3

Prevention

___________________

¹ This list is the rapporteurs’ summary of points made by the individual speakers identified, and the statements have not been endorsed or verified by the National Academies of Sciences, Engineering, and Medicine. They are not intended to reflect a consensus among workshop participants.

The third session of the workshop featured a firsthand account of a traumatic brain injury (TBI) before exploring selected examples of three major causes of TBI—sports injuries, motor vehicle collisions, and falls in older adults—with a focus on how technology and innovation can affect mechanisms of injury and the potential to personalize protection and prevention moving forward. Objectives of the session were (1) reviewing prevention strategies across several leading causes of TBI, (2) identifying innovations in both low- and high-tech approaches to TBI prevention in various contexts, and (3) exploring future opportunities for personalized approaches to mitigate injury. The session was moderated by Kristy Arbogast, R. Anderson Pew Distinguished Chair, Department of Pediatrics at Children’s Hospital of Philadelphia/University of Pennsylvania, who described preventing injuries as the most effective way to address the burden of TBI.

A LIVED EXPERIENCE PERSPECTIVE

Wesley Ilana Schnapp, TBI survivor, graduate student at the University of Arizona, and Christine Mirzayan Science & Technology Policy Fellow at the National Academies of Science, Engineering, and Medicine, provided a personal account of sustaining and recovering from TBI. As a teenager, Schnapp was a competitive downhill alpine ski racer. In April 2009, at the age of 15, she had an accident during routine training on Oregon’s Mount Hood. Having no memory of the actual accident, she has pieced together her memories of the morning with the accounts others have given her. Schnapp, who began skiing at age 3 and racing at age 7, was an experienced skier at this point in her adolescence. The slopes were icy that morning,

not yet having warmed in the sun. With helmet in place, she was following a friend downhill when she hit an unexpected bump. On her landing, she struck the strip of her forehead between helmet and goggles. She immediately lost consciousness and her friend called the ski patrol, who sedated her and transported her to the emergency clinic located on the slope. The clinic intubated her upon arrival, and she then received a life flight by helicopter to a hospital in Portland, Oregon. Schnapp remained in a coma in the intensive care unit (ICU) for 4 days.

Schnapp awoke from the coma and became responsive, but her personal awareness did not begin to emerge until about a week after her injury, when she was transferred to a rehabilitation hospital. She described the weeks she spent at the rehabilitation hospital as surreal and dreamlike, as if she were in a parallel universe in which she sometimes believed herself to be dead. Despite being able to move and perform basic functions, Schnapp could barely speak, understand other people, or have the sense of knowing who she was. She recalled being asked if she knew where she was and—after searching for cues about her location—Schnapp saw markers a friend had brought her and replied that she was at Office Depot. When the question was repeated, she responded that she was at Home Depot. Unable to comprehend her context and surroundings, she went through motions but felt as if her conscious being was separate from her body.

Schnapp described that after several weeks, “something clicked” and she suddenly began comprehending her surroundings, no longer believed herself to be dead when reminded about the accident, and was able to recall some memories from the morning of the accident. After being discharged from the hospital, achieving full recovery was a long process that required several years, said Schnapp, noting that the malleability of her 15-year-old brain likely aided her recovery. This experience fostered her interest in the brain, and she is currently studying neuroscience in a doctoral program.

Prevention efforts range from those taken to avoid injury to those that avert the worst outcomes after injury, Schnapp stated. Recalling that she was wearing a top-of-the-line ski racing helmet and was a trained skier taking a routine ski run, she emphasized that having safety precautions in place will not eliminate all injuries. However, immediate intubation at the emergency ski clinic was a substantial contributing factor to attaining full recovery, she asserted. Schnapp noted examples of ski racers who were not intubated immediately following their accidents, instead being flown by helicopter to ICU without intubation, and did not survive. Therefore, Schnapp believes that intubation helped prevent worse outcomes after her accident. She closed by highlighting that with the approaching anniversary of her TBI—which she considers a “second birthday” on which she celebrates her survival—she feels gratitude to be alive, to share her experience, and to dedicate her life’s work to the brain community.

TBI PREVENTION STRATEGIES IN PROFESSIONAL AMERICAN FOOTBALL

The risks of TBI associated with professional football have received significant attention. Jennifer Langton, senior vice president of Player Health and Innovation at the National Football League (NFL), described NFL programs to understand behaviors that lead to injuries in players, to develop safety innovations, and to incorporate knowledge and technology into play. Since the days of leather helmets, the NFL has prioritized health and head safety as a leader in professional sports, she said, and has recognized the need to research and reduce brain injuries for their athletes. Over the past decade, the NFL has accelerated its efforts in science, statistical rigor, and innovation and has created an injury prevention paradigm that combines engineering, education, and enforcement to keep players safe, she continued. By disseminating its insights with other organizations and industries, the league hopes to advance technologies and practices for head injury prevention and detection in ways that are instructive beyond the context of professional football—which may include youth and collegiate sports, she said.

