Alzheimer's

Poster

REACHED: Participant characteristics from a pilot pre-screening study designed to diversify clinical trial recruitment in understudied populations in Alzheimer's disease

Alexandria Wise-Brown, Frank Boess, Frank Perilli, Ira Goodman, Erin Beck, Karra D. Harrington, Nicklas Linz, Alison Buick, Nicola Maiorano, Ruth Croney, Kelly Coultes, Janice Smith

*Poster presented at CTAD 2026

Background:

• In the US, ~7.4 million people live with clinically diagnosed AD, with varying risk per race and ethnicity – Black or African American and Hispanic/Latinx populations face a higher risk of developing AD and other dementias than non-Hispanic White Americans

• Underrepresentation of populations across clinical research reflects a multitude of interconnected barriers, specifically:3,4 – A shortage of research-active specialists providing care for diverse populations – Protocols creating participation burdens for those facing socioeconomic challenges – Gaps in non-English language support

• Reduced therapeutic access and greater social determinants of health burden disproportionately affect populations understudied in clinical trials

• Initial barriers to clinical trial participation impact translatability of findings and may confound efficacy and safety expectations when treatment becomes available to the diverse, heterogeneous population – Understanding neighborhood-level socioeconomic disadvantage may help to inform recruitment initiatives for reaching understudied populations

Methodology:

• REACHED methodology has been presented previously8 Study design and population

• REACHED was a US-based observational, prospective database study conducted via a centralized virtual research site between December 2024−April 2025 (13 US states)

• Participants were aged 50–90 years with eAD (mild cognitive impairment [MCI] due to AD or mild AD dementia), mostly from understudied populations (Black/African American, Asian American, Native American, or Hispanic/Latinx ethnicity regardless of race), alongside a reference group of 10 non-Hispanic White participants Participant identification and enrollment

• Participants were identified using an AI-assisted search of a real-world database, based on study-defined inclusion/exclusion criteria and International Classification of Disease billing codes (ICD-9 or ICD-10) from EMRs

• Individuals recommended to take part via a treating physician received a Health Insurance Portability and Accountability Act (HIPAA) release of information (ROI) authorization – Those who signed the ROI were invited to complete a web-based eligibility questionnaire; completed questionnaires and medical records were reviewed by the Principal Investigator

• Upon signing an electronic informed consent form (eICF), participants were enrolled Remote data collection and assessments

• Remote participant characteristic data collection included: – Area Deprivation Index ([ADI]; categorized into low [1–4], moderate [5–6] and high [7–9] deprivation ranks per state deciles)7 – self-reported education – living arrangements – transportation barriers

• Participant characteristics included medical history, including prior AD biomarker (amyloid or tau) testing, clinically significant disease history, and medication use

• Cognitive performance was assessed using two remote neurocognitive tests: an automated phone-based voice assessment (Speech Biomarker for Cognition [SB-C])9,10 and a smartphone-based List Learning and Memory Assessment (LILA ) application11–13

• Following remote assessment completion, an exit survey was provided and potential participant Phase III eligibility was assessed

Results: Participant characteristics

• Overall, 61 participants, aged 53–89 years, were enrolled; 50 (82.0%) participants were from understudied populations (Table 1)

  • – Most participants were retired (n=50; 82.0%), and 36 (59.0%) had foundational educational experience (i.e., some college [non-graduate], some high school/high school graduate, or elementary school)

  • – Most participants lived with others; among those living alone (n=7), median ADI score was 7, indicating high neighborhood deprivation – A total of 33 (54.1%) individuals were classified within the moderate-to-high ADI deciles, regardless of race/ethnicity

  • – Most (n=42) participants had prior documented cognitive testing. Of those with a documented instrument, 18 had a Mini-Mental State Examination, 11 had a Montreal Cognitive Assessment and 1 had a St Louis University Mental Status examination

• Primary ICD-10 diagnostic codes were MCI (n=26), mild AD (n=16), other amnesia (n=15), and unspecified dementia (n=4)

Digital assessment feasibility and completion

• Cognitive assessments determined via the LILA application and the phone-based SB-C captured clinical severity gradients within the expected range 14,15 across diagnostic groups (Figure 3)

• Observable trends emerged in time from consent to remote assessment completion via LILA application

  • – Primary Spanish speakers experienced longer delays downloading and opening the application (median: 5 vs. 2 days) and longer overall completion times (Figure 4), compared with primary English speakers

  • – Median completion time was 7.0 days for both White non-Hispanic and Black/African American cohorts, and 9.0 days for the Hispanic/Latinx cohort, while the Asian and Native American cohorts recorded the longest times at 13.0 and 16.0 days, respectively

  • – Observed latencies were driven by language preference and structural access. These remain descriptive operational trends patterns, across very small subgroups without formal statistical hypothesis testing

Study completion and exit survey

• Study completion was 80.3% (n=49/61) overall, and 86.0% (n=49/57) among those who initiated cognitive testing. Remote eICF introduced minimal latency: same-day signature (median: 0.0 days; mean: 1.3±2.6 days)

• In total, 63.9% (n=39/61) of participants completed all assessments within the protocol-defined study timelines. Mean overall time from eICF to final assessment completion was 8.5 days

• Descriptively, completion was higher for the phone-based SB-C assessment than for the LILA application (91.8% [n=56/61] versus 83.6% [n=51/61]), respectively (Table 2)

• Exit surveys demonstrated high participant acceptability, with most (n=43; 87.8%) participants reporting satisfaction with the study participation process

Potential eligibility for Phase III trials

• When mapped against publicly available study criteria for early AD trials, a subset of participants (n=15–21; 24.6%–34.4%) shared baseline age, diagnostic, and preliminary cognitive characteristics consistent with cohorts evaluated in trial pre-screening; biomarker status confirmation would help determine eligibility

Discussion:

  • REACHED provides a broad view of greater diversity of participants with eAD, identified through AI-assisted EMR review and enrolled through a centralized virtual research site. Cognitive assessments captured expected clinical severity gradients across diagnostic groups, while self-administered smartphone and voice-based tools were feasible in this diverse cohort

  • REACHED demonstrates that AI EMR phenotyping must search broad amnestic umbrellas, rather than narrow AD codes, to successfully capture underrepresented groups – In real-world community clinics, these codes frequently capture vascular cognitive impairment, nutritional deficiencies, medication-induced delirium, or psychiatric pseudodementia, and primary care providers serving underserved populations frequently avoid specific AD codes due to diagnostic uncertainty and lack of specialist access 16

  • Findings highlight potential barriers to clinical trial participation related to socioeconomic context, language, digital workflow, and prior access to biomarker testing that may be relevant when considering decentralized or digitally enabled trial methodology. Further inquiry is required to evaluate potential barriers to remote assessment completion, ranging from digital literacy to workflow timing, to inform evidence-based solutions

  • Learnings may inform recruitment strategies of understudied populations in future clinical trials

Related Articles