We sought to determine whether physicians' estimates of critically ill patients' hemodynamics are accurate as compared with esophageal Doppler probe (EDP) measurements and whether provision of measured hemodynamic profiles produces treatment changes and changes in physician confidence. At an urban county ED, we enrolled 55 critically ill adults with pulmonary edema of unclear etiology, sustained hypotension, or lactic acidosis. Physicians estimated cardiac output (CO) and systemic vascular resistance (SVR), categorized shock, stated treatment plans, and rated confidence in assessment and treatment plans. Physicians were informed of EDP hemodynamic measurements (CO, SVR, stroke volume, and contractility), and they repeated their assessments and plans. Weighted kappa values between physician estimates and EDP measurements of CO and SVR were 0.57 (95% confidence interval [CI] = 0.77-0.36) and 0.40 (95% CI = 0.64-0.16), respectively. Shock characterization changed in 52%, confidence increased significantly (pre-EDP mean = 3.3 +/- 0.9; post-EDP mean = 4.0 +/- 0.6; P = .0001), and stated treatment plans changed in 68% of the patients. Chart review demonstrated that similar proportions of control subjects had treatment changes, with a mean difference of 20% (95% CI = -2 to 42). Physician assessments of hemodynamic variables were moderately accurate. We conclude that EDP hemodynamic profiles change assessments and increase confidence in assessments but may not alter treatment.