Agreement Between Serum Assays Performed in ED Point-of-Care and Hospital Central Laboratories

• Published in: The western journal of emergency medicine 2017-04, Vol.18 (3), p.403-409
• Main Authors: Dashevsky, Meir, Bernstein, Steven L, Barsky, Carol L, Taylor, Richard A
• Format: Article
• Language: English
• Subjects: Adult; Biological Assay; Biomarkers - blood; Blood Chemical Analysis - instrumentation; Blood Chemical Analysis - standards; Blood Urea Nitrogen; Cost-Benefit Analysis; Creatinine - blood; Emergency Department Operations; Emergency Medical Services - methods; Emergency Medical Services - standards; Female; Humans; Laboratories, Hospital - standards; Male; Middle Aged; Point-of-Care Systems - standards; Potassium - blood; Quality Assurance, Health Care; Reproducibility of Results; Retrospective Studies; Sodium - blood; United States
• ISSN: 1936-900X; 1936-9018
• DOI: 10.5811/westjem.2017.1.30532

Point-of-care (POC) testing allows for more time-sensitive diagnosis and treatment in the emergency department (ED) than sending blood samples to the hospital central laboratory (CL). However, many ED patients have blood sent to both, either out of clinical custom, or because clinicians do not trust the POC values. The objective of this study was to examine the level of agreement between POC and CL values in a large cohort of ED patients. In an urban, Level I ED that sees approximately 120,000 patients/year, all patients seen between March 1, 2013, and October 1, 2014, who had blood sent to POC and CL labs had levels of agreement measured between serum sodium, potassium, blood urea nitrogen (BUN), creatinine, and hematocrit. We extracted data from the hospital's clinical information system, and analyzed agreement with the use of Bland-Altman plots, defining both 95% confidence intervals (CIs) and more conservative CIs based on clinical judgment. Out of 163,661 patients seen during the study period, 14,567 had blood samples sent both for POC and CL analysis. Using clinical criteria, the levels of agreement for sodium were 98.6% (within 5mg/dL), for potassium 90.7% (0.5 mmol/L), for BUN 89.0% (within 5 mg/dL), for creatinine 94.5% (within 0.3 mg/dL), for hematocrit 96.5% (within 5 g/dL). Agreement between POC and CL values is excellent. Restricting the analysis to clinically important levels of agreement continues to show a high level of agreement. The data suggest that sending a serum sample to the hospital CL for duplicate assays is unnecessary. This may result in substantial savings and shorter ED lengths of stay.