A retrospective comparative chart audit was conducted to identify patient characteristics associated with falls in the acute care setting, to examine the extent to which the significant characteristics explained if falls occurred, and to test the ability of variables believed to be risk factors to predict falls. Patients aged 60 and older who fell during hospitalization (n = 331) were compared with a random sample of patients aged 60 and older who were hospitalized during the same time period but did not fall (n = 300). Two days of documentation were sampled: admission day and day preceding the fall for the fall group, and admission day and a random day of hospital stay for the no-fall group. Findings supported the idea that fall-prone patients can be identified and that significant differences between those who do and do not fall are evident at hospital admission. The findings also suggested an alteration in the constellation of characteristics nurses use to identify fall-prone pattents. Of 11 variables representing standard risk factors, only 6 were significantly related to fall status; 5 entered the regression equation as significantly contributing to the 22% explained variance. When potential predictor variables were expanded to include additional patient characteristics, the explained variances for fall status were 31% from the admission day data and 34.5% from the fall/random day data.