Accuracy of Physical Function-based Fall Risk Assessments Among Community-Dwelling Older Adults: An ROC Analysis

Physical function assessments, such as the Timed-Up-and-Go (TUG) and Short Physical Performance Battery (SPPB), are commonly used to evaluate fall risk among older adults, yet they may misclassify individuals. Center-of-pressure postural sway path length (PS) is a criterion for assessing fall risk, but it is unclear how TUG and SPPB compare. This study examined the sensitivity, specificity, and accuracy of these three assessments in classifying older adults as high or low fall risk.

We conducted a cross-sectional study of 234 community-dwelling older adults (women=203, age=75.0±7.0 years, height=159±7.9cm, Body Mass Index=30.0±6.3kg/m²). Cut off values for SPPB (<7 out of 12), TUG (≥20 seconds), and PS (>30 cm) were used to categorize participants as high or low fall risk. McNemar tests were used to compare fall risk assessment between assessments.

Participants had a mean TUG time of 10.2±6.3 seconds, SPPB score of 8.7±2.3, and PS of 34.7±21.3cm. Both TUG (high risk=10, low risk=224; X²=95.9, p<0.001) and SPPB (high risk=40, low risk=194; X²=55.7, p<0.001) significantly differed in fall risk assessment compared to PS (high risk=115, low risk=119). Compared to postural sway, TUG had a sensitivity of 4.3%, specificity of 95.8%, and accuracy of 50.9%. Similarly, SPPB had a sensitivity of 23.5%, specificity of 89.1%, and accuracy of 56.8%.

TUG and SPPB both demonstrated low sensitivity and accuracy, suggesting that they may be better as measures of physical function rather than fall risk. Clinicians may benefit from using PS or other criterion measures to screen for fall risk.