Although pulmonary artery (PA) catheterization is frequently employed in the management of critically ill patients, there is little documentation that the information obtained alters patient management. This study evaluated prospectively this question in 103 PA catheterizations. Before catheterization, physicians were asked to predict the range of several hemodynamic variables, the presumed diagnosis, and their plan for therapy. After catheterization, each chart was reviewed. The hemodynamics at the time of catheterization and therapy within 8 h of catheterization were noted. Pulmonary artery occlusive (wedge) pressure (WP) was correctly predicted 30% of the time. Cardiac output, systemic vascular resistance (SVR), and right atrial pressure (RAP) were correctly predicted approximately 50% of the time. There was no significant difference in the ability to predict hemodynamics of subgroups with either hypotension or impaired oxygenation. After catheterization, planned therapy was altered in 58% of the cases. Unanticipated therapy was added in 30% of the cases. This study documents the difficulty of predicting accurately hemodynamics based solely on clinical evaluation. Thus, the information obtained by catheterization often leads to alterations in the therapeutic plan. We suggest that PA catheterization is both indicated and useful in the management of critically ill patients.