2

Implementing an Innovation in TBI Care: Lessons Learned from TBI Biomarker Development and Considerations for Patient Engagement

___________________

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 second session of the workshop outlined the development and implementation of traumatic brain injury (TBI) biomarkers and devices, explored patient engagement across stages of the research and development life cycle, examined expert feedback on the use of TBI assessment tools in the military, and discussed implementation considerations related to public and private payers. The objectives of the session were (1) providing concrete examples of the critical role of patient-centric approaches in the translation of research to practice, (2) highlighting the role of TBI survivors and lived experience advocates as leaders, advisors, and partners across all stages of clinical research, and (3) offering considerations based on community feedback to improve the military’s uptake of diagnostic devices cleared by the U.S. Food and Drug Administration (FDA). These included issues related to reimbursement, portability, and environmental fit. Cynthia Grossman, subject-matter expert in the science of patient engagement and real-world evidence, moderated the session.

BLOOD-BASED BIOMARKERS FOR TBI: CONCEPTUALIZATION TO IMPLEMENTATION AND BARRIERS TO WIDESPREAD ADOPTION

Linda Papa, emergency medicine physician and director of academic clinical research at Orlando Health Orlando Regional Medical Center, discussed the incorporation of TBI biomarkers into diagnostic assessment, the potential ramifications on clinical practice of having a working diagnostic blood test for mild TBI, and the implementation and associated challenges of blood biomarker devices. There has been extensive interest in identifying blood-based and other types of biomarkers to inform assessment, classification, anticipated recovery trajectory, monitoring of treatment response and recovery from TBI, and to inform research efforts and clinical trials design (NASEM, 2023), with many contributors to this history. Papa’s talk highlighted a subset of these developments, with an emphasis on two protein markers that have been developed into commercially available devices (see also sources such as Kobeissy et al., 2024). She described TBI as ubiquitous since it affects all age groups in every part of the world. In the United States, TBI is diagnosed in emergency department (ED) visits for almost 3 million people, including almost 1 million children each year, with approximately 90 percent of these injuries classified as so-called mild TBI (that is, not requiring an extended stay in the hospital for critical care) (Korley et al., 2016; CDC, 2019). These numbers are likely an undercount, as instances of concussion (a form of mild TBI) or other TBI injuries may never be assessed in the ED.

TBI Clinical Decision Rules

Papa recalled working as an emergency medicine resident during the 1990s. At that time, and in the absence of TBI assessment guidelines, the ED managed assessment of mild TBI by determining the Glasgow Coma Scale (GCS) score,² conducting a neurological examination, and potentially performing a computed tomography (CT) scan. Using measures of responsiveness and verbal and motor functioning, GCS classifies TBIs as mild, moderate, or severe. She remarked that although GCS is a helpful triage tool, it has limited diagnostic usefulness. GCS score categories are imprecise—particularly in cases where a patient is intoxicated, has been given sedatives or neuromuscular blockers, or is experiencing seizures, distracting pain, or hypoxia—and they do not provide the information needed to accurately determine the severity of injury, Papa explained.

___________________

² The Glasgow Coma Scale measures impairments to consciousness and responsiveness after brain injury. On a scale of 0 to 15, higher values are indicative of less severe injury and lower scores are indicative of more severe injury.

In the early 2000s, two guidelines on CT scans for mild TBI were published. The New Orleans Criteria (NOC) and the Canadian CT Head Rule (CCHR) (Haydel et al., 2000; Stiell et al., 2001). The NOC recommended a CT scan when the patient presented with:

1. Headache,
2. Vomiting,
3. More than 60 years of age,
4. Drug or alcohol intoxication,
5. Anterograde amnesia,
6. Visible head or neck trauma, and
7. Seizures.

Meanwhile, the CCHR outlined a list of similar criteria, including:

1. GCS less than 15 after 2 hours,
2. Suspected open or depressed skull fracture,
3. Suspected basal skull fracture,
4. Vomiting two or more times,
5. Age 65 or older,
6. Severe mechanism of injury, and
7. Retrograde amnesia.

Both guidelines correctly identified mild TBI 97–100 percent of the time. Notably, the NOC elicited fewer false positives (10–20 percent) than the CCHR (30–50 percent), noted Papa.

The availability of these guidelines advanced management of mild TBI, but challenges remained. Firstly, interpretation of the clinical decision rules varied, said Papa. Secondly, the studies that informed the guidelines excluded certain patient populations, such as patients who had a bleeding disorder or were anticoagulated, had experienced syncope or seizures before the injury, had major trauma, had repeated head trauma, had TBI without loss of consciousness or amnesia, or were pregnant.

While serving as an emergency medicine attending physician in the early 2000s, Papa observed how blood biomarkers were commonly used to diagnose cardiac, kidney, liver, thyroid, pancreas, and prostate issues. However, an analogous blood biomarker test was not available for the brain. Papa and a group of scientists and clinicians were inspired to create a handheld device that, similar to a glucometer, would accept a patient’s blood sample via an inserted cartridge and test for the presence of elevated TBI markers.

