Development of a Modifiable Fall Risk Score for evaluating changes in fall risk over time

Accurately estimating fall risk is a key aspect of community-based and clinical fall prevention programmes. Current methods predict 5-year fall risk for individuals. Since such methods include static characteristics (eg, demographics and history), they are less sensitive to changes in modifiable characteristics resulting from fall prevention programmes. This study describes the development and evaluation of a Modifiable Fall Risk Score (MFRS) designed to predict changes in risk using modifiable predictors.

The National Health and Aging Trends Study (NHATS) is a nationally representative longitudinal cohort of Medicare beneficiaries in the USA. Participants with at least two consecutive years of complete fall history and risk factor data were included in the development of MFRS to predict 1-year fall risk.

Data from 5789 individuals aged 65 and older were used to create 28 206 1 year, fall risk predictions. Weighted risk scores and thresholds accurately identified non-fallers (sensitivity=0.895, negative predictive value=0.803) and fallers (specificity=0.877, positive predictive value=0.576). Fall history did not interact with modifiable predictors, demonstrating functional changes impact risk independent of fall history. Age was not a significant predictor of risk, further highlighting the importance of function for fall prevention.

The study provides novel insights into the assessment of fall risk and the importance of objective functional tests in risk estimation. The composite MRFS is suitable for estimating 1 year fall risk and changes in risk over time. The use of average- and high-risk thresholds offers advantages for fall risk screening.