Applying root cause analysis to improve patient safety: decreasing falls in postpartum women.

摘要

OBJECTIVE: To investigate the effectiveness of interventions to prevent falls designed through hazard analysis using root cause analysis. DESIGN: Prospective longitudinal study. Under preceding root cause analysis, root factors were classified into four major categories: environment and facilities, procedure, individual, and communication. Among them, communication, environment and facilities were recognised as the most vital factors to facilitate intervention accordingly. The fall prevention programme included first intervention with adding live demonstrations and offering a printed education sheet, and second intervention with adjusting rails at the bedside and in the ward, placing anti-sliding pads on the floors of the bathrooms and enhancing local light in the bathrooms. SETTING: Two large hospitals in Taiwan, both with an average of 200 or more childbirths per month, with an intervention group and a non-intervention control group. PARTICIPANTS: 2460 (intervention group) and 2451(control group) participants. MAIN OUTCOME MEASURES: The number of postpartum falls within 6 months before and after intervention, and the incidence after adjustment by patient-days. RESULTS: In the intervention group, the incidence of postpartum falls before intervention was 14.24 per 1000 patient-days and dropped to 6.02 per 1000 patient-days after intervention. The control group showed no marked decline in incidence with a rate of 13.72 and 14.05 per 1000 patient-days, respectively. Using the Mantel-Haenszel test to compare the incident rate, there were significant differences (p<0.001) between the incidence of the intervention group before and after intervention; and between the latter and each incidence of the control group. CONCLUSIONS: This study provides direct evidence that root cause analysis can be adopted in analysing causes and in formulating interventions to reduce the incidence of postpartum falls and improve patient safety.