The Search for TBI Blood Biomarkers

Using a neuroanatomic approach in a rodent model of TBI, the team searched for TBI biomarkers from different parts of the neuron and supporting structures (Papa, 2012). After brain injury, levels of the protein ubiquitin C-terminal hydrolaseL1 (UCH-L1) showed correspondence to brain injury severity. With a promising biomarker identified in an animal model, the focus shifted to clinical studies in human subjects. Papa described that the studies operated on a limited budget owing to difficulty in generating interest in TBI biomarkers. The studies initially focused on severe TBI, given the availability of human cerebrospinal fluid (CSF) bio-samples associated with this level of injury.³

Papa and colleagues wrote an institutional review board protocol to collect these CSF samples to search for biomarkers. As was the case in the TBI rodent model, they found that UCH-L1 was also present in human CSF samples following TBI. Moreover, in humans, UCH-L1 levels rose quickly within the first 6 hours after injury and gradually decreased over 7 days (Papa et al., 2010). Patients with increased intracranial pressure, hypoxia, or hypotension postinjury also demonstrated increased levels of UCH-L1; thus, Papa explained, the biomarker serves as an indicator of these postinjury complications.

Given the impracticality of collecting CSF samples from patients with mild and moderate TBIs, Papa and colleagues searched for a more accessible biofluid. In comparing human serum and CSF, they found a correlation of UCH-L1 levels in these two fluids. Identifying a TBI biomarker in serum enabled researchers to look for UCH-L1 in milder injuries. During this process, Papa and colleagues discovered another biomarker and histological stain, glial fibrillary acidic protein (GFAP), coming from the supporting structures of the neuron. Papa collected serum samples from patients in the emergency department with mild TBI in the early to mid 2000s. While exploring the data from these studies in 2006, Papa detected a high correlation between GFAP, the severity of injury, and the presence of lesions on CT scans and realized she had discovered another biomarker for mild TBI (Papa et al., 2008).

This discovery generated excitement from funding agencies and led to a more definitive, multicenter study of the UCH-L1 and GFAP biomarkers in patients with mild and moderate TBI. The study also enrolled patients with orthopedic injuries and patients who had experienced motor vehicle crashes without sustaining head injury. These groups served as the control groups, enabling Papa and colleagues to ensure that the biomarkers indicated TBI

___________________

³ When a person has a severe TBI, a clinical intervention to reduce intracranial pressure from bleeding and inflammation includes draining cerebral spinal fluid (CSF) from the ventricles.

rather than broader injuries. Community volunteers with no injuries also provided samples. Researchers found a graded response in both biomarkers, with patients with orthopedic injuries or motor vehicle crashes and no TBI showing slightly elevated levels relative to those of non-injured individuals; TBI patients also showed substantial increases over control groups, with levels reflecting the severity of brain injury. Moreover, levels of both UCHL1 and GFAP were significantly higher in patients with anatomic lesions visualized on CT scans than those with no lesions. Papa described both biomarkers as having excellent sensitivity and very good specificity—with GFAP demonstrating higher specificity than UCH-L1—and as being predictive of the need for neurosurgical intervention.

TBI Biomarker Studies

The discovery that GFAP and UCH-L1, both originally histologic stains, are TBI biomarkers occasioned excitement and motivated additional research in the field, said Papa. Differing biomarker studies reported obtaining samples at variable time points from injury, varying from hours to days after injury. This variability led Papa and her team to explore the time course and functionality of biomarkers systematically at defined times after injury. They collected blood samples at 20 distinct time points from mild TBI and trauma control patients over a 7-day period, with the first sample drawn within 4 hours of injury (Papa et al., 2016). In some cases, samples were drawn as early as 15 minutes postinjury.

The findings revealed that UCH-L1 levels increase rapidly after injury and quickly begin to decrease within the first few hours postinjury, making the biomarker a helpful indicator of TBI during the first 24–48 hours after injury (see Figure 2-1). In contrast, GFAP requires a few hours postinjury to elevate substantially, reaches its peak at approximately 20 hours, and decreases more gradually than does UCH-L1. She noted that even at 7 days postinjury, GFAP levels are discernable in patients with anatomic lesions on CT versus those without, with concentrations varying by approximately 2,000 picograms between the two groups. Furthermore, GFAP levels indicate whether mild TBI is present a week postinjury. Papa highlighted the value of determining whether a patient who enters the ED with a trauma has sustained a TBI. She emphasized that GFAP is particularly effective at indicating whether a patient has a mild TBI and whether a TBI features lesions.

Papa and colleagues also studied UCH-L1 and GFAP levels in patients with subconcussive injuries—i.e., patients who hit their heads but do not have symptoms of TBI—and in patients with orthopedic injuries at 20 points in time over a 7-day postinjury period (Papa et al., 2019). They found that GFAP levels remain close to zero in patients with traumatic

orthopedic injuries, whereas patients with subconcussive injuries and no TBI symptoms demonstrate a slight elevation of up to approximately 100 picograms in GFAP. Researchers compared serum GFAP levels in trauma controls with no concussion, patients with head trauma without TBI symptoms, patients with head trauma with TBI symptoms, and patients with intracranial lesions. They found a graded response, with increasing GFAP levels corresponding with the severity of head trauma. Papa emphasized that the biomarkers yield substantial information within 30 minutes of injury, with 98 percent sensitivity.

TBI Biomarker Technology

In 2018, the Brain Trauma Indicator was the first TBI biomarker test to gain FDA clearance to aid in the evaluation of TBI, said Papa. A milestone in TBI assessment, this UCH-L1 and GFAP blood test can serve as an aid to indicate CT lesions within 12 hours of experiencing mild to moderate

TBI. However, testing required specialized equipment and 3–4 hours to run assays, making the Brain Trauma Indicator impractical in many settings. In 2021, FDA granted clearance to the i-STAT Alinity TBI plasma test, the first rapid handheld point-of-care blood test of UCH-L1 and GFAP.⁴ It could produce a result in less than an hour. Papa explained that the i-STAT Alinity test indicates CT lesions within 12 hours of mild TBI and provides qualitative and quantitative GFAP and UCH-L1 levels.

