Poster Q11 - Older Adult Fall-Related Head Impact Velocities and Associated Impact Kinematics

Introduction: An estimated 22% of the 3 million older adults (those above 65 years old) who visit the emergency department after a fall suffer from a head injury, and the majority (50% - 80%) of traumatic brain injuries (TBIs) in this population are caused by falls. The increasing incidences of TBIs in older adults, resulting predominantly from falls, underscores a pressing public health challenge. Despite older adults experiencing the highest rates of TBIs, leading to severe outcomes compared to younger individuals, there remains a considerable gap in our understanding of the head impact kinematics during these falls. To comprehend the kinematics and resultant injuries from these falls, it's crucial to quantify real-world impact boundary conditions. This study first aimed to quantify the head impact velocity of older adult falls and determine if fall characteristics influenced head impact velocity. The second aim of this study was to reconstruct these impacts with a headform and measure the peak linear acceleration (PLA), peak linear velocity (PLV), peak rotational acceleration (PRA), and peak rotational velocity (PRV) resulting from impact and evaluate if fall characteristics effect impact kinematics. This study is the first to match individual fall characteristics to assess linear and rotational head kinematics. This research is the first step to linking older adult head impact kinematics to resulting injuries and paves the way for creating a biomechanical testing framework to assess protective wall padding and compliant flooring.

Materials and

Methods: This study used an open-source dataset of 300 fall videos from long-term care facilities involving individuals aged 58 to 98 years (82.8 ± 7.6). Head impacts were evaluated by direct observation or injury reports, and fall characteristics were recorded based on factors known to affect older adults' fall mechanics (Table 1). Head impact velocities, both normal and tangential, were tracked from perpendicular view videos (N=29) using model-based image mapping (MBIM) software. The effect of fall characteristics on head impact velocities was evaluated using a MANOVA. An inverted pendulum with a 50th percentile Hybrid III headform and neck was used to reconstruct the impacts (N=23) (Figure 1). Impact speeds tracked from MBIM were adjusted using conservation of momentum due to the pendulum arm's larger effective mass. A speed gate confirmed the drop speed, within 0.02 m/s, and the headform's impact location aligned with video data for each impact.
The headform included three accelerometers and an angular rate sensor (ARS), recording data at 20 kHz. Collected data were filtered using a 4-pole phaseless Butterworth low-pass filter with determined cutoff frequencies of 800 Hz for accelerometer and 255 Hz for ARS signals. Mass scaling laws were applied to the linear acceleration and rotational velocity captured, enabling the correlation of inverted pendulum captured data to that of a falling person's head. A linear model was used to examine the effect of fall characteristics head impact dependent variables: PLA, PLV, PRA, and PRV.

Results, Conclusions, and Discussions: 118 out of the 300 (39%) falls evaluated had an identified head impact (Table 1). This study found the average normal head impact velocity to be 1.76 ± 1.02 m/s and tangential to be 1.27 ± 0.95 m/s (Figure 2). The linear model revealed that no fall characteristic affected the head impact normal or tangential velocity. Previous research has identified injury tolerance thresholds for PLA, indicating risks for skull fractures (around 250-300 g), subdural hematomas (approximately 130 g), and mild-TBI (above 80-100 g). We found that PLA ranged from 14.8 to 151.5 g, averaging 50.2 g across the reconstructed falls (Figure 3). Similarly, PRA as low as 3.00 krad/s2 could lead to a mild-TBI, and values closer to 10.00 to 16.00 krad/s2 could result in a diffusion axonal injury. The reconstructed falls produced PRA ranging from 1.01 to 8.18 krad/s2, with an average of 2.91 krad/s2 (Figure 3). Notably, only four reconstructed falls resulted in PLAs above 80 g and six recorded PRAs exceeding 3.00 krad/s2 impacts that could produce a TBI. These are lower than what has been previously reported for reconstructed falls. However, the falls analyzed in this dataset did not have any resulting catastrophic injuries, which aligns with our findings. Impacts that occurred against the floor had a higher PLA (p = 0.201), PRA (p = 0.431), and PRV (p = 0.466) compared to wall impacts (Figure 3). Evaluating the influence of fall characteristics, direction (Figure 4) was most influential on head impact kinematics. Nonetheless, PRV remained consistent across different fall directions. This study advanced understanding of head impact kinematics from older adult falls but was limited by its small sample size of 29 tracked velocities and 23 reconstructed impacts, and a dataset that excluded falls resulting in top-of-head impacts or occurring outside common areas, omitting scenarios like impacts against furniture. Our findings highlight the need to accurately identify older adult head impact boundary conditions for laboratory reconstructions. This knowledge is essential for establishing biomechanical testing frameworks that can accurately assess the protective efficacy of wall padding and compliant flooring.

Acknowledgements (Optional):