Identifying key fall risk factors in older adults within a multifactorial intervention using feature selection methods

• Raul R. Teles Universidade de São Paulo (USP)
• Mariana L. de Melo Universidade Federal de São Carlos (UFSCar)
• Karina Gramani-Say Universidade Federal de São Carlos (UFSCar)
• Juliana H. Ansai Universidade Federal de São Carlos (UFSCar)
• Paulo H. S. Pelicioni University of New South Wales
• Jean R. Ponciano Universidade de São Paulo (USP)

DOI: https://doi.org/10.5753/sbbd_estendido.2026.249576

Resumo

Approximately 27% of older adults experience fall-related injuries each year. To reduce falls, multifactorial interventions assess individual risk factors and develop targeted care plans. While these interventions may reduce falls, their success depends on the accurate identification of risk factors and the implementation of appropriate plans. In this paper, we applied feature selection methods to identify relevant variables for predicting falls in three distinct moments of an intervention, achieving precisions of 0.6, 0.6 and 0.57 and recalls 0.71 0.75 and 1.0. Relevant factors found include mobility, anxiety and depression, fear of falling, home hazards, physical activity, and income. We expect these results to assist domain experts by enabling data-driven decision-making.

Palavras-chave: feature selection, fall risk, older people

Referências

Alberto, S. N., Ansai, J. H., Janducci, A. L., Florido, J. V. B., Novaes, A. D. C., Caetano, M. J. D., Rossi, P. G., Tavares, L. R. C., Lord, S. R., and Gramani-Say, K. (2022). A Case Management Program at Home to Reduce Fall Risk in Older Adults (the MAGIC Study): Protocol for a Single-Blind Randomized Controlled Trial. JMIR Res Protoc, 11(6):e34796.

Capodici, A., Fanconi, C., Curtin, C., Shapiro, A., Noci, F., Giannoni, A., and Hernandez-Boussard, T. (2025). A scoping review of machine learning models to predict risk of falls in elders, without using sensor data. Diagnostic and Prognostic Research, 9(1):11.

Ebrahimi, A., Wiil, U. K., Naemi, A., Mansourvar, M., Andersen, K., and Nielsen, A. S. (2022). Identification of clinical factors related to prediction of alcohol use disorder from electronic health records using feature selection methods. BMC Medical Informatics and Decision Making, 22(1):304.

Guyon, I., Weston, J., Barnhill, S., and Vapnik, V. (2002). Gene selection for cancer classification using support vector machines. Machine learning, 46(1):389–422.

Hopewell, S., Adedire, O., Copsey, B. J., Boniface, G. J., Sherrington, C., Clemson, L., Close, J. C., and Lamb, S. E. (2018). Multifactorial and multiple component interventions for preventing falls in older people living in the community. Cochrane database of systematic reviews, (7).

Li, Y., Hou, L., Zhao, H., Xie, R., Yi, Y., and Ding, X. (2023). Risk factors for falls among community-dwelling older adults: A systematic review and meta-analysis. Frontiers in medicine, 9:1019094.

Liang, H.-W., Ameri, R., Band, S., Chen, H.-S., Ho, S.-Y., Zaidan, B., Chang, K.-C., and Chang, A. (2024). Fall risk classification with posturographic parameters in community-dwelling older adults: a machine learning and explainable artificial intelligence approach. Journal of neuroengineering and rehabilitation, 21(1):15.

Maruf, M., Haque, M. M., Hasan, M. M., Farhan, M., and Islam, A. (2025). State-of-the-art review on fall prediction among older adults: Exploring edge devices as a promising approach for the future. Measurement: Sensors, 39:101878.

Remeseiro, B. and Bolon-Canedo, V. (2019). A review of feature selection methods in medical applications. Computers in biology and medicine, 112:103375.

Saunders, S., D’Amore, C., Hao, Q., El-Moneim, N. A., Richardson, J., Kuspinar, A., and Beauchamp, M. (2025). Risk factors for falls in community-dwelling older adults: An umbrella review. J AM MED DIR ASSOC, 26(9):105765.

Zhang, H., An, Y., Song, M., and Meng, Y. (2025). Dynamic fall risk prediction in hospitalized cancer patients: development and validation of a machine learning model using multidimensional clinical data to overcome over-sensitivity in traditional scales. BMC Medical Informatics and Decision Making, 25(1):377.