Clinically Meaningful Outcomes in AD/ADRD Trials

Day 1 | Tuesday, March 12

Read about the day 1 opening remarks and keynote presentation.

Welcome and Opening Remarks

Luke Stoeckel, Ph.D.; Richard Hodes, M.D.; & Walter Koroshetz, M.D.

Drs. Stoeckel, Hodes, and Koroshetz welcomed attendees to the three-day virtual workshop which was intended to (1) identify criteria for clinically meaningful outcomes in Alzheimer's disease (AD) and Alzheimer's disease related dementias (ADRD) clinical trials, (2) describe methods to validate and customize clinically meaningful outcomes, and (3) list practices that will ensure clinically meaningful outcomes are applicable to diverse populations. The workshop also aimed to foster robust multidisciplinary discussion between people with lived experience (PWLE), advocates, researchers, clinicians, funders, health economists, payers, and regulators. Each session started and ended with a research gaps and opportunities slide to help investigators identify innovative ways to use clinically meaningful outcomes to improve the design of AD/ADRD clinical trials.

Drs. Stoeckel, Hodes, and Koroshetz emphasized that clinically meaningful outcomes are subjective and vary from patient to patient. Therefore, investigators should include individuals living with AD/ADRD and their caregivers in the AD/ADRD clinical trial design process from the beginning to ensure that the clinically meaningful outcomes they prioritized are captured. Each AD/ADRD patient prioritizes unique clinically meaningful outcomes, and each cognitively impaired state requires different clinically meaningful outcomes to accurately assess cognitive change, including decline.

Keynote Presentation: Why Should ‘What Matters to People Living with Dementia’ Matter to Dementia Researchers?

Siobhan Reilly, Ph.D., University of Bradford

An outcome is a naturally or artificially designated point in the care of an individual or population suitable for assessing the effect of an intervention or lack of intervention. A patient-reported outcome (PRO) is any report of the status of a patient’s health condition that comes directly from the patient without interpretation of the patient's response by another party. Patient-reported outcome measures (PROMs) are instruments that are used to measure PROs and are most often self-report questionnaires. To identify appropriate outcome measures, investigators can access the COnsensus-based Standards for the selection of health Measurement Instruments (COSMIN) database, the Patient Reported Outcome Measurement Information System (PROMIS), and the Core Outcome Measures in Effectiveness Trials (COMET) Initiative. However, neither COSMIN, PROMIS, nor COMET identifies outcomes that are meaningful to individuals living with dementia. After collecting input from leading clinicians, researchers, and patients, the International Consortium for Health Outcomes Measurement (ICHOM) developed a standard set of outcomes for dementia patients that all health care providers should monitor. This standard set of outcomes includes measures for motor function, communication, sleep, learning disorders, aggression, caregiver burden, and overall quality of life. ICHOM measures are unique because they prioritize the outcomes that matter most to individuals living with dementia.

Measuring outcomes in dementia patients is challenging because many people living with dementia present with different symptoms and reside in a variety of settings. There is a high degree of variability of outcomes and tools across clinical trials, which hampers comparing effectiveness of each clinical trial’s intervention. It is unclear whether the outcomes being measured in clinical trials are more important to the people living with dementia or the dementia research community; previous systematic reviews, qualitative interview-based studies, and consensus exercises have found that individuals with dementia were rarely consulted before investigators chose a study’s outcome measurements.

To overcome challenges associated with measuring outcomes, the dementia research community should develop a core outcome set (COS) that reflects the most meaningful outcomes for individuals with dementia. In a COS, investigators conduct research to identify the outcomes that should be measured and determine how to measure them. A COS is applicable to all trials within the scope of the COS, which enhances comparability, responsiveness, and relevance of research. Dr. Reilly and her team developed COS for non-pharmacological community-based interventions for people living with dementia at home. The team aimed to identify outcomes that individuals living with dementia and their care partners would prioritize and conducted research to determine how these outcomes should be measured. The team interviewed focus groups of PWLE and identified 54 potential outcomes that could be measured in dementia-related trials. Then, individuals living with dementia prioritized the outcomes in a Delphi survey. Dr. Reilly and her team held a consensus meeting to finalize a list of 13 core outcomes. The outcomes were grouped into four domains: (1) self-managing dementia symptoms, (2) independence, (3) friendly neighborhood and home, and (4) quality of life. The core outcomes overlapped with social health, which refers to an individual’s ability to interact and form meaningful relationships with others; strong social health is associated with a reduced risk of cognitive decline and dementia. The overlapping of core outcomes with social health demonstrated a shift in outcome focus from symptoms deficit and disability toward the dementia patient’s mental capacity.

Dr. Reilly and her team reviewed 76 dementia patient-reported questionnaires to identify which included all 13 core outcomes; however, no questionnaires included more than 7 of the 13 outcomes, and most included fewer. These findings indicate that existing dementia PROMs do not include the outcomes that people living with dementia value but instead assess symptom reduction and quality of life. Investigators are unable to measure changes in clinically meaningful outcomes unless their assessments reflect the outcomes most valuable to PWLE. In the future, investigators from around the world should collaborate to design a searchable database for all outcome measures that are prioritized by dementia patients and their care partners and discover new technologies to support real-time measurement of outcomes.

Session 1: Cognitive Function

Moderator: Rosie Curiel Cid, Psy.D., University of Miami
Co-Chairs: Melissa Treviño, Ph.D., NIA & Kristi Hardy, Ph.D., NINDS

Equity in Cognitive Measurement

María Marquine, Ph.D., Duke University School of Medicine

Cognition encompasses various mental processes, including perception, attention, memory, reasoning, and problem-solving. These cognitive abilities are inferred from behavior through several methods, including self- and informant reports, observation, digital detection, and performance-based cognitive tests. Cognitive tests provide an estimate of brain function and can identify people with neurological diseases or injuries. However, performance on cognitive tests can be influenced by factors that are not indicative of brain dysfunction (e.g., sociocultural background, lived experience, emotional state). These factors need to be accounted for in order to accurately identify neurological disorders via cognitive tests, and normative data that consider sociodemographic characteristics can be one helpful tool to do so. Cognitive measures and their accompanying normative data have been largely developed by and are primarily for non-Hispanic, White, English-speaking people, which limits their utility in populations that have been underrepresented in the development of the science and leads to potential misdiagnosis in these groups.

Dr. Marquine and her team found that utilizing normative data developed for Hispanics/Latinos in the Uniform Data Set (UDS) Neuropsychological Battery significantly improved detection of cognitive impairment in this group. Using individuals as their own comparison is ideal when tasked with determining whether there is a newly acquired brain impairment and/or a clinically meaningful change for the individual. Yet, objective cognitive data before the onset of neurological disease are typically not available. Neuropsychological test norms serve as a standard to establish “baseline” expectations against which performance can be compared. Identifying clinically meaningful changes in cognitive test scores necessitates determining whether the changes are indicative of brain dysfunction and the degree to which they are reliably associated with a meaningful change in a patient’s clinical status, function, or quality of life. What is considered meaningful can change over time and may vary among individuals due to influences from different cultures, environments, and personal values.

Dr. Marquine outlined several guidelines and approaches to promote equity in cognitive measurement, including the (1) adaptation of tests and harmonization of batteries to ensure construct equivalence and accurate interpretation across cultures and languages, (2) careful consideration of cultural and language factors in test development and interpretation, (3) application of relevant statistical approaches and test formats to account for diversity, and (4) development of tools that are appropriate across cultures and languages from the outset. These guidelines are supported by various studies and initiatives, including the American Psychological Association's Presidential Task Force on Neuropsychological Test Norming in Diverse Populations.

Finally, Dr. Marquine discussed the potential of digital biomarkers, passively collected via mobile devices, to enhance the identification and tracking of neurocognitive disorders. These novel approaches can provide continuous, real-world data and hold promise for improving equity in cognitive assessment.

The Adaptive Measurement of Cognitive Ability

Robert Gibbons, PhD, University of Chicago

Measurement is defined as the process of obtaining the magnitude of a quantity relative to an agreed upon standard, but this definition is incomplete because it makes the incorrect assumption that researchers have developed an agreed upon standard. Cognitive ability is a latent (unobserved) trait, inferred from a set of measures with established psychometric properties, usually defined relative to a normative sample adjusted for age and education. Classic cognitive measurement models rely on observed scores without accounting for the difficulty of test items or the abilities of individuals. Item response theory (IRT) refers to a family of mathematical models that establish the relationship between latent traits (e.g., cognition) and observed variables (e.g., responses), organizing them on an unobservable continuum. IRT models allow for the assessment of individuals' positions on this continuum and can accommodate various response formats and dimensions.

