OBJECTIVES: Trunk position at 45 degrees from the horizontal is associated with a decreased risk of gastroesophageal aspiration. The objectives of this study were to determine the accuracy of trunk flexion estimates compared to a reference standard measurement, and to determine agreement about trunk flexion among ICU clinicians. DESIGN: Prospective observational study. SETTING: Two university-affiliated medical-surgical ICUs. PATIENTS AND PARTICIPANTS: Thirty-three mechanically ventilated ICU patients, seven residents, two fellows, three intensivists, and twenty-eight bedside nurses. INTERVENTIONS: Prospectively, concurrently, and independently during rounds, one bedside nurse, one resident, one fellow, and one intensivist clinically estimated the trunk flexion of mechanically ventilated patients. To record the reference standard, a trained investigator measured trunk position in the vertical plane using a goniometer. MEASUREMENTS AND RESULTS: We made 438 clinical assessments on 33 patients aged 57.2+/-19.4 (SD) years with an APACHE II score of 27.3+/-9.4. Mean trunk flexion estimates were: nurses 24.3+/-12.3 degrees from the horizontal, residents 20.2+/-13.7, fellows 20.3+/-10.8, and intensivists 21.1+/-13.1 compared to the reference standard measurement 16.2+/-9.0 degrees. The accuracy of trunk flexion estimates was fair to moderate [intraclass correlation for reference standard versus nurses (ICC 0.42), residents (ICC 0.52), fellows (ICC 0.36), and intensivists (ICC 0.55)]. The agreement among different groups of clinicians was moderate. CONCLUSIONS: In mechanically ventilated patients, we found that clinical estimates of trunk position were moderately good, agreement amongst caregivers was moderately good, but that all clinicians tended to overestimate the angle of semirecumbency.