In preparing to implement use of i-STAT Alinity blood testing, Papa and colleagues explored how clinical decision rules and biomarker testing correlate via a prospective study (Papa et al., 2022). Real-time clinical assessment of the clinical decision rules required physicians to identify clinical signs and symptoms of TBI and complete forms on all patients with suspected mild TBI. Blood samples were then tested to compare results of the assessment methods. Both the CCHR and NOC demonstrated 100 percent sensitivity, with 33 percent and 16 percent specificity, respectively. She explained that the biomarkers and clinical decision rules work synergistically, and together they achieved 100 percent sensitivity and 46 percent specificity. She emphasized that, in her opinion, this combination could reduce the number of costly diagnostic CT scans performed on patients with suspected TBI. Papa noted that Orlando Health is currently working to implement i-STAT Alinity testing into ED management of mild TBI patients. The updated assessment protocol will begin with a GCS score, a neurological examination, and clinical decision rules and then progress to blood testing and, if necessary, a CT scan.

Comparing a timeline of cardiac biomarkers for acute coronary syndrome to that of TBI biomarkers, Papa stated that TBI biomarkers are still in the early phase of clinical implementation. Although troponin was developed in 1990 as a cardiac biomarker, clinical implementation required 5–10 years before achieving frequent use (Garg et al., 2017). Biomarkers for TBI have been identified for approximately 20 years, but FDA approval for GFAP and UCH-L1 did not occur until 2018. The i-STAT Alinity device gained approval in 2021, and on April 1, 2024, FDA approved the first TBI biomarker device for whole blood. She underscored that this development enables TBI severity to be established at a patient’s bedside by placing a blood sample from a finger stick directly into a blood test device and obtaining results within minutes. She noted that currently, this assay is indicated to positively detect the subset of TBIs involving an anatomic lesion to the brain (as opposed to TBI from hypoxia and brain bleeds, but in the absence of physical tissue damage). In coming years, additional assays for additional TBI indications are likely to be developed, she continued.

___________________

⁴ See https://abbott.mediaroom.com/2021-01-11-Abbott-Receives-FDA-510-k-Clearance-for-the-First-Rapid-Handheld-Blood-Test-for-Concussions (accessed September 13, 2024).

Implementation Challenges and Considerations

The implementation of TBI biomarker testing to aid in evaluation and diagnosis of TBI in the Orlando Health ED has involved numerous challenges, Papa stated. For instance, changing decision-making patterns is difficult and integrating biomarker testing into the electronic health record (EHR) requires time. Laboratory implementation was particularly time-consuming, given the regulations involved. Turnaround time for test results and using the i-STAT Alinity for the correct indications have also posed challenges in implementation. Papa noted that some physicians, excited about a new device, experiment with using it in situations outside of its indications. Reimbursement and patient cost issues have been at play in the absence of a current procedural terminology (CPT) code. However, despite these challenges, emergency physicians at Orlando Health have been amenable to implementation efforts, she stated, indicating that 86 percent of surveyed emergency physicians were in favor of the TBI blood test (Papa et al., 2022). Papa added that during the implementation phase, TBI biomarker testing has demonstrated 100 percent sensitivity in detecting CT lesions in clinical use.

Moving forward, several implementation considerations could affect the use of diagnostic TBI biomarker testing, said Papa. Physician buy-in will require education efforts and real-world scientific evidence to increase awareness and confidence in the performance of the test. Additionally, clinicians will be faced with interpreting quantitative values and improving workflow to integrate the test into current practice. Papa noted that biomarker testing should be incorporated into practice guidelines. Endorsement from medical societies and involvement from third-party payers will be needed to advance implementation. Considerations regarding cost to patient, reimbursement, and CPT codes will need to be addressed, as well as any medicolegal issues. She pointed out that TBI biomarker testing is less expensive than CT scanning and requires fewer resources. Papa emphasized the importance of patient feedback, as improved patient care is the purpose of diagnostic advances. Approximately 30 percent of patients discharged from the ED with mild TBI and a clear CT scan continue to have substantial symptoms and TBI-related problems moving forward. Papa remarked that the progress in identifying diagnostic TBI biomarkers and advancing biomarker testing as an aid to evaluation and diagnosis is the beginning of a future that enables better TBI treatments and outcomes.

PATIENT-CENTERED RESEARCH DESIGN FOR HEALTH CARE INNOVATION

Suzanne Schrandt, founder and chief executive officer at ExPPect, explored two types of patient engagement and the benefits of partnering

with patients throughout the medical therapy development process. She described how ExPPect works to embed the lived experience and wisdom of patients and families into both clinical research and innovation and health care.

Integrating Patient Engagement in the Innovation Pathway

A development process that waits until a product is complete to consider implementation has failed to capitalize on opportunities, said Schrandt. Rather than consider implementation as a point in time, it should be an ever-present consideration throughout the innovation process, she added. Moreover, implementation extends beyond patient and family engagement, involving the engagement of frontline health care workers, clinicians, and all the people who interact with an innovation throughout the clinical workflow. Lack of familiarity is a common downstream implementation challenge, given the human inclination to resist change. Patients who are doing relatively well on a current therapy will likely require substantial motivation to try a new therapy even if it offers promise. By understanding this dynamic, innovators can address this downstream challenge by partnering with patients, families, and other parties during the upstream stages of innovation to identify concerns and needs. Schrandt remarked that ExPPect is guided by a belief that upstream engagement can never begin too early.

