[Development of a portable fall risk index for elderly people living in the community]

To develop a portable risk index for falls.

Risk factors were chosen from previously established factors then we added several environmental factors to the risk index; previous falls in the last 12 month, trippig or stumbling, inability to ascend or descend stairs without help, decreased walking speed, inability to cross a road within the green signal interval, inability to walk 1km without a break, inability to stand on one leg for 5 seconds (eyes open), using a cane, inability to wring out a towel, dizziness or faintness, stooped or rounded back, knee joint pain, visual disturbance, hearing disturbance, cognitive decline, fear of falling, receiving 5 or more prescribed drugs, sensation of darkness at home, obstacles inside, barrier on the carpet or floor, using steps daily at home, steep slopes around home.

The questionnaire sheet was completed by 2,439 community-dwelling elderly subjects (76.3 +/- 7.4 years old). The frequency of each items of fall risk index was compared between fallers (history of fall within one year) and non-fallers. Multiple regression analysis was performed to identify independent risk factors for previous falls.

Except barrier, step use and steep slope around home, all items in the fall risk index were more frequent in fallers. Multivariate analysis revealed that tripping or stumbling, inability to cross a road within the green signal interval, dizziness or faintness, obstacles inside, inability to wring out a towel, cane use and knee joint pain were independent risk factors for previous falls. These 7 selected items were further analyzed as predictors. The maximum sum of sensitivity and specificity was reached at the cut-off point of 2/3 (sensitivity 0.65, specificity 0.72) by receiver operating curve.

Portable fall risk index is useful for clinical settings to identify high-risk subjects.