Computerized adaptive testing (CAT), built on IRT, consists of a tailored set of items selected to be administered based on the responses of the examinee to the previously administered items. The main advantages of CAT include a reduction in testing time, improved accuracy with the same number of items compared to a fixed test, constant measurement precision across the ability continuum, and enhanced test security. Early CAT applications were based on unidimensional-IRT models for dichotomous data. New models allow for different response formats (e.g., ordinal, continuous) and assessment of more than one dimension using multidimensional-IRT (MIRT). These newer models have enabled the development of CATs for complex traits such as cognitive ability/impairment. MIRT-based CATs allow test takers with very little cognitive impairment to be measured with the same level of precision as a severely impaired test taker and eliminate unwanted practice effects by administering different questions over time. MIRT-based CATs are validated against in-person structured clinical neuropsychological assessments and can be used to identify cognitive impairments and measure cognitive decline over time. Additionally, MIRT-based CATs are highly accessible because they are cloud-based, Health Insurance Portability and Accountability Act (HIPAA)-compliant, can be administered anywhere in the world using any internet-capable device, and can be integrated into electronic health record (EHR) systems. MIRT-based CATs, such as the CAT-MH for adults and the K-CAT for children, are sensitive to differences between racial and ethnic groups and can quickly and accurately measure many different conditions, including depression, anxiety, psychosis, and suicidality.

Dr. Gibbons and his team have developed a prototype MIRT-based CAT to assess cognitive ability, known as pCAT-COG. The pCAT-COG has 57 items in its bank, and each one is associated with three tasks that are scored from zero to three based on the number of correct responses given. The investigators calibrated the pCAT-COG with 730 participants, 646 from an online system and 84 from the Rush Alzheimer's Disease Center (RADC). Cognitive impairment was assessed in the following domains: cognitive ability and flexible cognition, attentional control, episodic and semantic memory, processing speed, and language. The test demonstrated high correlation with total item bank scores and significant differences between RADC and online participants.

Funding from the National Institute of Aging will lead to the development of the full CAT-COG based on an expanded 500-item bank and full scoring of general cognition and specific sub-domains. Future versions of CAT-COG will use differential item functioning to determine whether a CAT-COG item used in the general population will have the same effect in a prespecified target group. Investigators can thus use CAT-COG to prevent biases and make high quality neuropsychological assessments widely available to underserved populations in the future.

Measuring Cognitive Function as a Dynamic Process in Health and Disease

Laura Germine, Ph.D., Harvard Medical School

As opposed to traditional paper-and-pencil tests, digital tools allow us to assess intraindividual change over time and across different contexts in an individual. Digital tools empower patients to conduct patient-initiated cognitive change monitoring in their own homes. Researchers can use these approaches and methods to estimate each patient’s dementia risk and assess momentary changes in cognition that could be used to aid in diagnosis, monitor age- and disease-related change, and detect responses to interventions in near real time. Momentary cognition uses time of day fluctuations, location, recent activity, health, sleep patterns, and environmental influences to capture how an individual’s cognition changes over minutes, hours, or days.

Dr. Germine discussed the use of two specific tests to measure momentary cognition: the Gradual Onset Continuous Performance Test (CPT) and the Digit Symbol Matching Test (DSMT). The Gradual Onset CPT assesses sustained attention and response inhibition, while the DSMT measures processing speed. These tests are ultra-brief (30 seconds to 1 minute), highly repeatable, and provide good reliability both between-person and within-person. Investigators can use both tests to identify cognitive differences between people and between time points within one individual. Cognitive variation from Gradual Onset CPT and DSMT is not random; therefore, the differences in the way an individual performs are predictable. Investigators can predict about 35 percent of the variance in CPT performance and 72 percent of the variance in DSMT performance when 20 time points are used.

Cognition is highly dependent on glucose fluctuations. To better understand this relationship, Dr. Germine and her team developed the GluCog study to assess cognitive changes in 200 adults with type 1 diabetes. For 14 days, the team used Gradual Onset CPT and DSMT to assess cognitive function and assessed glucose levels every five minutes using a continuous glucose monitor. Investigators found that low CPT scores immediately preceded hypoglycemic episodes, and that these low CPT scores continued even after the individual returned to a normoglycemic state. The research team found that (1) some individuals had stronger relationships between cognition and glucose fluctuations than others, (2) individuals with larger neck circumferences and microvascular complications had a stronger relationship between glucose levels and cognitive fluctuations, and (3) individuals who spent more time in hypoglycemic states experienced more cognitive fluctuations.

When investigators use momentary cognition assessments instead of traditional cognitive assessments, they can accurately approximate a patient’s response impulsivity. Traditional cognitive assessments could underestimate future performance on response impulsivity. Momentary cognition assessments can quantify cognitive differences between patients and capture practice effects better than traditional cognitive assessments. Additionally, momentary cognition assessments can capture information that traditional cognition assessments cannot, such as dynamic cognitive metrics, diurnal fluctuations, and volatility. Momentary cognition assessments enable investigators to derive novel, clinically meaningful cognitive outcomes that combine cognitive function with health variables to assess cognitive reactivity and identify early indicators of cognitive decline. Momentary cognition assessments enable researchers to derive stable, individualized phenotypes that reflect the dynamics of cognitive processes, providing a more comprehensive understanding of cognitive health.

Finally, Dr. Germine emphasized the potential of digital technology to revolutionize cognitive health assessment by (1) engaging patients, considering patient goals, and sharing research results, (2) increasing accessibility by providing measures that participants can complete in naturalistic environments, and (3) prioritizing low-burden monitoring to enabling patient-initiated cognitive change monitoring that can be shared with researchers and clinicians. These advancements promise to improve the accuracy, efficiency, and equity of cognitive assessments, ultimately enhancing the understanding and management of cognitive health in diverse populations.

Session 2: Everyday Function

Moderators: Elena Fazio, Ph.D., NIA and Benfeard Williams, Ph.D., NIA

Everyday Function and the Priorities of Individuals with Dementia Across Disease Stages

Andrea Gilmore-Bykovskyi, Ph.D., RN, University of Wisconsin-Madison

Everyday function is a multidimensional construct that considers physical, emotional, psychological, and social dimensions of functioning. Evaluation of Instrumental Activities of Daily Living (IADL) and Basic Activities of Daily Living (BADL) is commonly employed to assess everyday function. IADLs are closely associated with executive functioning and are essential for independent living, such as preparing and cooking meals, shopping, doing chores, and cleaning. BADLs are tasks that are essential for maintaining health and important for independent living, such as washing, eating, getting dressed, and using the bathroom. IADLs are more commonly affected in the early stages of cognitive impairment, whereas BADLs are more commonly affected at more advanced stages. Dr. Gilmore-Bykovskyi addressed changes in IADLs and BADLs in Stages: 1 (pre-symptomatic or preclinical AD), 2 (Mild Cognitive Impairment or MCI), and 3–4 (mild to moderate AD), with Stage 5–7 including more severe AD. It is important for investigators to distinguish between changes in IADLs and BADLs because changes in IADLs typically occur in early stages of cognitive disease (Stages 2–3) and changes in BADLs are more salient in moderate to advanced disease stages (Stages 4–7). However, it is difficult for investigators to reliably detect changes in IADLs and BADLs in the persons living with MCI (Stage 2), as decline can be masked by compensatory mechanisms. In the future, investigators could prioritize development of more sensitive measures of IADLs and BADLs to detect changes in functioning at MCI or early (i.e., preclinical) disease stages.

Dr. Gilmore-Bykovskyi reviewed two cognitive impairment studies that investigated how changes in IADLs vary across AD stages. In these studies, investigators observed a decline in IADL scores during Stage 1 of AD, discovered that functional decline accelerated in Stages 2–4, and learned that patients exhibited three classes of cognitive and functional decline: slow, moderate, and rapid. These findings indicate that the rate of cognitive and functional decline is not uniform across cognitive stages or individuals. It was suggested that the implications of variability in rate of decline be considered in outcome measurement, particularly when monitoring changes over time.