Outlining the stages of the innovation pathway, Schrandt highlighted questions to address through patient engagement. The first step of innovation is identifying an unmet need, and developer questions at this stage include whether the unmet need is important to patients, whether patients are aware of this importance, whether an educational or awareness gap is present, and whether the innovation will target the correct symptoms or advance diagnosis in a desired and actionable way. In the next innovation phase, ideation, considerations include whether patients are likely to be willing to use the new process or product, whether its function will align with patient needs, and whether any side effects or benefit-risk trade-offs are worth the gain that patients will experience. The protocol, study, or clinical trial phase should consider whether the design is feasible and attractive to patients and is likely to accrue and retain them, she noted.

Additionally, design should incorporate patient feedback to ensure that the study outputs will answer the questions patients are likely to ask once the therapy becomes available. Schrandt provided an example from the rheumatology disease space, in which new therapies or diagnostic tools have been developed without early patient engagement. Thus, upon introducing an innovation to patients, developers were often unable to answer questions such as “Will it manage my fatigue?” because they did not build fatigue as an endpoint in the study.

Engaging patients in the earliest aspects of innovation improves the ability of researchers to answer questions that are important to patients, Schrandt emphasized. Similarly, the regulatory review and approval phase benefits from consideration of the final phase—implementation and use—from the outset. Schrandt described implementation with a baseball analogy, likening the innovation to a skilled pitcher and implementation to the catcher ready to receive the pitch. Early patient engagement enables the catcher to be prepared for the various types of pitches—questions, concerns, and potential barriers—that an innovation may generate.

Types of Patient Engagement

Schrandt described an encouraging explosion of patient engagement during the past 2 decades, but she noted that although two distinct types of patient engagement exist, the term patient engagement is often applied to only one of them. She stated that most people use the term patient engagement in reference to awareness, education, and peer-support efforts that enable patients to engage in shared decision-making with their providers to coproduce health outcomes. Although critically important, this personal engagement is not a focus at ExPPect, said Schrandt. Rather, she and colleagues focus on system-level engagement in which patients apply their lived experiences to the improvement of systems or organizational processes. For example, lived experience can improve biotechnical development, patient safety, regulation, legislation, advocacy, or other aspects related to diagnosis, care, and research. The two types of patient engagement rely on different tools and skill sets, yet they are related to each other. Patients skilled at personal engagement have high patient activation scores, are able to self-advocate, and are often high-functioning given their disease state. Schrandt remarked that the best way to achieve such patient outcomes is via system-level engagement; thus, distinct types of patient engagement foster one another.

Return on Engagement

Drawing a parallel between the automobile and health care industries, Schrandt recounted the 1958 Ford Edsel as an example of failure to incorporate user input. At the time, market trends favored cars that were more streamlined, affordable, and fuel-efficient. Despite consumer desires, the Ford Edsel was large, had poor gas mileage, and was at a higher price point than some models. Discontinued after only 2 years, the Edsel line cost the Ford company $350 million, a huge sum at the time (Brooks, 1969). Schrandt noted that health care consumers are different than those of the auto industry, but the example highlights how disregarding user input in development can lead to product failure.

She then provided a health care example from several years ago, in which a company determined to develop an inhaled insulin formulation to eliminate the need for administration by injection. The developers did not involve patients in the innovation pathway. Schrandt stated that although injections can be cumbersome and painful, many patients had become adept at them and were able to self-administer discreetly. In contrast, the inhaled insulin device was large, awkward to use, and more difficult to use discreetly. Patients expressed concern at not having the ability to discern the exact amount of insulin received via inhalation. Despite regulatory approval, the product ultimately failed to achieve uptake in the market, and the company lost $2.8 billion in the venture. Schrandt stated that this loss could have been avoided at multiple points along the innovation pathway.

Not only does patient engagement lead to products that meet patient needs, it is also financially advantageous, said Schrandt. In calculating the “return on engagement,” researchers found that every $100,000 invested in patient engagement—incurred by costs such as those associated with creating and facilitating patient advisory panels and paying patients for their time—yielded a 500-fold return on investment (Levitan et al., 2018). Cost savings are accrued in avoiding protocol amendments, retaining study participants, and improving patient experience. She noted that while this figure pertains to the drug development space, it demonstrates value that is relevant across innovation in areas including clinical practice workflow, diagnostic safety, and diagnostic quality. For example, she worked on a project focused on racial and ethnic disparities in inflammatory bowel disease (IBD) diagnostic delay.

A survey of approximately 1,000 patients with IBD revealed breakdowns in the diagnostic pathways. The concept of normality featured in numerous responses. For instance, patients with juvenile onset IBD had experienced symptoms since early childhood and therefore did not know they were abnormal. Some respondents noted that when asked by clinicians whether there were any changes, they replied “no” because their symptoms remained the same. Schrandt and colleagues used these data to create learning modules for frontline clinicians on how to engage in conversation with patients about the range of normal and what patients are experiencing to improve diagnosis.

Tools and Methodology

Schrandt stated that in addition to using patient feedback to design studies that are more comfortable and convenient for participants, patient engagement has helped data scientists to better analyze data. In one example, data scientists were confused by claims data indicating that patients were frequently stopping and starting rheumatoid arthritis therapy and switching from one therapy to another. However, patients with rheumatoid

arthritis who were shown these data explained their lived experience of having to change treatment plans in response to infections, surgeries, side effects, insurance barriers, or other issues. In the absence of patient engagement, data scientists would have missed an opportunity to understand the barriers to treatment that patients experience.

