How emergency physicians use biomarkers: insights from a qualitative assessment of script concordance tests.

OBJECTIVES: Biomarkers have been developed in emergency medicine to improve decision at bedside using Bayesian approach. We intend to determine the cognitive process actually utilised by emergency physicians to incorporate biomarkers in clinical reasoning. DESIGN: We invited eight emergency physicians to answer eight script concordance tests. Interviews were tape-recorded and qualitatively analysed using predetermined categories until saturation. RESULTS: Emergency physicians mainly mobilised intuition and non-Bayesian reasoning to incorporate biomarkers for diagnosis or treatment strategies. CONCLUSIONS: Although biomarkers have been developed to be used in a Bayesian approach, emergency physicians mainly use other analytical and non-analytical cognitive processes to introduce these tools in their clinical reasoning.

Midregional pro-atrial natriuretic peptide for the diagnosis of cardiac-related dyspnea according to renal function in the emergency department: a comparison with B-type natriuretic peptide (BNP) and N-terminal proBNP.

BACKGROUND: Although renal dysfunction influences the threshold values of B-type natriuretic peptide (BNP) and N-terminal proBNP (NT-proBNP) in diagnosis of cardiac-related dyspnea (CRD), its effects on midregional pro-atrial natriuretic peptide (MR-proANP) threshold values are unknown. We evaluated the impact of renal function on MR-proANP concentrations and compared our results to those of BNP and NT-proBNP. METHODS: MR-proANP, BNP, and NT-proBNP concentrations were measured in blood samples collected routinely from dyspneic patients admitted to the emergency department. Patients were subdivided into tertiles based on their estimated glomerular filtration rate [eGFR, in mL · min(-1) · (1.73 m(2))(-1)]: tertiles 1 (<44.3), 2 (44.3-58.5), and 3 (≥58.6). RESULTS: Of 378 patients studied, 69% (n = 260) had impaired renal function [<60 mL · min(-1) · (1.73 m(2))(-1)] and 30% (n = 114) had CRD. MR-proANP, BNP, and NT-proBNP concentrations were significantly increased in patients with impaired renal function. In each tertile, all peptides remained significantly increased in CRD patients by comparison with non-CRD patients. By ROC analysis, MR-proANP, BNP, and NT-proBNP threshold values for the diagnosis of CRD increased as eGFR decreased from tertile 3 to tertile 1. Areas under the ROC curve for all peptides were significantly lower in tertile 1. Using adapted thresholds, MR-proANP, BNP, and NT-proBNP remained independently predictive of CRD, even in tertile 1 patients. CONCLUSIONS: Renal function influences optimum cutoff points of MR-proANP for the diagnosis of CRD. With use of an optimum threshold value adapted to the eGFR category, MR-proANP remains as effective as BNP and NT-proBNP in independently predicting a diagnosis of CRD in the emergency department.