Dr. Gilmore-Bykovskyi discussed research demonstrating a lack of alignment between commonly used measures of everyday function (IADL/BADLs) and what individuals with cognitive impairment identify as the most important aspects of functioning. Most commonly used functional assessment measures do not detect the outcomes that individuals with AD/ADRD prioritize, such as finding meaning and enjoyment in activities; connecting with their identities; connecting with others and having a sense of belonging; and managing and adapting to changes over time. Instead, existing cognitive assessment tools typically capture cognition and function outcomes that researchers consider to be clinically meaningful. Additionally, cognitive assessment measures are often performance-based, whereas measures of IADLs and BADLs rely heavily on self- and proxy reports, and there are often discrepancies between these two different measurement approaches. Dr. Gilmore-Bykovskyi recommended that investigators systematically integrate performance-based assessments with other measurement strategies, including self- and proxy reports, across longitudinal cognitive aging cohorts. Investigators should use validated assessment tools, such as those available in PROMIS, to improve the representation of psychological, emotional, and social dimensions of everyday function that are not as well represented in current IADL and BADL assessments to capture changes in function that are meaningful to people with dementia. In the future, investigators should design AD/ADRD studies that capture behavioral symptoms (given their predominate role in impacting function), determine how social factors could help participants achieve their priority outcomes, and account for individual preferences and cultural values.

Clinically Meaningful Effects in Progressive Diseases with Disease Modifying Treatments: Time Saving Estimates

Suzanne Hendrix, Ph.D., Pentara Corporation

Many AD/ADRD studies fail to achieve significance on their primary outcomes yet demonstrate slowed disease progression. This discrepancy often arises because these studies choose outcomes that are acceptable for FDA submission rather than more progressive outcomes. The Food and Drug Administration (FDA) often prefers scales with large categories of response over scales with a continuous response because the former support perceived clinical meaningfulness, but they can also obscure true treatment effects. Dr. Hendrix emphasized the need to assess multiple aspects of the disease, including functional, cognitive, behavioral, and global aspects. Disease-modifying treatments are designed to slow disease progression and provide lasting benefits, even after treatment discontinuation. In contrast, symptomatic treatments provide immediate benefits but do not alter the disease's progression.

Disease-modifying treatments slow disease progression by affecting multiple aspects of disease; however, disease modifying effects are often more subtle than symptomatic effects. Although disease-modifying treatments are critical for managing neurodegenerative diseases, these treatment effects are difficult to establish for several reasons, including (1) FDA-acceptable scales are not sensitive to change and have a lot of variability, (2) patients must be treated early enough to observe a treatment effect, but late enough to measure progression, and (3) investigators must treat patients with neurodegenerative diseases before their symptoms meet the established symptomatic definitions of clinical meaningfulness. These challenges make it more difficult for investigators to design adequately powered studies of disease-modifying treatments than symptomatic treatments.

When assessing multiple outcomes, requiring significance at a strict alpha level (e.g., 0.05) for each outcome can be overly stringent, especially if the outcomes are uncorrelated. To address this, Dr. Hendrix suggested using combination outcomes, such as composite scores and global statistical tests (GSTs), and using time saving for interpretation. GSTs combine scores statistically using ranks, percentiles, or z-scores; to make GSTs understandable to patients, investigators can translate treatment effects into estimated time saving. Time saving is defined as the amount of progression in a disease that is delayed by a drug, which corresponds to preservation of function in earlier stages of disease. To calculate and interpret time saving, investigators can convert each outcome to time saving and then combine all time saving estimates into one overall time saving estimate. This leads to interpretable metrics that are meaningful to patients such as "X months saved of progression with Y months of treatment."

As disease-modifying treatments become more common, investigators must develop tools to better detect cognitive and other clinical changes in the early stages of neurodegenerative disease. In both disease-modifying and symptomatic treatments, investigators have often used fixed point values on clinical scales, established in the context of symptomatic treatments, to define clinical meaningfulness. However, clinical meaningfulness should consider overall disease progression and should be interpretable by patients and their families. Patients and their families can use time saving to weigh the benefits and risks of an intervention to determine whether participation in a study is worthwhile.

Everyday Function: A Search for Meaning across Clinicians, Regulators, Payers, and People

David Cella, Ph.D., Northwestern University Feinberg School of Medicine

Dr. Cella outlined a three-step process to ensure that measures are meaningful to the target population through content verification, anchoring, and prospective demonstration (validation). Content verification involves demonstrating that the measure's content is meaningful to the target population, typically through qualitative methods such as: (1) concept elicitation interviews to identify outcomes meaningful to the target population, (2) debriefing to confirm that the assessment’s items measure what is intended, and (3) literature reconciliation to ensure that the assessment aligns with existing scientific literature. PROMIS is one example of a measurement system that has used this approach to provide a comprehensive structure for self-reported health, encompassing social, mental, and physical health, symptoms, function, affect, behavior, cognition, and relationships.

Anchoring involves linking cognitive change scores to other similar measures with conceptual and statistical relationships. Clinical or person-centered anchors should have a conceptual relationship to the measure and a statistical relationship (r > 0.3-0.5), while bookmarking/standard setting exercises can help identify expert-defined thresholds for mild, moderate, and severe cognitive changes. Distribution-based methods, such as effect sizes and standard deviations, can determine whether a patient’s outcome has changed. However, FDA guidance states that distribution-based methods do not incorporate the patient’s voice and should not be the primary evidence to establish within-patient clinical meaningfulness. FDA guidance instead recommends that investigators observe the average change in patients who clinically improved or worsened by administering a self-report measure known as the Patient Global Impression of Change (PGIC). However, because the patient’s global impression of change is heavily influenced by their current status, investigators should not use PGIC as a primary endpoint, but instead use it as an anchor to cognitive change scores.

To measure cognitive change within an individual patient, researchers often use the standard error of measurement, reliable change index, or coefficient of repeatability (CR). Dr. Cella suggested that investigators identify likely change rather than absolute change by relaxing their confidence intervals and CRs to ensure significant individual change is not higher than the anchor change threshold, which captures patient-identified clinically meaningful outcomes. Investigators can support estimates of reliable change with prospectively collected anchor-based data and with participant or expert judgment regarding the magnitude of change considered to be meaningful.

In future studies, investigators should continue to use anchor-based methods to incorporate patient-identified clinically meaningful outcomes into study designs. Additionally, investigators should determine, on a case-by-case basis, when they should relax confidence intervals around individual change scores so as to maximize the probability of correct assignment (improved, unchanged, worsened). For example, a 90–95% confidence interval might be maintained in a trial where a concern about false positive assignment exceeds a desire to maximize classification accuracy. Implementing these practices will allow investigators to identify clinically meaningful and statistically significant changes as a result of disease-modifying treatments.

General Discussion

Read about the day 1 general discussion.

Meaningful Change

Current cognitive assessments often fail to capture outcomes that are meaningful to patients with cognitive impairment. Studies should prioritize incorporating patient-identified meaningful outcomes alongside disease-modifying treatment outcomes. The time saving measure can be a valuable tool for integrating these outcomes into study designs, helping patients achieve their health goals. Alternatively, item banks such as PROMIS, or perhaps a new item bank that calibrates preexisting cognitive scales using a MIRT-based CAT, may help address the need for patient-centered meaningful outcomes that are responsive to treatment.

Composite Endpoints and Measures

In proof-of-concept studies, investigators should use composite endpoints, which are more sensitive to changes in disease progression than primary endpoints and can identify which aspects of a disease are impacted by an intervention. For Phase III clinical trials, primary endpoints should be used to assess the consistency of effects observed in proof-of-concept studies across all outcomes. Instead of unit-weighted composite measures, investigators should derive composite measures using a MIRT model; one example is a bifactor model, which evaluates the empirical plausibility of subscales and the practical impact of dimensionality assumptions on test scores. A partial least squares regression model can be used to align composite measures with disease progression time.

Practice Effects

Practice effects are a valuable data source because they help investigators understand how well patients learn. CATs can be used to eliminate unwanted practice effects, such as question redundancies that create response biases. To identify practice effects determined by differential item functioning, cognitive assessments could begin with the same question or task. Investigators should also consider the differential impacts of systematic effects, stereotype threats, and test-taking familiarity on participants from different backgrounds.

Patient-Identified Measures

Patients with mental health concerns have indicated that using mean values to interpret or explain their mental state invalidates the variability of their experience and dismisses their clinical needs. They prefer assessments that consider their best and worst days, showing timepoints of their fastest and slowest disease progression rates. Wearable technologies can collect these data, and it is essential to develop methods to determine an intervention’s effect on a patient’s best and worst days.

Science Communications

Investigators often struggle to explain cognitive assessment measures and the utility of various data points to participants. Clinicians find it challenging to discuss a patient’s prognosis and explain expected changes and disease progression rates. Collaborating with science communication experts can help ensure clear phrasing of questions in cognitive assessments and effective communication of test results. It is imperative to consider language and cultural differences and to tailor communication strategies to each patient’s health literacy level.

Lived Experiences

Two PWLE shared that they were not accepted into AD research studies due to high cognitive scores, despite experiencing cognitive decline and difficulty with daily tasks. Their decline was not reflected in cognitive assessments because they were familiar with the questions asked. This would not have occurred using CAT-based assessment. Investigators must improve screening methods for cognitive research studies to avoid overlooking individuals who would benefit from treatment.

Day 2 | Wednesday, March 13

Read about the day 2 keynote presentation below.

Keynote Presentation: Clinical Meaningfulness in the Community

Ronald Petersen, M.D., Ph.D., Mayo Clinic

Dr. Ronald Petersen's presentation focused on evaluating clinical measures, imaging biomarkers, and plasma biomarkers in the general population to assess cognitive decline and the efficacy of interventions. Dr. Petersen emphasized that reducing amyloid in the brain does not necessarily result in clinically meaningful cognitive improvement. Interventions should target multiple aspects of the disease, including functional, cognitive, behavioral, and global aspects, to achieve meaningful outcomes. In addition, many clinical trials do not have population-representative samples, which presents a challenge for understanding how clinical measures, cognitive and behavioral assessments, and biomarkers might perform in the general population.

Dr. Petersen and his team conducted a population-based study involving 6,900 participants without dementia, aged 30 to over 90 years, from Olmsted County, Minnesota. The study aimed to identify clinically meaningful transition points during cognitive decline and the associated changes in clinical measures and biomarkers. Participants underwent comprehensive assessments, including nurse interviews, neurological evaluations, psychological assessments, and biomarker analyses. Based on the findings of previous AD trials, Dr. Petersen and his team designed their study to assess the clinical meaningfulness of two transition points that occur during cognitive decline: the transition from cognitively unimpaired to MCI and the transition from MCI to dementia. Over the duration of the trial, the team divided their participants into three cohorts: (1) cognitively unimpaired participants who transition to MCI and are amyloid-negative, (2) cognitively unimpaired participants who transition to MCI and are amyloid-positive, and (3) mild cognitively impaired participants who transition to dementia. Dr. Petersen and his team began by assessing clinical measures. In Cohorts 1 and 2, there was a gradual disease progression after the participants were diagnosed with MCI and in Cohort 3 there was a rapid disease progression after the participants were diagnosed with dementia. Behavioral and psychological symptoms, such as depression and anxiety, developed gradually in Cohorts 1 and 2 after the participants received their MCI diagnoses. However, there was relatively no change in depression or anxiety symptoms in Cohort 3. Dr. Petersen proposed that the patients in Cohort 3 were experiencing anosognosia, a neurological condition that occurs when the patient is unaware of their cognitive deficit or psychiatric condition. Similarly, the patients’ informants in Cohort 3 were significantly more aware of the patients’ cognitive changes than the patients were. In Cohorts 1 and 2, the patients’ informants detected cognitive changes before their MCI diagnoses and had more awareness of patients’ cognitive decline after their diagnoses.

The trial also assessed biomarkers. In Cohorts 1 and 2, amyloid levels in the brain increased steadily and tau levels also increased slightly in Cohort 2. In Cohort 3, amyloid and tau levels plateaued at the time of dementia diagnosis and then remained the same thereafter. In Cohorts 1 and 2, the beta amyloid ratio 42/40 decreased over time, indicative of cognitive decline.

The transition from cognitively unimpaired to MCI and the transition from MCI to dementia were clinically meaningful transition points. Dr. Petersen and his team discovered that all of the clinical measures and biomarkers changed significantly during these transition points, except for behavioral measures. Changes in depression and anxiety symptoms were less dramatic; thus, they should not be used to identify transition points in cognitively impaired subjects or used to assess the severity of cognitive impairment. By understanding these transitions and the associated biomarkers, researchers and clinicians can develop more effective interventions that address the multifaceted nature of Alzheimer's disease and related dementias.

Session 3: Biomarkers

Moderator: Christopher Weber, Ph.D., Alzheimer’s Association

Considerations for Clinically Meaningful Outcomes Among Diverse Populations

Sid O’Bryant, Ph.D., University of North Texas Health Science Center

Dr. O’Bryant’s presentation focused on the unique challenges and opportunities in using biomarkers to evaluate AD/ADRD among diverse populations. There are disparities in risk for AD/ADRD, with African Americans having twice the risk and Hispanics have 1.5 times the risk of developing AD/ADRD compared to non-Hispanic Whites. In addition, non-White populations tend to experience more health disparities than non-Hispanic White populations, such as delays in diagnosis, more severe pathology, greater likelihood of being diagnosed with psychiatric conditions, living longer with disease, and greater health care costs. Although minoritized individuals have a higher risk of developing AD/ADRD, over 90 percent of individuals in the AD research that led to the identification of AD biomarkers, known as amyloid, Tau, neurodegeneration (AT[N]), were non-Hispanic Whites. AT(N)-based AD biomarkers are more strongly associated with cognitive outcomes among non-Hispanic Whites compared to African Americans and Hispanics, who have other factors that potentially contribute to cognitive impairment, such as social and structural determinants of health (e.g., socioeconomic status). Thus, investigators do not fully understand how the AT(N) framework applies to diverse populations.

The Health and Aging Brain Study Health Disparities (HABS-HD) has approximately 4,000 participants enrolled from the general population. About 65 percent of this population are from African American and Hispanic communities. HABS-HD investigators examine biomarkers within a sociocultural and health disparities context and have discovered that amyloid positivity rates are lower among African Americans and Hispanics compared to non-Hispanic Whites. Additionally, APOE4 gene prevalence rates vary across diverse populations. HABS-HD investigators have also found that nine of the top 11 AD risk-conferring single nucleotide polymorphisms vary in allele frequency among different ethnicities. For example, APOE4 is less prevalent among Hispanics and the relationship between APOE4 and cognitive outcomes varies between ethnic groups. Vascular status, which is determined by white matter burden and predicts the rate of cognitive decline, also differs across various ethnic populations. In general, white matter burden is higher among African Americans and Hispanics compared to non-Hispanic Whites. Hispanics also tend to experience the first signs of cognitive impairment 7 to 10 years earlier than their non-Hispanic Caucasian counterparts, which aligns with an earlier onset of MRI-based neurodegeneration among Hispanics. The expression and sequence of imaging biomarkers (amyloid, tau, neurodegeneration), the values of plasma biomarkers, the relationship between imaging and plasma biomarkers, and clinical outcomes all vary among racial and ethnic groups. Social determinants of health (SDOH) and medical comorbidities impact these biomarkers differently across populations.

Dr. O’Bryant closed with a call to action for the research community related to recruitment and community engagement, evaluation of existing and new biomarker measures, and development of novel research approaches to address health disparities in AD/ADRD. This includes increased representation by aiming for 50 percent of research participants from non-White populations, reflecting the U.S. population, engaging with communities directly rather than solely recruiting from dementia specialty clinics, setting and adhering to specific recruitment goals, using exact numbers rather than percentages, investigating how imaging and plasma biomarkers vary across diverse populations to inform clinical trials and interventions, studying the combined genetic and epigenetic risk for AD/ADRD in diverse populations, and using these findings to develop targeted prevention and intervention strategies at both the population and trial levels.

Considerations for Use of Fluid Biomarkers in Clinical Trials of ADRD

Fanny Elahi, M.D., Ph.D., Icahn School of Medicine at Mount Sinai, James J. Peters, VA Medical Center, New York

Dr. Elahi’s presentation focused on readiness for the integration of fluid AD/ADRD biomarkers into clinical care, outside of the research setting. Proteomic fluid biomarkers represent measures of pathological or imbalanced physiological processes detected from biological fluids. The discovery phase included unbiased quantification of numerous molecules, but, for clinical utility, the list has been trimmed to a set of molecules that are reliably measured (from a technical perspective) and demonstrate a high level of biological validity within a given context of use. For clinical trials, biomarkers can provide a noninvasive, scalable means for risk stratification, diagnostic classification, and prognostication of course (pace of progression), and they can serve as surrogate outcomes of interventions. In clinical care, biomarkers can identify individuals who would benefit from disease-modifying therapeutics and help prognosticate disease course.

To achieve technical reliability, targeted biomarkers must have a consistent and reliable performance and be broadly usable across clinical settings. On the other hand, unbiased molecular omics measurements can be used to identify future targeted biomarkers and novel therapeutic targets, or they can connect human disease to well-suited translational models (connecting human biology to model systems). To achieve biological validity, the targeted biomarkers must be sensitive to core pathologies being captured and, if indicated by context of use (e.g.: differential diagnosis), exhibit disease specificity. At this time, biomarkers are also tested for associations with clinical outcomes. However, this ask may be limiting the discovery of biomarkers to select molecules with greater proximity to later disease stages.
Thus, given the underlying pathological heterogeneity of neurodegenerative disease symptoms (and syndromes), targeted biomarkers have the potential to identify contributing pathologies across the spectrum of symptomatology (i.e., from asymptomatic to dementia). As such, the use of biomarkers could reduce clinical trials’ sample sizes and improve patient care. However, most research on targeted biomarkers was performed in research cohorts that lacked real-world diversity (i.e., from medical, socio-economical and racial/ethnic perspectives). Therefore, more research is needed to understand the effects of real-life factors within our complex societies in order to equitably implement these game-changing tools within clinical care systems that serve diverse populations.

Highly validated blood-based biomarkers (BBMs) for brain degeneration and AD/ADRD include neurofilament light chain (NfL, an indicator of neuronal axonal degeneration), glial fibrillary acidic protein (GFAP, a marker of astroglial activation and degeneration), and hyperphosphorylated tau (p-tau, a marker of the dual proteinopathy in AD). Interestingly, hyperphosphorylated tau is a better predictor of amyloid burden in the brain than tau, but it is not yet understood why its levels are increased in the plasma of patients with AD dual-proteinopathy and not primary tauopathies. From this class of biomarkers, p-tau217 is the most promising, recently validated biomarker to predict amyloid burden in AD. Current AD BBMs perform best in non-Hispanic White (NHW) participants with a low burden of vascular and systemic comorbidities, which highlights the need for real-world studies of these biomarkers in richly diverse populations.

In the next few years, every major neurodegenerative disease proteinopathy will be detected with a blood-based biomarker. Promising results show that TDP-43-associated cryptic peptides could provide new biomarkers for this pathology; these biomarkers would capture aberrations in alternative splicing driven by TDP-43 mislocalization. Molecular measures related to vascular plasticity, such as the placental growth factor (PlGF) and vascular endothelial growth factors (VEGFs), are also emerging as promising biomarkers that capture vascular contributions to brain degeneration and vascular dysfunctions that could contribute to brain dysfunction and degeneration across the spectrum of proteinopathies. As such, the future use of BBMs in clinics will likely leverage combinations of biomarkers to capture the many contributing pathologies that drive brain dysfunction and, ultimately, the incidence of dementia.

In summary, investigators must collect real-world data by building study cohorts that reflect the diversity of the societies in which the biomarkers will be used. Research is needed to elucidate the impact of medical comorbidities on AD/ADRD BBMs and understand whether these BBMs are as predictive in diverse populations as they have been in NHW research cohorts. For instance, adjustments and corrections may be needed for comorbidities. In light of the many factors that impact the interpretation of BBM levels and the distribution of values observed in AD/ADRD studies, Dr. Elahi suggested that intraindividual trajectories of BBMs be used in clinical care instead of strict cutoffs. Coarse cutoffs may be useful for entry into clinical trials, but in clinical care where efforts are focused on modification of longitudinal trajectories of a given patient (n of 1), a cutoff is not only artificial, considering the wide distribution of values in research populations, but it is also likely less useful since pathologies and clinical symptoms are both on continuums rather than strict thresholds. Intraindividual trajectories also present a potential solution to the inaccuracy of predictions that are impacted by comorbidities. Therefore, new approaches may need to be implemented for maximizing the benefits of BBMs and reducing misdiagnoses and harm in clinic. The next few years will be very informative and, eventually, transformative for clinical care.

An Insider’s View on Research: Sharing Biomarker Results And Community Engagement

Sarah Walter, Alzheimer’s Clinical Trials Consortium & Barbara Boustead, Wisconsin Alzheimer’s Disease Research Center

Ms. Walter interviewed Ms. Boustead—a research participant for several AD studies—to understand what drove her to participate in AD studies, determine how researchers could improve the recruitment process, and learn how to establish and maintain a trusting relationship with research participants. Ms. Boustead became interested in AD studies while working as a clinical social worker assisting older adults to manage their financial affairs. As Ms. Boustead continued her work, she learned that her friend had been diagnosed with early onset AD, which encouraged her to get more involved with the Alzheimer’s Association and to participate in numerous workshops and walks. Ms. Boustead was hesitant to become an AD research study participant because she was concerned about the medical procedures involved, such as lumbar punctures and MRIs. However, she eventually decided to become an AD research study participant after attending informative brunches with staff at the Wisconsin Alzheimer’s Disease Research Center for several years and learning how the staff conducted their studies.

Ms. Boustead emphasized the importance of AD researchers immersing themselves in the communities where they intend to recruit research participants. Researchers should not wait to recruit research participants from hospitals or clinics because some people do not have reliable access to health care. Additionally, researchers must aim to establish a relationship with their research participants’ communities. To establish solid relationships with these communities, researchers must understand each community’s values and how they process information. For example, if a community values religion and enjoys learning in group settings, then researchers could recruit research participants for their AD studies in churches and host in-person information sessions to explain the importance of participating in AD research studies. Importantly, researchers must clarify that they are visiting different communities to learn from them, not to exploit them. Researchers should prioritize the well-being of all members of the community before, during, and after the research study to establish trust.

Ms. Boustead emphasized that it is important for researchers to share a participant’s results with them upon request. If researchers intentionally withheld information from participants, this could make participants wary of the entire AD study and lose trust in the researchers. It is important that researchers provide all participants with a Research Participant Bill of Rights, which will allow them to choose whether they want to receive their results or not, and have conversations with participants to identify which parts of the results would be valuable to them. It is equally as important to personalize the communication of these results as it is to personalize the quantification of these results.

Session 4: Payers, Regulators, and Health Economics

Moderators: Joseph Hutter, M.D., CMS, and Teresa Buracchio, M.D., FDA

Assessing Meaningful Change in Clinical Trials

Jeffrey Cummings, M.D., Sc.D., University of Nevada, Las Vegas

Assessing meaningfulness in clinical trials is a timely and necessary conversation for neuroscience research because of the number of new therapies approved recently for AD, ALS, and Friedreich’s ataxia. The balance of efficacy, safety, and cost of new therapies involves input from many stakeholders with varying priorities, and no single measure can incorporate all stakeholder values related to meaningfulness. Current measures of meaningfulness include clinical trial outcomes, biomarkers, quality of life outcomes, health economic measures, and PROs. Dr. Cummings noted PROs as an especially important area for further development and attention. Analyses related to meaningfulness include those that evaluate effect size, responder analysis, time to event/time saved, transition probabilities, and quality-adjusted life years (QALYs). Novel analyses may include evaluations of cumulative benefit (e.g., increasing drug-placebo differences over time) and predictive benefit (e.g., use of a predictive biomarker).

Dr. Cummings presented results from various clinical trials that evaluated monoclonal antibodies for AD, and he discussed how outcome measures can showcase a treatment’s ability to slow disease progression compared to placebo. He then presented data from a Phase III study of lecanemab that showcased the utility of the Clinical Dementia Rating Sum of Boxes score for quantifying a slowing of disease progression over time. The trial team also used data to predict other possible outcomes, including the need for future institutional care or mean time to dementia. In a separate study, patients identified several meaningful outcomes, including improving and restoring memory (66.7%), stopping AD progression (58.3%), slowing AD progression (33.3%), remaining independent and not feeling like a burden (10%), remembering family (20%), and removing plaques and tangles (10%). Patients have also underscored their right, as outlined in the Patient’s Bill of Rights, to try any FDA-approved therapy without an outlined perceived magnitude of benefit. Dr. Cummings concluded his presentation by highlighting the importance of creating a policy pathway through which breakthrough advances can be made available to eligible patients in a timely manner.

A Health Economics Perspective of Meaningful Benefit

Julie Zissimopoulos, Ph.D., University of Southern California

Treatments have both health and financial impacts on patients and families, and assessment of meaningful benefit should consider both. Measuring the various benefits and costs (e.g., earnings, hope, medical costs, pain, death risk) associated with a treatment will support decision making about its use by patients and their families. For example, a disease-modifying therapy (DMT) that delays cognitive and functional decline at early stages of disease would likely support workforce productivity of patients and caregivers, increase social enjoyment, reduce time in need of long-term support and services, and provide hope. However, the same DMT may also impose high out-of-pocket (OOP) costs for patients and families, impose high costs on payers, lead to suffering via treatment administration or side effects, and, if the DMT leads to increased risk of treatment-related death, shorten lifespan.

Modeling clinical trial outcomes can help quantify a DMT’s long-term meaningful benefit. Dr. Zissimopoulos presented an example in which endpoint data from a lecanemab study was applied to a larger patient population with similar characteristics to the trial population and analyzed via a model to quantify the long-term value of the treatment to patients and their families, including impact on medical care costs, quality of life, and caregiving hours. She presented the results from three scenarios: (1) an 18-month lecanemab trial that had a 27% slowing of cognitive decline, (2) a 48-month lecanemab trial that maintained the 27% slowing, and (3) a hypothetical 48-monthDMT trial that had a 50% slowing of cognitive decline. Dr. Zissimopoulos then reported on the lifetime economic value associated with the treatment’s effect on QALYs, informal caregiving hours, and medical care savings for each of these three scenarios. She found that all treatment scenarios were associated with an increase in lifetime value relative to no treatment. Lifetime value also increased significantly with longer treatment times and higher efficacy therapies. This analysis demonstrates that, while clinical trials may show short-term benefits, meaningful benefits, including QALYs and cost savings, accumulate over a longer period. Dr. Zissimopoulos emphasized that the way trials measure meaningful benefit influences research investments and the development of future treatments. Treatments should be evaluated for their health and financial impacts on patients, families, payers, and society, including both short- and long-term meaningful benefits to patients and society.

Healthcare Coverage, Payers, and Policy: Implications for People with Dementia

Jose Figueroa, M.D., M.P.H., Harvard University

People with dementia are predominantly insured through public (i.e., taxpayer-based) insurance programs, including Medicare and Medicaid. Estimates suggest that approximately $345 billion is spent on formal care for people with dementia; two-thirds of this cost is covered by Medicare and Medicaid, whereas approximately 25 percent is paid OOP by patients and families. Key challenges in caring for with people with dementia include failure of care delivery and care coordination, overtreatment, low-value care, abuse, and administrative complexity, in addition to the high costs of health care in the U.S. Thus, health care utilization and spending are generally higher for people with dementia, whereas clinical outcomes, quality of life, and patient experience are lower.

Emerging challenges facing people with dementia include: (1) increased enrollment into Medicare Advantage (MA), (2) a lack of integrated care for low-income, socially marginalized people with dementia (e.g., individuals dually eligible for Medicaid and Medicare), and (3) coverage and affordability of new therapeutics. Enrollment into MA has increased from 19 to 51 percent of all Medicare beneficiaries during from 2007 to 2023. Dr. Figueroa emphasized that MA enrollees can face unique obstacles, including very restrictive physician networks for some MA plans, complicated benefit designs that require prior authorizations for many services, and “choice overload” that can cause beneficiaries to face substantial choices when choosing plans. Data suggest that people with dementia are more likely to disenroll from MA because of these obstacles. Dr. Figueroa noted that a concern has emerged that minorities with dementia may be likely to select lower quality MA plans, possibly because restrictive networks lead to fewer care options.

Approximately 24 percent of people with dementia are covered by both Medicare and Medicaid (i.e., dual-eligibles), but the majority of these individuals must navigate these plans separately because of the lack of an integrated model. Lack of integration forces dual-eligible enrollees to figure out how to navigate two different sets of state-specific coverage rules, which leads to cost-shifting (e.g., MA plans may be less willing to invest in costly therapies that save Medicaid dollars) and inefficient and confusing health care. Integrated care models have the potential to improve care quality, clinical outcomes, and equity for dual-eligibles; however, more testing of these models is needed.

Further, affordability of new and emerging therapeutics is a major challenge and concern for the Medicare program. For example, current short-term costs of lecanemab treatment include the list price ($26,000 per person), administrative costs ($82,500 per person), and coinsurance (more than $5,000 per person). Out of concern for the projected high demand of aducanumab, Centers for Medicare & Medicaid Services (CMS) increased Medicare Part B premiums in 2022 by 15%. Estimates indicate that lecanemab will likely become one of the top-spending Part B drugs. The costs of these therapeutics pose equity concerns in uptake that lead to high OOP costs and overall coverage concerns among payers.

A related topic is prescription digital therapeutics (PDTs), which are generally not covered by Medicare even if the product is FDA-approved. Medicare covers software embedded in durable medical equipment, but it does not cover software in personal devices; thus, patients and families largely incur the costs of PDTs. CMS has tasked the Medicare Payment Advisory Committee (MedPAC) to help review PDTs; MedPAC has expressed concerns that PDTs undermine the integrity of payment bundles, limit competitive forces that drive down cost of services, lead to prescription overuse, and shift financial pressures from providers to Medicare.

General Discussion

Read about the day 2 general discussion.

Current Gaps

Dr. Buracchio asked each speaker to discuss what current repertoire of tools is missing that could provide more information on clinical meaningfulness in clinical trials and health economics outcome studies. Dr. Cummings noted that PROs are a current gap and asked whether PROs could be viewed as primary outcomes. Dr. Burrachio confirmed that PROs could be used as primary endpoints, but that challenges arise when patients progress further into disease and lose their ability to report on their own cognition. In these cases, observer-reported outcomes would be more appropriate than PROs.

Dr. Zissimopoulos noted that identifying and measuring factors that patients and families care about, which are not bought and sold in the health care market space, are major gaps; the field currently has the tools to identify these factors, but it has lacked the collection of such data, particularly from minority populations with lower incomes.

Dr. Hutter noted that treating appropriate patients and collecting real-world data along the way would help not only future patients (which is the goal of research), but also current patients by providing them with more information about how risks and benefits may change over time and when they should stop or restart treatment.

CMS VRDC

Dr. Gilmore-Bykovskyi noted that researchers often link Medicare data with other research and registry data (e.g., exposome data) to investigate health care utilization outcomes; she asked speakers to share thoughts on the CMS Virtual Research Data Center (VRDC) proposed shift. Dr. Figueroa noted that the VRDC’s change from physically distributing data to its digital-only infrastructure has raised concerns in the academic research community because using a virtual environment is expensive and could mean that some researchers cannot afford to use it. Researchers are also concerned about the security of potentially identifiable Medicare data.

Pharmacoeconomic Modeling

Drs. Zissimopoulos and Figueroa were asked whether their pharmacoeconomic models account for DMT versus symptomatic therapy. Dr. Zissimopoulos noted that the time saving that can be gained from a therapeutic are included in her team’s model, but the model’s assumptions can be changed based on new evidence.

AD Pathologies and Diversity

Participants reflected on the previously presented evidence that White individuals exhibit more amyloid-beta pathology as compared to non-White individuals. Dr. Elahi noted that amyloid-beta pathology is observed less frequently in real-world data than in clinical studies. She emphasized the need for better biomarkers aligned with key pathologies (e.g., TDP-43, micro-vascular disease) to optimize effect sizes for pathology-targeted therapeutics. Participants also noted that non-White individuals are excluded from trials more often than White individuals and that clinical trial screening tools may not be optimized for non-White populations. Dr. Hendrix recommended the research field not rely heavily on traditional cutoffs for amyloid-positivity inclusion criteria in non-amyloid-targeting clinical trials to avoid unnecessarily excluding minority patients.

Care Coordination

Navigating research studies, health care, and insurance programs can be difficult, particularly for an individual with dementia. Thus, care coordination can be a critical resource to receive appropriate and consistent care. Dr. Figueroa noted that primary health providers and governmental agencies should help provide care coordination.