Schrandt noted that engaging patients in data analysis entails efforts and methodology, but the payoff can be substantial. Repositories of methods and tools for this purpose are available, she said. Moreover, domestic and international coalitions and organizations are dedicated to medical product development in the preventative, diagnostic, and treatment spaces. Examples include the Medical Device Innovation Consortium, Patient-Focused Medicines Development and Patient Engagement Synapse, and the Patient-Centered Outcomes Research Institute (PCORI). Schrandt highlighted that PCORI’s research funding centers on meaningful and authentic patient engagement. It has developed fundamental expectations for research partnerships, a database of engagement in research literature, and the PCORI Engagement Tool Resource Repository, which includes guides, frameworks, and other resources.⁵

Schrandt emphasized that patient engagement is not a symbolic exercise; its purpose is the betterment of science. To this end, FDA has been developing guidelines for incorporating patient engagement in medical product development (FDA, 2024). She stated that patient engagement can be daunting for researchers who are unfamiliar with the processes and methodologies involved. This unfamiliarity—coupled with budgetary and time constraints—sometimes motivates introductory, surface-level efforts. Schrandt cautioned against such superficial engagement efforts, remarking that they yield unsatisfactory results. In contrast, comprehensive patient engagement benefits all steps of any medical product, she asserted, including the most technical steps of the innovation pathway. Myriad resources are available, she added, and patients and patient groups are eager to partner in improving the systems upon which they rely. Schrandt reiterated that engagement is not conducted for the sake of engagement, but in the pursuit of better science.

CONSIDERATIONS REGARDING UNDERUSE OF APPROVED TBI DIAGNOSTIC DEVICES IN MILITARY ENVIRONMENTS

Kathy Lee, director of the Warfighter Brain Health Initiative at the U.S. Department of Defense (DoD), reviewed feedback on adoption and useful-

___________________

⁵ The PCORI Engagement Tool Resource Repository is available at https://www.pcori.org/engagement/engagement-resources/Engagement-Tool-Resource-Repository (accessed May 24, 2024).

ness within the military environment of several types of devices cleared by FDA as aids in TBI assessment. She emphasized that her presentation is not an official DoD position but rather an informal application of subject-matter expertise to explore implementation and barriers to use of TBI diagnostic devices. On a quarterly basis, DoD updates its TBI Advisory Committee, comprising more than 100 members, on FDA-cleared devices. Currently, eight devices have been cleared, including neurocognitive assessment tools, eye-tracking devices, biomarker measuring devices, and brain monitoring and measuring devices.

Neurocognitive Assessment Tools

Lee reported that three neurocognitive tools have been cleared by FDA for the assessment of head injury: Automated Neuropsychological Assessment Metrics, Immediate Post-Concussion Assessment and Cognitive Testing (ImPACT), and Defense Automated Neuropsychological Assessment⁶. Subject-matter experts within DoD, including consultants to the surgeon general, provided informal feedback on these cognitive monitoring tools. Experts expressed concerns that information from these tools could be misinterpreted or misused by providers, who might rely solely on the test to diagnose TBI, Lee said. Instead, she said, data from these assessment aids should inform a larger clinical picture, and feedback indicated that better guidance is needed on integrating these tools into patient care.

Lee noted that lack of interoperability between new tools and existing information technology systems poses challenges to accessibility and receiving and incorporating results. Experts noted that the test should be mobile and be capable of storing baseline data for analysis. DoD has a longstanding predeployment neurocognitive testing program, and clinicians need to access these data at the time of injury to enable comparison. However, retrieving these results is currently a cumbersome, multistep process that limits usefulness and would be improved via an accessible, application-based tool, added Lee.

Eye-Tracking Devices

Use of a TBI assessment tool in isolation lacks integration of a broader clinical perspective and appears to pose a barrier to use, Lee noted. The feedback solicited about monitoring tools reflected skepticism about the

___________________

⁶ See https://www.army.mil/article/202517/pre_deployment_screening_establishes_baseline_to_fight_tbi; https://impactconcussion.com/ and https://mrdc.health.mil/index.cfm/media/articles/2014/FDA_clears_mobile_app_to_diagnose_head_injuries (both accessed September 13, 2024).

value of isolated quantitative TBI assessment measures, such as eye tracking. Experts also expressed concerns about investing in technologies that do not meaningfully contribute to TBI management with actionable data or those that fail to outperform existing clinical methods. In the example of eye tracking, Lee highlighted the identified need for manufacturers to make the intended use clear for end users. Additionally, experts expressed the need for better quality evidence derived from larger, unbiased studies (i.e., independent studies not funded by the manufacturer) in specific settings to determine the context and usefulness of devices. For example, even if a device is not validated as a clinical diagnostic, it may yet serve as a useful informational tool as part of the broader picture of patient functional evaluation and care.

Feedback emphasized the need for collaboration with vision and performance communities to assess eye-tracking devices from the perspectives of these disciplines. She added that the absence of a gold standard for eye-tracking metrics poses challenges in determining whether a device captures the correct indices. Furthermore, devices may provide impressions or reports that require a specialist to interpret, and lack of easily understandable impressions creates a barrier to adoption, said Lee.

Biomarker Measuring Devices

Feedback from DoD experts reflected an overall view that blood biomarker measuring devices are useful in determining the need for CT scans based on negative predictive value for intracranial bleeds, Lee stated. Additionally, experts acknowledged that these devices hold value for battlefield care and in decisions related to scenarios involving prolonged field care and lack of air superiority. TBI assessment in combat areas sometimes involves transporting patients on dangerous convoys to have CT scans performed; she noted that more than 740 of these CT scans—and the risky travel by convoy they entailed—could have been avoided if this type of technology and capability had been available. Furthermore, the potential for a future fight with adversaries sometimes prohibits air travel, necessitating prolonged field care. Therefore, strategic planning examines methods of housing patients on the battlefield and providing the best care possible in those circumstances.

Biomarker testing that requires phlebotomy to draw the blood sample as well as a centrifuge to “spin down” the blood sample involves substantial resources in an austere military environment. Lee said that although TBI biomarker testing could replace CT scans in some cases, the unfamiliarity of the new devices gives rise to concerns of missing a diagnosis and the potential legal issues involved. Confidence in alternatives to CT scans will require time and training, she added. In January 2024, the National Institute of

Neurological Disorders and Stroke held a workshop on TBI classification, and Lee expressed optimism that outcomes from that meeting will advance the use of biomarkers by codifying clinical symptoms, biomarkers, imaging, and modifiers associated with the TBI-mild complex. Moreover, the DoD Joint Trauma System developed a clinical practice guideline that incorporates blood biomarker values for downrange use. Lee predicted that the ability to test whole blood samples for TBI biomarkers—thereby eliminating the need to travel to a centrifuge—may be transformative.

Brain Monitoring and Measuring Devices

Two brain-monitoring and measuring devices, BrainScope and Infra-Scan, have FDA clearance for aid in TBI assessment, said Lee, and DoD also sought feedback on the extent to which these devices were being used. She highlighted unique issues the military faces. Military medical providers contend with space and weight limitations and must carry equipment to far-forward environments; as a result, they limit devices to those that yield high value in assessment, care, and treatment. Brain-monitoring headsets can be impractical in the prehospital or far-forward environment, Lee stated. As was the case for eye-tracking devices, expert feedback underscored the need for better quality evidence derived from larger studies in specific settings. She noted that data on devices are often generated by industry-funded studies and that determining appropriate settings for using the device can be difficult, given the variety of settings to which troops are deployed and the variance in requirements of these settings.

Better understanding and clarifying how a device is most useful—as part of assessment, diagnosis, treatment, recovery, or prognostication, for example—and for which populations and in which settings is important. Trying to use a device for multiple purposes across care settings simultaneously can backfire, as people may generalize device failure for one type of use or for one setting and assume it will not work in others, Lee remarked. Determining the most appropriate settings for device use and communicating with those targeted end users is important. Additionally, DoD has medical assets across battlefields and fixed facilities, and tools must fit the military requirements for use.

High-Level Considerations to Maximize Use

Maximizing device use within military environments necessitates clear requirements, said Lee, noting that DoD explicitly states its requirements for the studies and research activities it funds. Needs assessments can aid in identifying tools and techniques necessary for unique populations, various environments of care, and stages of care (e.g., assessment, diagnosis, treatment, recovery, prognostication). The battlefield population has different

needs than the general population, and a clinic environment requires tools that vary from those needed in an ED setting. Moreover, some devices claim utility across stages of care, but this rarely proves to be the case, Lee remarked. The process of codifying need for a device—in terms of the type of care environment and the phase of the care continuum in which it will be used and its target population—elucidates requirements; innovators that meet these requirements during development will address many identified barriers, she added. Furthermore, Lee stated that maximizing usefulness will necessitate the creation of a centralized process for submitting devices into care, a process that should include evaluation of suitability and field testing and prioritize TBI devices.

PAYER CONSIDERATIONS

Shari M. Ling, deputy chief medical officer in the Center for Clinical Standards and Quality at the U.S. Centers for Medicare & Medicaid Services (CMS), and Leslie Wise, chief executive officer at EvidenceMatters, outlined Medicare and commercial coverage considerations for diagnostic tools and devices.

Medicare Considerations

Ling explained that CMS, the largest payer for health care services globally, covers approximately one in three Americans via two distinct programs, Medicare and Medicaid. Medicare provides coverage to individuals who are aged 65 years or older, who have a disability, or who have end-stage kidney disease. She stated that evidence generation (aka research) considers the population that treatments and services are targeting as well as the coverage, coding, and payment processes required by the payer that serves the population of interest. The Social Security Act delineates Medicare coverage requirements and processes, specifying (1) the items and services covered, (2) who is eligible to receive payments for items and services, and (3) the vehicles for payment delivery. Specifically, Medicare coverage is limited to items and services that are reasonable and necessary for the care of qualifying Medicare beneficiaries and have a Medicare benefit category.

Ling remarked that CMS coverage benefit requirements focus on the recipients of care and services as well as the context in which services are delivered to ensure safety, effectiveness, and clinical utility. She underscored that clinical utility is grounded in the factors that matter to the people receiving care and the goals of care that can be assessed by tracking health outcomes that are meaningful to the people experiencing them. Thus, Medicare coverage requires that supporting evidence must reflect the health concerns of the population being served. Ling noted that evidence may be sufficient to meet FDA requirements but not include the clinical context that

defines reasonable and necessary for Medicare coverage purposes. Therefore, these contextual aspects need to be integrated into research designs and methods, she said, so as not to fall short of the evidence needed to subsequently meet coverage purposes. This is increasingly important given the complex health conditions of the people covered by Medicare.

Federal law also establishes expectations and limitations for Medicare coverage of diagnostic tests. Diagnostic tests should yield information that is used in the management of Medicare beneficiaries, which should translate into improved health outcomes, she explained. The evidence is reviewed to understand the usefulness of diagnostic tests as applied to people with Medicare, said Ling, adding that the management of an individual condition is important, but health outcomes are often affected by other diseases that are present. Ling emphasized that the research and evidence supporting a Medicare coverage policy must apply to the beneficiary population or sub-populations for whom the treatment or service is intended. She added that Medicaid also values and depends on evidence generated, but authorities are distributive, and coverage and payment decisions are at the discretion of each state, within broad federal guidelines.

Ling stated that in navigating complex care systems, attention should be given to care processes, workflow, and the context in which services are provided. This awareness is important in delineating if observed impediments to access or service delivery are specific to the management of people with TBI or are also faced by the general patient population. It is important to methodically examine the workflow of TBI care at multiple levels, from the policies to the interpretation of those policies by health systems, down even further to the level of individual decisions made by clinicians and providers in those systems, said Ling. This yields insights into system effectiveness and efficiency in providing treatments and services to TBI patients.

Commercial Payer Considerations

Wise spoke from her experience in commercialization, market access, and reimbursement policy and echoed the importance of innovators’ understanding who the payer will be for the solution being developed and what evidence that payer will require. Given that Medicare requires evidence on patients who are at least 65 years old, disabled, or on dialysis, omission of such patients from a study will make it difficult to obtain Medicare coverage for a drug, device, or diagnostic tool. Most commercial payers want strong, clear clinical outcomes with little homogeneity in the studies, she said; thus, research on diagnostic tools should demonstrate clinical usefulness. In cases where a test does not add additional clinical utility to patient care, most commercial payers will not be willing to cover it. For instance, if clinicians continue to perform both CT scans and biomarker testing, commercial payers might be unwilling to cover both. Wise noted that within

the next decade, the establishment of a large body of data may demonstrate strengthened clinical utility from such a test, at which point commercial payers would become more likely to provide coverage.

Among all payers, commercial insurance requires the most rigorous research in providing coverage, she stated. Prior to health care reform, most commercial payers covered devices upon FDA clearance. However, once health care reform required commercial payers to cover all patients, companies increased limitations on coverage. Therefore, Wise explained, commercial coverage involves a high barrier to entry that requires well-designed studies indicating clinical utility and effects of diagnostics on patient outcomes.

Wise highlighted that the term mild traumatic brain injury sounds contradictory and does not have a diagnosis code. She encouraged the alignment of language to approved diagnosis codes and testing that reflects the diagnosis codes addressed. For instance, a variety of possible diagnosis codes include concussion with or without losing consciousness, loss of consciousness lasting for different lengths of time, and other TBI codes. Wise noted that most professionals evaluating evidence for payers do not have medical or doctorate degrees and are trained in the GRADE system to evaluate evidence, which does not include the nuances of any condition. Using language that is accessible to payers and aligned with their understanding of diseases and diagnostics could facilitate coverage.

Wise stated that payer coverage is of paramount importance to implementation. She emphasized that FDA criteria need to be considered throughout the research and development process, but these criteria pertain to safety and effectiveness, with effectiveness indicating that a technology does what it was designed to do. Effectiveness in this sense does not necessarily indicate improved clinical outcomes or subsequent payer coverage of the innovation. Thus, evidence of clinical utility should also be considered from the early stages of development. Furthermore, obtaining a CPT code involves a multiyear process and multiple studies. Wise commented that she has witnessed companies go out of business before obtaining FDA clearance, but that factoring considerations of safety, effectiveness, and clinical utility into study design can help minimize such situations.

DISCUSSION

Biomarker Implementation Status

Ramon Diaz-Arrastia, professor at the University of Pennsylvania, asked about the current state of TBI biomarker testing at Orlando Health Orlando Regional Medical Center and about the likelihood that GFAP and UCH-L1 tests would be performed while treating a person after a car crash. Papa replied that Orlando Health is on the cusp of TBI biomarker testing being common

clinical practice. Currently in the process of switching from plasma cartridges to whole blood cartridges, Orlando Health has made necessary changes to the EHR and has conducted organizational activities to facilitate the new equipment. The Orlando Health care system is prepared to incorporate whole blood cartridges into clinical practice upon equipment arrival, she said.

TBI Biomarker Testing in Children

Darío Villarreal, head of science and technology at Toyota Way Forward Fund, noted that Papa’s presentation highlighted that approximately 1 million children visit the ED for TBI each year. He asked whether children have been involved in biomarker studies and, if not, what the potential ramifications of their absence from studies might be. Papa replied that pediatric biomarker studies have been conducted on children (aged 0–18 years), and the GFAP and UCH-L1 biomarkers can be used in children as well as adults. She added that a multicenter pediatric trial is currently underway using both biomarkers, with the results expected over the next 2 years.

Medicare Patient Engagement

Noting that clinical trials are often not patient-driven, Corinne Peek-Asa, vice chancellor for research at the University of California San Diego, asked for clarification on how patient experience drives Medicare discovery and coverage. Wise responded that CMS is a statutory body, as is FDA, and the Medicare standard is that a technology is reasonable and necessary and fits into a benefit category. For instance, if a diagnostic is software only administered in the home, Medicare cannot cover it—despite whether patients like it or whether it is FDA-approved—because no benefit category exists for digital health. She noted that Medicare has an interest in establishing such a category, but this requires policy change from Congress. The Medicare Evidence Development and Coverage Advisory Committee (MEDCAC) gathers input from clinician societies and patient organizations to better understand the outcomes that patients and providers feel are meaningful. The committee publishes questions and answers, collects comments, and offers opportunities for interested parties to present at MEDCAC meetings. Medicare then uses these data to shape policy, and patients are thereby included in the process, said Wise.

End User Terminology Considerations

Villarreal commented that the end user of TBI technologies is often not the patient but the clinician; patients, in turn, benefit from clinician use. He asked how viewing the clinician as the end user affects the process of

patient engagement through the product development pathway. Schrandt contended that given that the ultimate benefit or harm of a medical treatment accrues to the patient, the patient is an important end user even if not the person physically using the device. For instance, surgeons deployed the morcellator tool to remove uterine fibroids, but later learned that in some cases, this practice led to the spread and increased severity of an undiagnosed cancer and harmed the patient.

Determining how the patient community can play a role in scientific endeavors such as identifying a biomarker may seem difficult, but Schrandt asserted that a nexus between patients and the development of any treatment is always present. Additionally, patients can be involved downstream, and their feedback can be incorporated into the care and clinical pathway. She noted that although this practice occurs regularly, it has not yet extended to all areas of medicine. Several efforts are currently underway to revise diagnostic guidelines in partnership with patients and patient groups. The data generated from patients cannot be collected only from clinical trial data, claims data, or EHR data, she noted. Schrandt specified that observing social media commentary is not a replacement for active engagement with patients in identifying gaps and better addressing needs.

Innovation Pathway Considerations

Kristy Arbogast, R. Anderson Pew Distinguished Chair, Department of Pediatrics at Children’s Hospital of Philadelphia and University of Pennsylvania, said that presentations pointed to strong, unbiased evidence as a gateway to implementation. Highlighting that federal funding is often limited for use in implementation, she asked whether industry-funded research could be approached in such a way as to decrease potential bias. Her question was followed by a query from Christopher Loftus, chief medical officer at the FDA Division of Neurological and Physical Medicine Devices, who commented on the tendency to view the emergence of a new medical device as reaching the “end of the line” and shifted focus to adoption. Highlighting the secondary stages of implementation, such as evaluation, financing, and coverage, he asked whether more focus should be given to these aspects of innovation adoption.

Grossman recalled working at the National Institutes of Health (NIH) where Anthony Fauci, then director of the National Institute of Allergy and Infectious Diseases, often described treatment development as a relay race in which NIH developed tools and tested HIV interventions and then handed them off to the Centers for Disease Control and Prevention or the U.S. Agency for International Development. In this analogy, success depends on the entire team and all legs of the race. Grossman asked about key stakeholders that should be brought into the process early on.

Wise offered her perspective from an industry background, noting that industry currently funds 70 percent of clinical research. Remarking that industry funding constitutes too large a portion of available funds for all researchers to disengage from industry-funded studies, Wise stated that she does not believe this funding can be totally unbiased. Companies answer to investors, and investors would not consider funding studies that do not move commercial goals forward to be an effective use of funds. Industry may provide grants or use other funding mechanisms, but the need to advance commercial goals is always present, she emphasized. The relay race analogy does not hold in the pharmaceutical industry, where numerous activities occur simultaneously, Wise described. However, the capacity for simultaneous activity is limited in some startup companies because of smaller workforces and budgets.

In advising startup companies, Wise encourages innovators to begin the process by considering whether money can be made from their idea. If an innovation can address an unmet clinical need, and the market is seeking a solution, then engaging patients can improve the development process. She noted that she encourages innovators to participate in design thinking and consider all elements—such as FDA requirements, payer requirements, and background data—from the start. Patient engagement enables understanding the outcomes that matter most to patients. Wise pointed out that clinical care guidelines have shifted to clinical outcome guidelines in response to payer emphasis on outcomes.

Biomarker Implementation in Prehospital Settings

Jeremy Kinsman, emergency medical services (EMS) specialist at the National Highway Traffic Safety Administration, asked whether an opportunity exists for civilian EMS providers—including emergency medical technicians and paramedics—to use point-of-care blood biomarker tests in a similar fashion to their use of glucometers and lactate tests to triage head injury patients to trauma centers or EDs that specialize in TBI care. Papa replied that having whole blood cartridge capability will make TBI biomarker testing much more accessible for varied settings, including the opportunity for prehospital personnel to use this device. She remarked that biomarker testing would be particularly helpful for communities located far from a neurosurgical care or trauma center in determining whether a long trip to specialized care is necessary. The biomarkers are evident soon after injury and can therefore be used in the prehospital setting to triage patients and determine whether a neurosurgical intervention or CT scan is needed. Papa added that point-of-care devices would also be helpful in other varied settings such as communities worldwide that lack access to CT

scanning technology and the military, as well as on sporting event sidelines to determine whether an athlete should be removed from play.

Exploring the Prognostication Value of TBI Biomarkers

Noting that many TBI patients diagnosed with mild TBI later experience prolonged challenges with cognition and everyday tasks, Kinsman asked whether biomarkers could be helpful in identifying patients who present with ostensibly mild TB in the acute care setting but are at risk for developing more severe, long-term symptoms. Papa responded that mild TBI is a misnomer, and the nomenclature should be changed to better reflect this reality. In 30 percent of cases, mild TBI causes significant symptoms or problems. Papa replied that these cases surface in follow-up studies. In her opinion, evidence is not yet sufficient to be certain that biomarker results can be used to prognosticate, said Papa. Data indicate that higher initial levels of certain biomarkers are associated with patients having a worse recovery trajectory after TBI (for example, Frankel et al., 2019), she continued, and this offers promise that biomarkers could potentially have prognostication value, but she believes the direct evidence remains insufficient. Papa noted that this is an important question that many researchers in the TBI biomarker space are currently working to answer.

Closing Reflections

Grossman closed the session by emphasizing the value of designing innovation with the end in mind and bringing stakeholders into the process early on. The innovation pathway is not a straight road, she said, and stakeholder engagement does not guarantee success nor predict the final product. Rather, she emphasized that engagement enables ongoing learning that can be integrated into design and implementation.

This page intentionally left blank.