The NFL Engineering Road Map, developed in 2016 in collaboration with biomechanical engineering and injury prevention experts, is a comprehensive effort to better understand how head injuries happen on the field and to adopt learnings from areas beyond sports to catalyze the design of protective equipment.² The program’s initial goal was to make advances in head protection by understanding and measuring on-field injury mechanisms. The next phase was the creation of laboratory tests for helmets under realistic conditions. Data collected from these efforts are shared with manufacturers to enable improved helmet design. Ultimately, the NFL aims to achieve innovative, position-specific helmet designs, said Langton.

Helmet Performance Ranking

Helmet properties have improved by nine times since the NFL began helmet testing in 2015, Langton stated, with better performing helmets contributing to an average 25 percent reduction in concussion (a form of mild TBI). Each year, the NFL and the NFL Players Association engineering partners collaborate to conduct laboratory tests to determine which helmets are best able to manage impacts sustained by players on the field. The helmets are then ranked from best to worst performance. The NFL created a poster with a graphic of the ranked helmets, including those indicated

___________________

² See https://www.nfl.com/playerhealthandsafety/resources/fact-sheets/nfl-engineering-roadmap-fact-sheet (accessed June 4, 2024).

as “not recommended” and “newly prohibited,” and displays this poster in all locker rooms across NFL facilities. Clubs, equipment managers, and medical personnel use the poster to educate players and encourage them to choose the safest helmet available to them. Helmets prohibited because of poor performance are indicated in red and are not eligible for use. Langton noted that helmet innovation is one reason for changes in the helmet rankings over time, with some helmets that once topped the list becoming prohibited within a few years.

Helmet Manufacturers and Marketplace

Manufacturers understandably work to avoid NFL prohibition of their models, said Langton. In an effort to bolster manufacturers’ ability to continually develop helmet models that demonstrate increased performance, the NFL shares data, test methods, helmet models, and insights with leading helmet manufacturers. The granularity of data is such that manufacturers can now design position-specific helmets for lineman and quarterback positions that offer protection tailored to the type of contacts players in these positions typically sustain. To help stimulate the marketplace, the NFL expanded collaboration efforts into advanced manufacturing and materials science, areas that have revolutionized design. For example, the NFL conducted crowdsourcing challenges to engage engineers and materials scientists who are not necessarily involved in sports manufacturing. These challenges yielded forward-thinking helmet designs via 3D printing and customizations that will be used during NFL play in the 2024 season. One such design, the Guardian Cap, is an add-on device for the exterior of the helmet that has demonstrated a 50 percent reduction in TBIs sustained during training camp over the past two seasons for players mandated to wear the device. In response to these data, the NFL is expanding the mandate to include almost all positions³ during all contact practices throughout the 2024 season; players will also be permitted to wear it during games, Langton noted.

Prevention Benefits and Next Steps

Langton reiterated the importance of helmet improvements since 2015. Innovation has led to 12 new top-performing helmets submitted for NFL testing in 2024, the highest annual number since testing began, she said. Of these 12 helmets, 8 are position-specific, providing linemen and quarterbacks with an available range of tailored helmets. Position-specific helmets for other positions are expected to become available in the next few years.

___________________

³ Use of the device is optional for positions where full contact with tackling defensive players is against the rules of play. This includes quarterbacks, kickers, and punters.

She remarked that the NFL has entered a new era in injury prevention, prevention technology, and performance optimization. In 2019, the NFL and Amazon Web Services (AWS) introduced the Digital Athlete, technology that uses artificial intelligence and machine learning to build a virtual representation of each player. This “digital twin” captures a complete record of the player’s data from training, practice, and game play, drawing on video, sensors, and other sources to capture and analyze plays and impacts. These data feed risk mitigation models that help teams understand the precise needs of their players in terms of staying healthy, recovering quickly, and optimizing their performance. Langton described this development as the next generation of player health and safety in the NFL.

