External validation and further exploration of fall prediction models based on questionnaires and daily-life trunk accelerometry

bioRxiv (Cold Spring Harbor Laboratory)

Abstract

Abstract Background Ambulatory measurements of trunk accelerations can provide valuable insight into the amount and quality of daily life activities. Such information has been used to create models that aim to identify individuals at high risk of falls. However, external validation of such prediction models is lacking, yet crucial for clinical implementation. We externally validated three previously described fall prediction models (van Schooten et al., 2015a). Methods Complete questionnaires and one week of trunk acceleration data were obtained in 263 community-dwelling people (mean age 71.8 years, 68.1% female). To validate models, we first used the coefficients and optimal cut-offs from original cohort, then recalibrated the original models, as well as optimized parameters based on our new cohort. Results Among all participants, 39.9% experienced falls during 6-month follow-up. All models showed poor precision (0.20-0.49), poor sensitivity (0.32-0.58), and good specificity (0.45-0.89). Calibration of the original models had limited effect on model performance. Using coefficients and cut-offs optimized on the external cohort also had limited benefit. Last, the odds ratios in our cohort were different from those in the original cohort and indicated that gait characteristics, except for index of harmonicity ML, were not statistically significantly associated with falls. Conclusions Prediction of fall risk in our cohort was not as effective as in the original cohort. Recalibration as well as optimized model parameters resulted in limited increase in accuracy. Fall prediction models are highly specific to the cohort studied. This highlights the need for large representative cohorts, preferably with an external validation cohort.

Authors 5

1. Yuge Zhang

   Vrije Universiteit Amsterdam

   Affiliation as printed

   Department of Human Movement Sciences, Faculty of Behavioural and Movement Sciences, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands

2. Roel H. A. Weijer Aachen Department of Neurology, Leiden University Medical Center (LUMC)

   Vrije Universiteit Amsterdam · Leiden University Medical Center

   Affiliation as printed

   Department of Human Movement Sciences, Faculty of Behavioural and Movement Sciences, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands

   Department of Neurology, Leiden University Medical Center, Leiden, The Netherlands

3. Kimberley S. van Schooten

   UNSW Sydney · Neuroscience Research Australia

   Affiliation as printed

   Neuroscience Research Australia, University of New South Wales, Sydney, Australia

   School of Population Health, University of New South Wales, Sydney, Australia

4. Sjoerd Matthijs Bruijn corresponding

   Vrije Universiteit Amsterdam · Fujian Medical University

   Affiliation as printed

   Biomechanics Laboratory, Fujian Medical University, Quanzhou, Fujian, PR China

   Department of Human Movement Sciences, Faculty of Behavioural and Movement Sciences, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands

   Institute of Brain and Behavior Amsterdam, Amsterdam, The Netherlands

5. Mirjam A. G. M. Pijnappels

   Vrije Universiteit Amsterdam

   Affiliation as printed

   Department of Human Movement Sciences, Faculty of Behavioural and Movement Sciences, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands