering in adults. Two independent reviewers extracted data on functionality, study outcomes, and context and assessed the quality of included studies.

Results:
Of the 23 articles included, ten interventions targeted "high cost imaging," which usually included CT and MRI and occasionally nuclear medicine tests, four interventions targeted pulmonary CT angiography, two studies targeted chest x-ray, four interventions targeted multiple different types of radiologic studies, and three studies had other radiologic targets. There were 3 randomized trials, 7 time-series studies, and 13 pre-post studies. Thirteen studies provided moderate-level evidence that CCDS improves appropriateness (effect size, ?0.49 [95% CI, ?0.71 to ?0.26]) and reduces use (effect size, ?0.13 [CI, ?0.23 to ?0.04]). Interventions with a "hard stop" that prevent a clinician from overriding the CCCDS without outside consultation, as well as interventions in integrated care delivery systems, may be more effective. Harms have rarely been assessed but include decreased ordering of appropriate tests and physician dissatisfaction.

Implications:
Limitations include potential for publication bias, insufficient reporting on harms, and poor description of context and implementation. Computerized clinical decision support integrated with the electronic health record can improve appropriate use of diagnostic radiology by a moderate amount and decrease use by a small amount. Before widespread adoption can be recommended, more data are needed on potential harms.

Impacts:
These results will help the Workgroup in their effort to implement some of the segments of the "Choosing Wisely" campaign.