The league’s prevention priority is reducing TBI, and the NFL is also working to reduce the quantity and severity of head impacts, said Langton. The Digital Athlete enables each hit a player sustains to be identified and counted. The NFL shares these data with coaches and players to encourage certain techniques and behaviors that have demonstrated decreased head impacts. Furthermore, the NFL adjusts on-field rules in an effort to prevent unnecessary head contact, such as the recently announced kickoff modification and use of the helmet rule (Kasabian, 2024; Madani, 2024). These changes aim to reduce the head impacts most likely to cause player injury. Through Digital Athlete analytics, the NFL used machine learning to model the new kickoff rule, which is expected to increase the number of returns while simultaneously reducing injuries. Study of over 2,000 on-field TBIs informed the creation of injury criteria for predicting injury. In helmet evaluations, these criteria have enabled correlation of testing with actual on-field injury rates. These examples by Langton highlighted how the NFL leverages findings from its research to inform safer play.

AUTOMATION, CUSTOMIZATION, AND TBI TRENDS IN MOTOR VEHICLE CRASHES

Richard Kent, Frederick Tracy Morse Professor of Mechanical and Aerospace Engineering, Biomedical Engineering, and Orthopedic Surgery at the University of Virginia, discussed anticipated changes in the automotive environment and how they may affect TBI, with a focus on vehicle automation and customization of protection systems. Describing the biomechanics involved in brain injury, Kent stated that hitting one’s head in the automotive environment carries potential for TBI, and the harder the hit, the greater the risk of both injury and severe injury.

He presented a video of a crash test in which air bags deploy to prevent the dummies sitting in the front car seats from hitting their heads. Rather than a full-frontal impact, the crash represents a scenario in which another car crosses the lane and strikes the test car with a slight offset of approxi-

mately 15 degrees of obliquity to the driver’s side. In the video, the dummy in the driver’s seat is thrown to the left of the steering wheel, and its head slides between the steering wheel airbag and the side airbag. Kent noted that although the dummy’s head narrowly misses hitting the door sill or the A pillar (i.e., the pillar of the car frame between the windshield and the driver side window), the outcome could be different if the dummy was 4 inches taller or shorter, was obese, had breast tissue between the seat belt and rib cage, or was leaning to change the radio station. Should any of these circumstances lead to a driver striking their head on the vehicle interior, the airbags would likely not prevent a TBI, Kent explained.

Vehicle Automation and TBI Risk

Partially automated vehicles are currently on the road, and marketing materials highlight that automation reduces the need to pay attention to the road, said Kent. Were a fully automated driving experience developed, he noted, passengers may be able to read, play cards, or rotate seats 180 degrees to sit backward and stretch their legs or talk to passengers in the back seat. Kent showed a marketing illustration in which a passenger is lying down on a long seat resembling a bed, propped on his elbow as he reads a book. Describing himself as a “passive safety guy,” Kent stated his horror at this idea, given that researchers have dedicated decades of study on restraining people inside of vehicles, and nothing is yet known about how to restrain a person in a relaxed, partially prone position. Emphasizing this point, he noted that crash test dummies able to assume such postures are yet to be designed. The Hybrid III crash test dummy, the most commonly used model, features a molded pelvis that is unable to recline⁴. Kent added that implementing a new dummy model—one able to sit in a reclined position—into federal compliance standards is a lengthy, challenging process. Moreover, vehicle evaluations are performed using repeatable methods. Hours are spent positioning each dummy within a millimeter of where it needs to be placed for repeated tests, and no methodology currently exists to repeatedly test dummies in the array of positions envisioned by innovators and marketers, said Kent.

Illustrating the effects of reclined positions on the effectiveness of current prevention equipment, Kent presented two videos of a cadaver in a simplified crash simulation focusing on head kinematics. He noted that this testing used cadavers because of the inability to position current crash test dummy models into reclined positions. The cadaver was positioned for two frontal impact tests in an open car seat with a seat belt over the right

___________________

⁴ For information on several commonly used crash test dummies, including Hybrid III models, see https://www.nhtsa.gov/nhtsas-crash-test-dummies (accessed September 24, 2024).

shoulder. In the first, the cadaver is seated in an upright position, which Kent termed a “neutral” position. The second test placed the cadaver sitting in the car seat but leaning 20 degrees to the left. He emphasized that this position is much closer to the neutral position than is a person reclining 45 degrees reading a book. The tests were performed using a motion capture system using an array of markers mounted to the cadaver’s bones, which enabled array tracking. After the crash tests, researchers processed the motion capture system data to re-create the skeletal kinematics and performed CT scans to obtain comparison imaging. The cadaver head ended each crash test in a similar location regardless of whether it began in the neutral or leaning position. However, the kinematics during the crash tests varied between the positions.

The head that began in the leaning position had a longer forward trajectory, traveling more than 2 inches beyond the trajectory of the head in the neutral position. This translates to a 27 percent increase in forward excursion and results in a higher risk of impact with the vehicle interior or another occupant. Furthermore, the head that began in the leaning position had a higher velocity than the head in the neutral position, demonstrating a 40 percent increase in relative velocity and a two-fold increase in kinetic energy over the neutrally positioned head. Thus, should the head come into contact with another object, higher kinetic energy increases the severity of the head strike, Kent explained. He summarized that leaning 20 degrees to the side increases the risk of a passenger hitting their head and increases the likelihood that it will be a hard hit.

Kent and colleagues have also studied the kinematics of car crash impact on a person leaning backward in a reclined car seat, again using cadavers owing to the limitations of crash test dummies. He presented a simulation video in which a person leaning back in a reclined car seat is pitched forward upon impact. The simulation shows that the pelvis slides forward in the seat and, because the spine began the trajectory in a reclined position, the abdomen is exposed to impact and is crushed by the lap belt. The spine then flexes forward, with the head pitching so far forward that it strikes the thighs and sternum. Kent emphasized that possible injuries from this scenario could include a lacerated liver, perforated colon or bowel, lumbar spinal fractures, and cervical spine fracture. A seat belt prevents a person in a typical seated position from hitting their head on their sternum, but the kinematics of the trajectory of a reclined person enable this type of impact. Kent stated that researchers for automotive companies are working to solve this problem in order to market autonomous cars. He cautioned that relying on autonomous cars to eliminate crashes is unrealistic, and he highlighted that this technology entails moral and ethical decisions moving forward. Solutions are needed to protect people in autonomous vehicles when impact does occur, said Kent.

Customization in Injury Risk Prediction

Airbag technology and 90 percent seat belt use rates have greatly reduced the incidence of subdural hematomas, massive skull fractures, and diffuse axonal injuries associated with car crashes, Kent stated. With these preventative gains in place, the physical factors that can influence injuries warrant attention. Researchers are beginning to study the effects of body shape on impact reaction through the use of obese dummies. The role of biological sex—and the associated differences in size, shape, and other factors—in crash injuries is not well understood. Improving prevention efforts for people of various ages and physical conditions is another area of study.

Kent reviewed data from a regression model charting full body car crash injury risk from 2002 to 2018 (Forman et al., 2019). During this time span, the risk of moderate injury for a seat-belted, 50-year-old man in a frontal 56 kilometer per hour (35 miles per hour) impact has decreased from approximately 25 percent to 16 percent. The risk for severe injury decreased from approximately 11 percent to 5 percent. Kent remarked that severe injuries are more easily reduced than subtle injuries. In contrast, the risk of injury for women, although declining, remains much higher than that for men. In 2002–2018, the risk of moderate injury for a seat-belted, 50-year-old woman in a frontal 56 kilometer per hour impact has decreased from approximately 41 percent to 29 percent—almost twice the risk rate for men—and the risk for severe injury has fallen from nearly 16 percent to approximately 8 percent. Thus, a significant sex effect has remained fairly constant in this regression, Kent highlighted.

Scientists do not yet fully understand the causes for the car crash injury risk disparity between men and women, which is driven largely by extremity injuries, said Kent. Men tend to drive heavier vehicles than women, and this may factor into injury rates. In addition, most testing is performed with male crash test dummies, and therefore research gaps on female bodies could be at play, he noted. Moreover, the greatest sex disparity in car crash injuries is seen in lower speed crashes and less severe injuries, areas that historically have not been studied as thoroughly as more severe crashes, he added. Research indicates that sex has no effect on skull fractures. However, the odds ratio for moderate brain injury is 1.7 for sex (Forman et al., 2019), indicating a fairly dramatic effect of sex on brain injury, he explained. Research in this area is new and the reasons for this effect have not been established.

Kent remarked that research on skull fractures is far more robust. Researchers at Wayne State University measured acceleration on cadavers dropped down elevator shafts in the 1960s; the data from those studies were used in the Head Injury Criterion, a primary head injury prediction tool (Wang et al., 2021). The tool predicts skull fracture risk based on linear

acceleration, and subsequent car design has greatly reduced the risk of skull fracture. However, Kent stated that tools are not yet in place for predicting risk of moderate TBI, particularly given that variables such as sex appear to play a significant role in risk level.