Biomarkers for Preclinical Screening

Dr. Elahi emphasized that current treatment strategies for AD are administered too late, possibly because of cost of treatment, concerns related to toxicity, and difficulties in identifying preclinical individuals. She asked how close the field is to using a disease progression biomarker to identify and stratify clinical trial participants. Dr. Buracchio noted that FDA aims for researchers to use both biomarkers and clinical outcome assessments to better identify and risk-stratify patients, and that FDA recently published guidance related to evaluating products for early AD.

Day 3 | Thursday, March 14

Read more about sessions 5 and 6 below.

Session 5: Lived Experience Panel

Explore session 5, the lived experience panel.

The Lived Experience Perspective on Clinical Meaningfulness: "What does this mean for me?”

Sarah Walter, Alzheimer’s Clinical Trials Consortium

The Alzheimer’s Clinical Trials Consortium (ACTC) has a Research Participant Advisory Board comprised of a small group of individuals that have either research participant experience (clinical trials or observational studies) or are curious about research. Two of the members have been diagnosed with memory impairment and another two members have memory concerns but no diagnosis. Most members have provided care or support to an individual diagnosed with dementia. Ms. Walter thanked the members of the advisory board for their service and shared preliminary data collected when the group was asked to select between two hypothetical treatments, one of which slowed disease-related brain changes by a small amount and supported memory and thinking capabilities and the other that only supported memory and thinking. Many responses addressed the complex nature of such a prompt, noting the choice is influenced by safety concerns, accessibility and affordability, evidence supporting the drug as treatment, and whether alternatives exist. Board members emphasized that extending life without a good quality of life is not a motivator. In addition, people are looking for a treatment that can stop the progression of AD, and they hope to receive treatment that extends life by months to years, not days or weeks. One major influence on treatment decisions is who is recommending the treatment and how trustworthy the patient believes the physician to be. Board members also emphasized that treatments must be tested in all patient populations so that patients know how the drug works in patients with a similar background. Ms. Walter then introduced a panel of speakers with lived experience.

Barbara Boustead

Ms. Boustead, a retired clinical social worker and daily money manager, is a research participant in the Wisconsin Registry for Alzheimer's Prevention (WRAP), the Wisconsin Alzheimer’s Disease Research Center African-Americans Fighting Against Alzheimer's in Mid-Life (AA-FAiM), and a member of the University of Wisconsin ADRC Black Leaders in Brain Health. She has not been diagnosed with AD but has become involved in AD studies and advocacy because her career as a daily money manager led her to work with older adults and veterans. Her support of AD research was sparked by working with clients with AD and observing how AD impacts patient lives firsthand, including a close friend who passed away from early onset AD.

Rochelle Long

Ms. Long has taken part as a study participant in the Generations and Alzheimer’s Disease Neuroimaging Initiative (ADNI), serves on the Minority Advisory Board at the Cleveland Alzheimer’s Research Center and the ACTC Research Participant Advisory Board, and has served as a caregiver for multiple family members for more than 20 years. She shared that she inherited APOE4 alleles from both parents and has seen how AD impacted many family members, including her parents, aunts, and uncles. She also shared that while the majority of her health care experiences have been good, in the last year, she had one poor experience in which she was mistreated because of her race; she reported the incident and vowed not to allow another patient to feel similarly. Ms. Long stated that participating in clinical studies and advocating for AD is her way of becoming part of the solution to better treat and prevent AD.

Cynthia Huling Hummel

Rev. Dr. Huling Hummel first began experiencing memory issues in 2003 at the age of 49 and initially went through a myriad of possible diagnoses, including menopause and brain injury, before receiving the diagnosis of MCI 8 years later. In addition to being a retired reverend and singer in a country and rock band, she has been a research participant in ADNI for 14 years; serves on the National Advisory Council on Aging, ACTC Research Participant Advisory Board, and Dementia Advisory Board for the World Health Organization; and was a remote care partner for her mother, who lived with dementia. She is passionate about AD research and regularly speaks to service organizations about how to live with dementia. She hopes that what is learned from ADNI and other studies can help protect her children and future generations.

Panel Discussion

What has been your experience in participating in research studies?

Ms. Boustead shared that her experience participating in research studies has been welcoming, and she has been treated as though she is part of a community that can talk about fears and concerns. She added that providing better experiences in clinical visits, such as allowing family members to be close by during MRI scans, is key to ensuring a good clinical experience. She expressed that she is typically not asked too many questions during her visits because she functions well and does not appear ill, and she plans to use the knowledge she has gained during this meeting to ask her own questions regarding test results and procedures.

Ms. Long shared that her earlier experiences with clinical studies felt very warm and familial, with many participants discussing their children and recipes and celebrating each other’s successes—like when Ms. Long won the Global Alzheimer's Platform Citizen Scientist Cornerstone Award. However, the COVID-19 pandemic and her experiences in joining ADNI have changed that perception; her experiences participating in ADNI have not been positive, with visits leaving her feeling unvalued and detached from the overall study.

Rev. Dr. Huling Hummel shared that her participation experiences have been mixed. She shared similar experiences to Ms. Long in that her experience with ADNI felt lonely and isolated from other research participants. When she joined the ACTC Research Participant Advisory Board and met Ms. Long, she realized that she was not receiving the same information as other ADNI participants who received regular research updates from their local site. Until recently, Rev. Dr. Huling Hummel’s site would not share the same updates. She emphasized that patients have the right to discuss their procedures and test results with physicians, and read the following proposed Study Participant Bill of Rights:

STUDY PARTICIPANT BILL OF RIGHTS

1. I have the right to receive my individual results, collected in the course of my research participation, if I so choose; I can also ask how to receive them. This may be done in person or by tele-health, and either alone or with a person of my choosing.
2. I have the right to exercise this right in an informed manner, including receiving information on validated decision-making tools if they are available, knowing who can access my results, and how the law does or does not protect me after receiving my results. In order to protect myself, I may need to finish any care and legal and financial planning in advance of receiving my results.
3. I have the right to be told what my test results mean, in easy-to-understand terms and with sensitivity, compassion, and patience. This information should also be provided in writing, so that I may review it later.
4. I have the right for my questions to be answered to the best of the researcher’s knowledge and to take all the time I need to process the information I received.
5. I have the right to be contacted or decline to be contacted to check on my well-being after receiving a result suggesting increased risk of dementia, and to be referred to additional resources for more information and support related to my health and well-being.
6. I have the right to decide what actions I take after receiving my test results, such as pursuing medical and/or psychological care, engaging in legal or financial planning, and informing my family and friends of my results.
7. I have the right to turn my results into action for my own well-being and the betterment of others, by exploring additional research studies that I may qualify for.

The above rights should apply regardless of my cognitive status.

Which outcomes are clinically relevant to you, your family, and your community?

Ms. Boustead emphasized the importance of using less invasive BBMs to predict possible future care needs and clinical progression. She is most interested in outcomes related to quality of life and time left before the disease progresses. Ms. Long agreed that less invasive testing is helpful and that patients should know who has access to their donated samples and how those samples will be used. Ms. Long added that, because participants are donating these samples, they should feel that the physician is giving them something back, particularly by discussing test results and providing research updates. Rev. Dr. Huling Hummel noted that any outcome that can monitor her progress and provide insight on her current and future life is important to her so that she can understand why certain tests are needed and what they can reveal.

What would you like to see changed?

Ms. Boustead called for more culturally sensitive and relevant practices for non-White participants, and she recommended that researchers provide care and study participation within the community more often rather than only in the clinic. Ms. Long agreed that researchers must make more of an effort to enter and form long-term relationships with communities to show patients how important they are to these studies. Rev. Dr. Huling Hummel recommended that researchers treat patients like partners in the clinical study rather than participants as well as provide more remote care options. Rev. Dr. Huling Hummel concluded the panel by recommending that researchers integrate the Research Participant’s Bill of Rights .

Session 6: Breakout Sessions and Reports

Moderator: Luke Stoeckel, Ph.D., NIA

Workshop participants concluded the meeting by identifying the following prioritized gaps and opportunities that could be considered for future research efforts. In addition, the Lived Experience Panel provided a summary of their priorities.

