BACKGROUND AND OBJECTIVE To describe the diagnostic accuracy and practical application of the Peter James Centre Falls Risk Assessment Tool (PJC-FRAT), a multidisciplinary falls risk screening and intervention deployment instrument. METHODS In phase 1, the accuracy of the PJC-FRAT was prospectively compared to a gold standard (the STRATIFY) on a cohort of subacute hospital patients (n = 122). In phase 2, the PJC-FRAT was temporally reassessed using a subsequent cohort (n = 316), with results compared to those of phase 1. Primary outcomes were falls (events), fallers (patients who fell), and hospital completion rates of the PJC-FRAT. RESULTS In phase 1, PJC-FRAT accuracy of identifying fallers showed sensitivity of 73% (bootstrap 95% confidence interval CI = 55, 90) and specificity of 75% (95% CI = 66, 83), compared with the STRATIFY (cutoff > or = 2/5) sensitivity of 77% (95% CI = 59, 92) and specificity of 51% (95% CI = 41, 61). This difference was not significant. In phase 2, accuracy of nursing staff using the PJC-FRAT was lower. PJC-FRAT completion rates varied among disciplines over both phases: nurses and physiotherapists, > or =90%; occupational therapists, > or =82%; and medical officers, > or =57%. CONCLUSION The PJC-FRAT was practical and relatively accurate as a predictor of falls and a deployment instrument for falls prevention interventions, although continued staff education may be necessary to maintain its accuracy.