Kent offered an overview of brain injury prediction formulas used in the automotive world, which he noted are in part based on NFL research. Angular velocity is used to predict risk of injury from impact that is fast, and angular acceleration is a better predictor of risk for impact that is slow. The Diffuse Axonal Multi-Axis General Evaluation (DAMAGE) algorithm accounts for various tolerances and sensitivities within different anatomical axes in the brain (Gabler et al., 2018, 2019). Moreover, DAMAGE incorporates the entire history of the motion of the skull to predict strains within the brain. The algorithm can predict risk from simulations of impact to an NFL helmet, which lasts about 10 milliseconds, to a pedestrian struck by a car with a 300–400 millisecond acceleration of the head toward the windshield. Kent remarked that the ability to predict injury from impact ranging from 10 to 400 milliseconds reflects a robust tool to effectively predict TBI risk. He described recent research progress using the DAMAGE algorithm with crash test dummies (Gabler et al., 2018, 2019; Reynier et al., 2022), and correlating these calculations with strain within brain tissue and along axonal pathways in the brain. The goal of this research is to develop a tool able to predict axonal strain from crash test dummy measurements to enable evaluating a vehicle for TBI risk.

Approximately 2 years ago, a collaboration between Kent’s laboratory and the Insurance Institute for Highway Safety (IIHS) yielded a DAMAGE injury criterion specifically for the small female, enabling injury prediction for both males and females (Reynier et al., 2022). IIHS sponsored this research in hopes of using the DAMAGE injury criterion in ranking vehicles according to safety. Kent maintained that using such tools can help address the TBI risk disparities in actual vehicles. He noted that the DAMAGE algorithm is being evaluated by the European New Car Assessment Programme—a group that issues crash safety ratings for every vehicle in Europe—for use in new car assessment programs. Kent recommended that the DAMAGE algorithm should eventually be incorporated into federal compliance tests.

INNOVATIVE APPROACHES TO FALL PREVENTION

Thurmon Lockhart, MORE Foundation Professor of Life in Motion in the School of Biological Health and Systems Engineering at Arizona State University, provided an introduction to issues related to falls in older

adults, mechanisms related to fall accidents, and current approaches to reducing falls, including a wearable assessment tool.⁵

Consequences of Falls in Older Adults

Significantly influenced by aging, TBI disproportionately affects older adults, with people aged 75 years and older making up approximately 40 percent of TBI-related hospitalizations and TBI-related deaths (CDC, 2021). The leading cause of TBI in this age group is falls, with approximately 64 percent of TBI resulting from a fall on the same level from slipping, tripping, or stumbling. Furthermore, rates of TBI in older adults have increased substantially since 2001, with fall accidents increasing annually (Harvey and Close, 2012). Lockhart noted that the U.S. government expects 52 million falls with 12 million resulting in hospitalization annually by 2030. Approximately 39,000 older adults died from falls in 2021 (Kakara et al., 2023), while about 20 percent of older adults who experience a hip fracture die within a year (Lu-Yao et al., 1994). Many older adults who survive falls experience a diminished quality of life after developing a fear of falling that leads to reduced activity and outings, which in turn fosters degradation. He highlighted that after experiencing a fall, 20–36 percent of older adults develop a fear of falling (Vellas et al., 1997) and 25 percent require nursing home care within a year (Magaziner et al., 2000). Thus, prevention efforts to avoid a first fall are needed, said Lockhart.

Fall Risk Factors and Assessment

Fall intervention is divided into fall protection and prevention. Fall protection includes personal protective equipment, such as helmets and harnesses. Fall prevention is complicated by environmental context and the coefficient of friction, which change with humidity and other factors that are nearly impossible to standardize, he explained. Interventions such as strength training and balance exercises can reduce falls. Lockhart highlighted the importance of assessment in implementing appropriate interventions. For instance, a person may be more prone to falling because of an issue with gait, balance, or musculoskeletal degradation. However, current fall assessment tools, such as the Morse Fall Scale,⁶ do not provide the underlying cause of fall risk and therefore cannot be used to selectively

___________________

⁵ Lockhart noted that the research he discussed was supported by grants from the National Science Foundation, National Institute for Occupational Safety and Health, and National Institutes for Health.

⁶ See https://www.ahrq.gov/patient-safety/settings/hospital/fall-prevention/toolkit/morse-fallscale.html (accessed September 13, 2024).

target interventions to specific weaknesses, he explained. Moreover, Lockhart stated these existing tools are somewhat qualitative and feature low sensitivity and specificity, which limit the usefulness of these assessments.

