Prospective validation and refinement of a decision rule to obtain CXR in patients with non-traumatic chest pain in the ED

Abstract

Objectives

To prospectively validate and refine previously published criteria to determine the potential utility of chest x-ray (CXR) in the evaluation and management of patients presenting to the emergency department (ED) with non-traumatic chest pain.

Methods

A prospective observational study was performed of patients presenting to three EDs in the US with a chief complaint of non-traumatic chest pain. Previously defined high-risk history and exam elements were combined into a refined decision rule and these elements were recorded for each patient by the ED physician. CXR results were reviewed and analyzed to determine the presence of clinically significant findings including pneumonia, pleural effusion, pneumothorax, congestive heart failure, or the presence of a new mass. Odds ratios for each history and exam element were analyzed as well as sensitivity, specificity, and negative predictive value of the rule overall.

Results

1111 patients were enrolled and 1089 CXRs were analyzed. There were 70 (6.4%) patients with clinically relevant findings on CXR. The refined decision rule had a sensitivity of 92.9% (CI 83.4%-97.3%) and specificity of 30.4% (CI 27.6%-33.4%) to predict clinically relevant findings on CXR, with a negative predictive value of 98.4% (CI 96.1%-99.4%). Five CXRs with clinically significant findings would have been missed by application of the refined rule (3 pneumonias and 2 pleural effusions). Applying these criteria as a CXR decision rule to this population would have reduced CXR utilization by 28.9%.

Conclusions

This study validates previous research suggesting a low clinical yield for CXR in the setting of non-traumatic chest pain in the ED. This refined clinical decision rule has a favorable sensitivity and negative predictive value in a patient population with low incidence of disease. Further validation is needed prior to use in practice.

This article is protected by copyright. All rights reserved.

http://ift.tt/2sq6AaK