Paper
Saturday, July 14, 2007
Validation of a three step fall risk assessment tool
Cathy R. Robey-Williams, BSN, MS, Behavioral Health, Spartanburg Regional Hospital System, Spartanburg, SC, USA and Kathy Lynn Rush, PhD, RN, Nursing, University of South Carolina Upstate, Spartanburg, SC, USA.
| Learning Objective #1: Describe the research design utilized to compare a new three step fall risk assessment tool with the Morse fall risk screening tool. |
| Learning Objective #2: Analyze findings of determined fall risk from two tools with actual incidence of patient falls in acute hospitalized patients. |
Fall prevention continues to be a top safety priority for healthcare organizations. This facility undertook an extensive literature review to identify an evidence-based assessment tool that was reliable, valid, and easy to use. Three major elements of fall risk were identified in the literature and were incorporated into the Spartanburg Fall Risk Assessment Tool (SFRAT). The hospital internal review board approved a comparative study between the Morse and SFRAT fall risk assessment tool. Nursing units were randomized and the admitting nurses performed the fall risk assessments on all patients included in this study. The admitting nurse’s maintained consistency with the units selected for study by completing both assessments on the same patient within minutes of each measure. Study RN’s were consistent throughout the study period. The initial reliability of the tool demonstrated strong statistical significance (Kappa .90) and prompted wide adoption of the tool for clinical use. Data collection continued for 2 years until the number of patient falls in the study sample of assessments met the minimum sample size for a statistical power (.80). The total sample of patients included in the study was 4278. A total of 118 patients had at least one fall during their hospital stay. The sensitivity of the SFRAT was found to be 88% and the specificity 24.62%. The Morse tool was completed at the same time and with the same patient as the SFRAT which demonstrated a sensitivity of 71.19% and specificity of 53.08%. Utilizing a binomial proportion test the SFRAT was significantly higher than Morse tool in sensitivity to risk for fall (p=0.0012). Demographics of the two populations of patients (fall versus non fall) found that the patients that fell had a mean age of 5 years older and length of stay was twice the non fall sample.