To address this gap, Lockhart and colleagues obtained a National Science Foundation grant to develop a technology to assess gait, posture, and musculoskeletal degradations. Using inertial measurement unit sensors, the wireless device is able to collect data on patients in their home environments and provide it to their physicians. Characterizing fall risk for older adults before a fall occurs is valuable, he said. Achieving this goal requires understanding the mechanisms that increase fall risk for older adults in comparison to younger individuals. Degradation of vision and the vestibular and proprioceptive systems occurs with aging, and he noted that no interventions are available to stop that process. Cognitive impairment and gait adaptation also occur with age. Lockhart emphasized the importance of understanding how these various factors integrate to influence falls in older adults.

Lockhart and colleagues have conducted perturbation experiments over the past 25 years to understand reactive behavior to various fall scenarios. For instance, bathtub testing has revealed that risk of falls is higher when stepping out of the bathtub than when stepping in, he said. Conducting research with younger and older participant groups yielded an understanding of gait adaptation that occurs with age. The gait of fall-prone older adults features changes in gait characteristics such as shorter step length, a wider base of support, slower acceleration, and slower transition of whole-body center of mass. These gait adaptations were thought to improve one’s stability; however, closer study revealed that these gait adaptations lead to increased fall risks in older adults (Lockhart et al, 2003). Furthermore, in experiments that posed slip or trip hazards to participants wearing safety harnesses, the initiation of the fall and the middle of the slipping or tripping action were very similar between young and older adults, he continued. The key difference between the groups was reactive recovery, which was 70–120 milliseconds faster in younger adults, enabling them to interrupt the trajectory of the fall and regain balance (Kim et al., 2005). Lockhart attributed this delayed reactive behavior to sensory degradation and musculoskeletal weaknesses, and he added that the initial condition of a participant—such as presence of fatigue—influenced the outcome of the fall.

Stability Assessment and Temporal Variabilities

Lockhart and colleagues modeled a person’s dynamic stability during walking. He explained that regardless of how advanced a measurement device is, it will yield simple qualities such as mean values, standard deviations, and variance. However, the gait cycle involves variability as a person

walks. System dynamics may not be measured at the point in time that yields the most helpful data or may only describe one instant in time, such as when the heel strikes the ground. Thus, a measure such as the standard deviation of heel contact does not capture all temporal information, and he maintained that complete time-series data are needed to more accurately assess fall risk. Therefore, the system Lockhart and colleagues developed uses nonlinear dynamics, the Lyapunov exponent, and Floquet multipliers to capture temporal information and illuminate variability that affects fall risk.

The gait-cycle data Lockhart and colleagues obtained indicate significant variability in the leg angle involved in the heel strike and less variation in the middle of the stride. Applying the Maximum Lyapunov exponent, a gait-stability metric, revealed much higher instability in older adults than in younger ones, he said. Lockhart also explained that there is age-related loss of biocomplexity in the structure of giant pyramidal Betz cells found in the motor cortex of the brain (Lipsitz and Goldberger, 1982). In a healthy, younger adult, the cell structure is complex and features many branches; in an older adult, it has far fewer branches. The branch complexity enables faster reactions to perturbations, he said, and therefore reduces the risk of fall accidents.

Fall Risk Prediction Tool

Understanding and assessing linear and nonlinear gait parameters fueled development of a wearable fall risk prediction system that uses inertial measurement unit sensors and a smartphone, said Lockhart. He and colleagues tested this system by modeling 171 community-dwelling older adults. On measures used to evaluate algorithms’ predictive ability, the best performing model (which included both linear and nonlinear dynamics) achieved areas under the curve (AUC) specificity and sensitivity, rates of approximately 90 percent when testing 10-meter gait (Doshi et al., 2023; Lockhart et al., 2021). Predictive accuracy of the model was further tested with 44 adults who were measured and followed for 6 months after testing. In that time, nine individuals experienced falls, and eight of these were predicted by the fall risk prediction system.

This system is a digital health platform called MyACTome, said Lockhart.⁷ The app measures gait speed, postural stability, and dynamic stability, and it generates a Lyapunov exponent. He recalled that less than a

___________________

⁷ More information about MyACTome is available at https://www.myactome.com/ and https://pitchbook.com/profiles/company/541550-71 (accessed May 31, 2024). Lockhart indicated that the iPhone application version of this tool is called the Lockhart Monitor and is available free of charge.

decade ago, obtaining such measurements required several hours; now these measurements can be generated by a smartphone in seconds. Lockhart’s group began collaborating with industry partners. After 2 years of validation, Healthcare Outcomes Performance Company acquired MyACTome,⁸ and this development is expected to foster clinical use of the tool. Lockhart stated that the company views the tool as a technology to promote safety and patient–provider communication. He continues to work with the company to improve the system. Additionally, he and colleagues are developing a skin patch to calculate the frequency of muscle fiber contractions. In older adults, type II A and B muscle fibers—the fast twitch muscle fiber required for reactive recovery—decrease, increasing proneness to falls. This patch will monitor muscle fiber frequency to inform fall risk assessment unobtrusively. Furthermore, Lockhart and colleagues are developing a physics-informed machine learning model, which will require far less data than that required by a data-driven machine learning system for humans.

