Objective To compare the effectiveness of the STRATIFY falls tool with nurses' clinical judgments in predicting patient falls. Study Design and Setting A prospective cohort study was conducted among the inpatients of an acute tertiary hospital. Participants were patients over 65 years of age admitted to any hospital unit. Sensitivity, specificity, and positive predictive value (PPV) and negative predictive values (NPV) of the instrument and nurses' clinical judgments in predicting falls were calculated. Results Seven hundred and eighty-eight patients were screened and followed up during the study period. The fall prevalence was 9.2%. Of the 335 patients classified as being "at risk" for falling using the STRATIFY tool, 59 (17.6%) did sustain a fall (sensitivity = 0.82, specificity = 0.61, PPV = 0.18, NPV = 0.97). Nurses judged that 501 patients were at risk of falling and, of these, 60 (12.0%) fell (sensitivity = 0.84, specificity = 0.38, PPV = 0.12, NPV = 0.96). The STRATIFY tool correctly identified significantly more patients as either fallers or nonfallers than the nurses (P = 0.027). Conclusion Considering the poor specificity and high rates of false-positive results for both the STRATIFY tool and nurses' clinical judgments, we conclude that neither of these approaches are useful for screening of falls in acute hospital settings