Lived Experience Panel

Sarah Walter, Barbara Boustead, Rochelle Long, Cynthia Huling Hummel, Alzheimer’s Clinical Trials Consortium

1. Clinical Relevance is Three-Dimensional.
   • Gap: Clinical relevance can be extended to any factor that is relevant to brain health, family, friends, and the community.
   • Opportunity: Composite outcomes (e.g., cognition, motor, function, biomarkers), better testing of intervention efficacy and safety in representative populations, and more intervention studies (including head-to-head comparisons and combination drug/lifestyle studies) are needed.
2. Long-Term Partnerships with Communities.
   • Gap: Understanding what is clinically meaningful to an underrepresented community requires building long-term partnerships with that community.
   • Opportunity: Instead of researchers always meeting patients in clinics, researchers could enter the community to meet its members where they live, conduct health monitoring, and share research results.
   • Gap: Researchers should be in regular contact with study participants.
   • Opportunity: Study teams should share individual results and study-level findings with study participants, even if the meaning remains unclear.

Cognitive Function

Rosie Curiel Cid, Psy.D., University of Miami, Kristi Hardy, Ph.D., NINDS, and Melissa Treviño, Ph.D., NIA

1. Equity in Cognitive Measurement
   • Gap: There is a need for skillful and informed adaptation and cross-cultural validation of cognitive tests that measure clinically salient outcomes. Currently, there is a lack of harmonized longitudinal assessments of diverse populations in the U.S.
   • Opportunity: Develop novel cognitive testing approaches and technologies sensitive enough to detect and track clinically meaningful changes over time indicative of disease progression. This includes creating scalable digital measurement tools that allow dynamic assessment models accounting for real-world states impacting performance.
2. Adaptive Measurement of Cognitive Abilities.
   • Gap: Traditional fixed-length cognitive assessments often produce practice effects and lack precision.
   • Opportunity: Utilize large cognitive item banks for adaptive administration tailored to a person’s cognitive ability level. This approach enables high-frequency, precise, and practice-effect-free longitudinal assessments across a person's lifetime. Facilitate collaborative research to optimize and accelerate the availability of valid and reliable cognitive measurements for diverse older adults.
3. Measuring Cognitive Change Dynamics Using Digital Tools.
   • Gap: There is a need for novel, clinically meaningful outcome measures that combine cognitive function with health and physiological factors.
   • Opportunity: Develop paradigms and tools that are more naturalistic and ecologically relevant. Provide follow-up results to participants in an understandable format, enabling them to integrate findings into their lives. Employ adaptive testing to tailor items to the participant’s current ability, reducing humiliation and enhancing personalized assessment. Utilize differential item functioning to identify and eliminate biased items across languages, cultures, and subcultures.

Everyday Function

Elena Fazio, Ph.D., NIA, and Benfeard Williams, Ph.D., NIA

1. Researcher-Based Approaches.
   • Gap: Existing assessments may not fully capture the multidimensional nature of everyday function, including emotional, psychological, and social dimensions.
   • Opportunity: Develop multidimensional assessment approaches that incorporate domains prioritized by persons living with dementia, such as maintenance of self, belonging, and engagement in meaningful activities. Integrate multiple assessment strategies, including self-report, reliable informant report, and performance-based assessments.
2. Ongoing Community Input.
   • Gap: Outcomes measured are not always meaningful to people with AD/ADRD.
   • Opportunity: Refine and apply approaches to ensure that measured outcomes align with what matters to individuals with AD/ADRD. Gather qualitative and quantitative evidence from self-report and performance measures, identify thresholds for meaningful change, and apply these measures in clinical research and practice.

Biomarkers

Christopher Weber, Ph.D., Alzheimer's Association, Amber McCartney, Ph.D., NINDS, and Kristina McLinden, Ph.D., NIA

1. Inclusion of Diverse Communities. Advancement of AD biomarkers must include diverse communities.
   • Gap: More data is needed to understand racial and ethnic group differences in imaging and fluid biomarkers.
   • Opportunity: Increasing community representation requires community engagement and site selection strategies to be part of a study’s design, in order to avoid the inclusion of racial and ethnic groups only after enrollment begins.
2. Real-World Data. Real-world data will play a crucial role in the development of biomarkers for AD.
   • Gap: Real-world data will play a crucial role in the development of biomarkers for AD/ADRD.
   • Opportunity: The value of real-world data in developing biomarkers in AD/ADRD lies in its ability to provide a more comprehensive, diverse, and representative understanding of the disease’s complexity and progression.
3. Outside of the “Amyloid” Box.
   • Gap: Amyloid has been the major focus for biomarker development and trials.
   • Opportunity: Researchers should consider blood-based biomarkers capturing AD/ADRD pathologies other than amyloid that may be more relevant to diverse populations.

Payers and Health Economics

Joseph Hutter, M.D., CMS, Michelle Campbell, Ph.D., FDA, Nada Radoja, Ph.D., NIA, Priscilla Novak, Ph.D., NIA, and Marcel Salive, M.D., M.P.H., NIA

1. Patient- and Care Partner-Reported Outcome Measure.
   • Gap: We have measures of function, cognition, behavior, and health outcomes, but no patient/care partner-based outcome that has been widely accepted and integrated into trials.
   • Opportunity: Develop a patient/care partner reported outcome that could become a standard clinical trial measure.
2. Cost.
   • Gap: Newer DMTs are costly.
   • Opportunity: Overcome the growing challenge of covering and affording newer, costly drugs and digital therapeutics among an increasingly aging population.
3. Measures of Short- and Long-term Benefit.
   • Gap: We have measures of short-term (e.g., 18 months), but not long-term benefit in DMTs.
   • Opportunity: Develop measures and models that quantify the short- and long-term meaningful health and financial benefits of treatments, which could be used by various stakeholders.

General Discussion

Read more about the day 3 general discussion below.

Non-Competitive Partnerships

Dr. Campbell emphasized the need for more non-competitive partnerships that enable the convening of stakeholders across the AD research landscape and the quick sharing of data to answer key questions in the field.

Neuropsychiatric Symptoms and Social Health

Meeting participants noted two critical components that were not previously discussed in this meeting: neuropsychiatric symptoms and social health. AD is often discussed as a condition primarily of memory deficits, but patients and caregivers are typically more concerned about neuropsychiatric symptoms like hallucinations, depression, or delusions.

In addition, meeting participants emphasized the critical need to capture more measures of social health and behavior. Patients may be worried about their future ability to participate in interpersonal activities and social relationships.

Patient-Led Innovations

Meeting participants suggested considering more pathways for patient-led innovations in science. Increased information and data sharing could lead some patients to perform their own analyses and uncover new findings. Meeting participants noted that a pathway to allow these discoveries should be taken seriously because study participants are uniquely motivated to advance AD research.

Time Savings to Bridge Multiple Outcomes

Dr. Hendrix proposed that many clinically meaningful outcomes capture data that can be used to create a time saving measure; combining the time saving measures from various clinically meaningful outcomes into one comprehensive outcome may help compile findings across studies and align outcomes more closely to what patients find meaningful.

Pathways to Develop and Validate New Outcomes

Meetings participants highlighted the Clinical Path Institute (C-Path) as a pathway for changing clinical trial outcomes. Dr. Campbell noted that C-Path has developed an AD-specific consortium that is using a regulatory lens to look into clinical outcome assessments, biomarkers for AD, and novel outcomes for rare and neurodegenerative diseases. C-Path representatives added that data are often the rate-limiting factor to measure development and validation; more randomized and non-randomized clinical data are needed for these efforts.

Dr. Petersen emphasized that the AD research field needs mechanisms to validate instruments that capture clinically meaningful outcomes. Industry is hesitant to invest significant financial resources into trials with unvalidated measures, regardless of how meaningful an intervention may promise to be. He added that C-Path or partnerships with NIA may be the pathway to creating these validation mechanisms.

Dr. Cummings highlighted that it is critical to determine an outcome’s proposed context of use prior to development and validation.

Wrap-Up, Next Steps, and Closing Remarks

Luke Stoeckel, Ph.D., NIA

Following the conclusion of this workshop, the recordings will be made 508-compliant and posted online for viewing. NIA will support the drafting of a workshop summary that will also be shared online for review. Additional follow-up opportunities include a special journal issue on clinically meaningful outcomes, which will include a summary of the workshop and finalized gaps and opportunities from the workshop.

Dr. Stoeckel thanked all meetings participants, speakers, and organizers for their efforts in supporting and participating in the workshop.