DISCUSSION

Research on Nonimpact TBI Mechanisms

Christopher Loftus, chief medical officer at the U.S. Food and Drug Administration’s Division of Neurological and Physical Medicine Devices, and Uzma Samadani, founder of Oculogica and neurosurgeon at the Minneapolis Veterans Administration Medical Center, both asked about mechanisms other than direct head impact that may cause TBI, such as rapid chest deceleration from a seat belt or airbag and decreasing jugular venous return that elevate intracranial pressure, including whether research models have monitored intracranial pressure. Kent replied that he is unaware of research on transient increases in blood pressure and intracranial pressure in this context. Although other mechanisms for acquiring TBI exist, his prevention research—and the field’s as a whole—focuses on head impacts, said Kent.

He added that the models he uses do not model cerebral vasculature, and that even the most complex finite element models of the brain are only now approaching accurate boundary conditions. Emphasizing the difficulty in accurately modeling the brain stem and brain interactions with the skull, Kent stated that research is examining stress distributions around blood vessels in the brain, but understanding how blood pressure changes that increase intracranial pressure place mechanical strain on brain tissue and whether that leads to TBI is not likely in the near future.

___________________

⁸ See https://www.azpei.org/media-hub-articles/weartech-applied-research-center-leads-fall-risk-technology-to-major-acquisition (accessed September 16, 2024).

Repetitive Head Impacts

Noting that professional athletes are one of several population subgroups with an elevated risk of sustaining multiple TBIs, Loftus asked how the NFL is addressing concerns around chronic traumatic encephalopathy (CTE). Langton commented that her role with the NFL Innovation Platform is very different from that of the NFL chief medical officer, to whom she defers on matters related to CTE. Noting that in the past, NFL technicians spent 4 days calculating the number of head impacts in 1 day—an activity that Digital Athlete technology performs in less than 1 minute—she emphasized that the NFL is working to reduce TBIs by providing head impact data to coaches and players. The goal is to help teams better understand the number of head impacts each player has sustained since 2015, and to identify scenarios where the frequency of head impacts is high. Analysis of the player behaviors associated with frequent head impacts adds to an understanding of when head impacts result in a TBI, said Langton.

Prevention Equipment Implementation and Innovation Accessibility

Kristy Arbogast, R. Anderson Pew Distinguished Chair of Pediatrics at Children’s Hospital of Philadelphia, queried how preventative innovations for athletes, automobile passengers, and older adults can be widely disseminated to enable access for all.

Schnapp described adoption of helmets in professional sports such as the ski racer environment as one in which competitors understand the risks involved and always wear helmets in races, speculating that failure to wear a helmet is likely more common in the general population than among professional skiers; this may contribute to accidents with poorer outcomes. Although the adoption of helmets seems to have improved among the general population, she continued, not all helmets provide equal protection and more protective equipment may be more expensive, contributing to inequities in access to and use of highly protective ski helmets.

Lockhart agreed on the importance of innovation dissemination and accessibility to the general population. To this end, he referred back to his group’s creation of a digital health platform (the Lockhart Monitor, or MyACTome, see above) for detecting a suspected TBI. This free tool is a mobile phone application that assesses gait, posture, and balance. Arbogast asked about efforts to ensure that people are aware of the app and use it correctly. Noting that distribution is a challenge, Lockhart shared his hope that the company that acquired MyACTome will increase distribution rates.

Kent remarked that the implementation he focuses on does not target the person in the vehicle as the end user of the technology. Instead, he considers implementation in terms of enabling other innovators to use and

incorporate knowledge and technology into the design and development of their own products. For instance, the National Highway Traffic Safety Administration provides access to its crash test biomechanics database free of charge. The NFL Helmet Challenge funds innovators to advance helmet design and shares its findings with other industries and at forums such as this workshop, said Langton.

