Validation of the LASA fall risk profile for recurrent falling in older recent fallers

Abstract

Objectives The fall risk profile developed in the Longitudinal Aging Study Amsterdam (LASA) identifies community-dwelling elderly at high risk for recurrent falling. This study assessed the predictive validity of this profile in older persons seeking care after a fall. Study Design and Setting The LASA fall risk profile was completed for 408 persons of 65 years and older who consulted the emergency department or general practitioner after a fall. Falls were prospectively reported with a calendar during 1 year. Recurrent falling was defined as ≥2 falls within a period of 6 months. Results During 1 year of followup, 76 (18.6%) participants became recurrent fallers. The area under the receiver operating characteristic curve was 0.65 (95% confidence interval [95% CI]: 0.58–0.72). At a cutoff value of 8, the sensitivity was 56.6% (CI: 51.8–61.4), the specificity was 71.4% (CI: 67.0–75.8), the positive predictive value was 34.1% (CI: 29.5–38.7), and the negative predictive value was 85.6% (CI: 82.2–89.0). Conclusion The discriminative ability of the LASA fall risk profile was moderate. The predictive validity of the LASA fall risk profile to identify recurrent fallers is limited among older persons who consulted the emergency department or general practitioner after a fall.