Predicting which patient will fall again ... and again

Abstract

In order to compare the characteristics, preventive interventions and outcomes of single and multiple fallers, a retrospective cross-sectional study was conducted in a 680-bed acute-care hospital in Western Australia. Fifty patients falling more than once (multiple fallers) were randomly selected from all patients reported to have fallen between 1 July 1989 and 31 December 1989, and age-sex matched with 50 patients falling once in the trial period (single fallers). In total, 382 in-patients were reported to have sustained 578 falls in the 6-month trial period. Fifty-two per cent of these falls involved multiple fallers. An analysis of the 100 single and multiple fallers showed that single fallers were more likely to have fallen from their bed; be discharged home from hospital; and be clinically deteriorating at the time of the fall. Multiple fallers were more likely to be transferred to a long-term nursing facility after discharge from hospital; suffer blindness/poor vision; be sedated post fall; be ordered to be restrained following a fall; and be hospitalized for longer periods. There was also a tendency for multiple fallers to repeat the type and location of the fall on successive falls. Stepwise logistic regression showed that falling from the bed on the first fall predicted remaining a single faller. Being ordered to be restrained following the first fall and hospitalized for longer periods predicted the patient would fall repeatedly. Further analytical research incorporating an expanded number of independent variables is needed to allow confident assertions of causality. To test the effectiveness of preventive measures, a prospective longitudinal study is required.