Safety Threshold Considerations

Leslie Wise, chief executive officer at EvidenceMatters, shared that her brother-in-law played in the NFL for 10 years during the 1970s. Three years ago, he died from CTE while in his late sixties. Her nephew was a Texas all-star Division I football player but quit the sport after witnessing his grandfather contend with CTE. Acknowledging that protective equipment has improved in the decades since her brother-in-law played professional football, she asked whether there would be a point where data on football risks could lead to a declaration that the sport is not safe. Wise drew an analogy to smoking, asking whether any type of smoking is considered safe and whether prevention can truly be created in football, given the repetitive impacts. She noted that long before players reach professional status, they are children playing on teams that lack the resources of the NFL.

Reflecting an earlier remark made by Schnapp that an important component of her recovery was abstaining from activity that would increase the likelihood of reinjury within the first year after her TBI, Wise pondered whether a guideline could state that once a child sustains a serious concussion, they should be discouraged from returning to sports where concussion risk is high, like football. In response to the mention of football, Langton reflected that recent years have seen efforts to make football safer as a contact sport, including improvements to protective equipment and research efforts to understand the play conditions where TBIs occur.

From here, Kent turned the dialogue to briefly comment on TBI prevention efforts in youth sports. He reflected that, broadly speaking, children’s mental and physical health benefit from participation in athletics, but some sports do involve a higher risk of concussion. Arbogast emphasized the importance of disseminating prevention knowledge and equipment throughout youth sports and described education efforts at various levels of sports as an ongoing challenge. Highlighting the increasing popularity of flag football and its inclusion in the 2028 Olympics, she commented that changes to how sports are played could potentially increase safety.

Safety Feature Considerations in Autonomous Vehicles

Given that federal motor vehicle safety standards do not prohibit drivers from using aftermarket protective devices, Max Sevareid, emergency

medical services specialist at the National Highway Traffic Safety Administration, asked whether countermeasures in motor vehicles beyond seat belts, such as helmets, should be considered to improve passenger safety in autonomous vehicles. Kent replied that people do not consistently comply with voluntary safety measures and that prevention mechanisms that protect people without requiring active engagement tend to be more effective. Thus, asking people to wear helmets inside cars is likely unrealistic. Furthermore, a helmet increases the space a person’s head occupies and may not necessarily increase safety if it increases the frequency of impact, he added.

Research efforts are currently directed at developing new safety systems. For instance, a model is being developed in which part of the seat pan would rise to grab hold of the ischial tuberosities, or bones of the pelvis that you sit on, while the lap belt pulls the pelvis downward. A mechanism that locks the pelvis in place holds potential in addressing the poor restraint performance of standard seat belts on a person in a reclined position. Kent emphasized that due to the strength of the pelvis, seat belts can apply large forces directed at the pelvis resulting in associated decreased movement of the head. He added that the automotive industry has the advantage of market forces driving safety innovation, in contrast with football. Helmet manufacturers have far smaller research budgets than those of Toyota or General Motors. Kent noted that collaborations between the sports and automotive industries may yield beneficial innovations.

Professional Football Research Efforts

In response to a question about what is included in the scope of the NFL’s research activity in TBI, Langton mentioned a few relevant initiatives. First, the NFL hosts helmet technology and performance innovation challenges.⁹ Second, the NFL supports a research portfolio that includes longitudinal studies on neuroscience and pain management. This work is supported by a scientific advisory board and recipients of NFL research support are listed on the NFL Player Health and Safety Portal.¹⁰ She further noted that the NFL and the NFL Players Association collaborate on crowdsourcing challenges and grant investing. Arbogast added that the NFL Players Association does fund some projects examining the long-term health of their player population. Further, Langton described how, in 2024, the NFL held its first Health and Safety summit to engage different disciplines within the NFL teams in education about the league’s research findings.

___________________

⁹ See https://www.nfl.com/playerhealthandsafety/equipment-and-innovation/innovationchallenges/ (accessed September 16, 2024).

¹⁰ See https://www.nfl.com/playerhealthandsafety/ and https://www.nfl.com/playerhealthandsafety/resources/ (accessed September 16, 2024).

Innovation Disparity Considerations

Corinne Peek-Asa, vice chancellor for research at University of California San Diego, remarked that a cataclysmic interaction between health disparities and technology disparities lies ahead, involving technology access issues and considerations regarding the computing power needed to sustain future innovations. Arbogast echoed this view, noting that epidemiological data clearly indicates disparities in TBI injury and outcomes, and unequal access to innovations such as those discussed during the session may further emphasize such disparities. Noting that the second session of the workshop mentioned the lack of a digital health benefit category in Medicare policy, rendering the program unable to cover digital health technologies, Peek-Asa closed by reflecting that dissemination of TBI innovations into widespread use remains a daunting but important challenge.