Validity of the STRATIFY risk score of falls in nursing homes

Objective. We investigated the performance of the STRATIFY risk assessment for fall risk in nursing home patients. Methods. This prospective study was conducted at the nursing home Antonius Binnenweg in Rotterdam from June 1 to December 31, 2000. Subjects included newly admitted patients of whom a STRATIFY score was obtained. Patients were followed for falls during a person-week observation period of 13 weeks. Measurements were age, sex, admission category, STRATIFY score, length of stay and number of falls. Results. Of 120 patients (75 women, 45 men, average age 74.5 years), a STRATIFY score was available. Three months after admission 24 of these patients (20%) had died, 56 patients (47%) had been discharged and 40 (33%) still resided in the nursing home. Of the 120 patients included, 36 patients (30%) had fallen during the observation period. The risk of falling in the patient group with a STRATIFY score of 2 or more was 2.35 (95% confidence interval: 1.22, 4.52) times that of patients with a lower score. The sensitivity of the STRATIFY score was 50% (95% CI: 32, 67), specificity was 76.2% (95% CI: 65, 87.5). Conclusion. The STRATIFY risk score of falls gave disappointing results in our setting.