1. Background
Blood-based Alzheimer's diagnostics has emerged as a rapidly developing IVD area over the past several years, with multiple plasma biomarker programs reaching FDA-cleared status and several more in late-stage development.2 Verification work on plasma biomarkers in this area requires specimens drawn across the cognitive-status spectrum — cognitively normal, mild cognitive impairment (MCI), and Alzheimer's-disease dementia — so that assay performance can be characterised across the clinical decision range, not just in research-volunteer cohorts where the marker distribution is narrower.
The cognitive-status stratification at the protocol level imposes a specific operational requirement: each enrolled subject must be classified to a cognitive-status group through a structured cognitive screening at or near the time of specimen collection. Site selection therefore weights memory-clinic infrastructure and on-staff cognitive screening capability heavily, in addition to general clinical-research throughput.
2. Constraints encountered
Three constraints shaped the execution plan:
-
Cognitive-status mix is the primary protocol-level constraint
The protocol specified target enrollment ranges across three cognitive-status strata. The MCI and AD-dementia strata are the rate-limiting cohorts: cognitively normal subjects can be drawn from broad clinical-research recruitment, while MCI and AD-dementia subjects are reached almost exclusively through memory-clinic and cognitive-disorder care relationships. Site selection had to weight memory-clinic patient-panel disclosure heavily against the easier general-population enrollment frame.
-
Cognitive screening at the visit level
Each enrolled subject required a structured cognitive screening at or near the time of specimen collection — the screening drives the cognitive-status classification, which is the protocol-level stratification variable. Sites with on-staff cognitive screening capability completed consent, screening, and draw in a single visit; sites without on-staff capability scheduled screening and draw across two visits, with operational losses concentrated at the screening-to-draw transition.
-
Clinical-status documentation drift
Cognitive-status classification has a temporal-stability property: a subject classified as cognitively normal at the time of draw may be reclassified later if cognitive trajectory diverges. The protocol’s cognitive-status documentation therefore had to capture the classification basis at the time of draw rigorously enough that downstream sponsor analysis could anchor on the documented status rather than re-classifying subjects retrospectively.
3. Methods
RDI executed the program as full-service CRO. Three operational moves are reported here in order of implementation.
3.1. Three-lead-memory-clinic anchor
Three memory-clinic and cognitive-disorder-research sites were activated as the program lead network. Site A (147 subjects) is a dedicated memory-clinic operation with on-staff cognitive screening, an MCI-and-AD-dementia patient panel measured in thousands, and a standing clinical-research apparatus. Site B (100 subjects) is a neuro-cognitive clinical-research organization with cognitive-screening capability and an MCI-track research panel. Site C (98 subjects) is a multi-specialty clinical-research organization with a cognitive-disorder track and external referral relationships into local memory clinics.
3.2. Three-site supplementary network for cognitively-normal enrollment
Three additional sites were activated for cognitively-normal enrollment — Site D (31 subjects, primary-care research clinic), Site E (19 subjects, primary-care clinical-research organization), Site F (8 subjects, multi-specialty research clinic). The supplementary network was deliberately weighted toward cognitively-normal subjects rather than MCI and AD-dementia, since memory-clinic infrastructure is the rate-limiting capability for the harder cognitive-status strata.
3.3. Single-visit consent-screening-draw workflow at lead sites
The three lead sites completed consent, cognitive screening, and specimen draw in a single visit. The single-visit workflow is the principal operational driver of the lead-site throughput; sites without on-staff cognitive screening required a two-visit workflow and saw operational losses at the screening-to-draw transition. The consent-to-draw window at the lead sites was typically same-day; at the supplementary sites it ranged from same-day to a 14-day window depending on cognitive-screening logistics.
General clinical-research recruitment frame
Multi-site network across primary-care or general clinical-research sites. Cognitive-status classification done at the laboratory back door from chart abstraction. Cognitively-normal enrollment runs at full pace; MCI and AD-dementia enrollment runs slowly.
Memory-clinic-anchored frame (this program)
Three lead sites with established memory-clinic infrastructure + a three-site supplementary network for cognitively-normal volume. On-staff cognitive screening at lead sites enabling single-visit consent-screening-draw workflow. Cognitive-status mix at the protocol level achieved without a recruitment-funnel penalty.
4. Results
Cumulative enrollment crossed 35 by end-June 2024 (program open + first lead site SIV), 149 by end-November 2024, 257 by end-January 2025 (program peak month: +72), 378 by end-March 2025, and reached 403 at program close in June 2025 (Fig. 1). Monthly cadence ramped sharply through the second half of 2024 as the three lead sites came up to full cognitive-screening throughput, peaked at +72 in January 2025, then unwound through Q2 2025 as the cognitive-status enrollment caps approached saturation.
Site distribution at program close: Site A 147 subjects (36%), Site B 100 (25%), Site C 98 (24%); supplementary network (D, E, F) 58 combined (14%) (Table 1; Fig. 2). The three-lead-site share of approximately 86% was tighter than the typical lead-site concentration in the general specimen-collection programs in this engagement series — a structural consequence of memory-clinic infrastructure being the rate-limiting site capability for two of the three cognitive-status strata.
| Site | Setting | Enrolled | Share |
|---|---|---|---|
| Site A | Dedicated memory clinic with on-staff cognitive screening; MCI/AD-dementia patient panel | 147 | 36% |
| Site B | Neuro-cognitive clinical-research organization with MCI-track research panel | 100 | 25% |
| Site C | Multi-specialty clinical-research organization with cognitive-disorder track | 98 | 24% |
| Lead-site subtotal | 3 memory-clinic sites carrying the bulk of program volume | 345 | 86% |
| Site D | Primary-care clinical-research clinic; cognitively-normal enrollment | 31 | 8% |
| Site E | Primary-care clinical-research organization; cognitively-normal enrollment | 19 | 5% |
| Site F | Multi-specialty research clinic; cognitively-normal enrollment | 8 | 2% |
| Supplementary network subtotal | 3 sites for cognitively-normal volume | 58 | 14% |
| All sites at program close (Jun 2025) | 6 sites total | 403 | 100% |
5. Discussion
Three observations bear noting. First, blood-based Alzheimer's biomarker specimen collection is operationally a memory-clinic-throughput problem rather than a general-population recruitment problem. The cognitive-status stratification at the protocol level forces site selection toward sites with on-staff cognitive screening and existing cognitive-disorder patient panels; sites without that infrastructure cannot deliver the cognitive-status mix the protocol requires within the available enrollment window. The typical clinical-research-organization site profile that delivers well in general specimen-collection programs is structurally weak in this disease area.
Second, the lead-site concentration of 86% is structurally tighter than the 60–70% concentration typical in the rest of this engagement series. A program that depends on a specific clinical-care infrastructure (memory clinics, harm-reduction programs, transplant centers) has fewer eligible sites available; the resulting site-selection funnel is narrower and the per-site-per-month variability is higher.
Third, the on-staff-cognitive-screening capability at the lead sites is the single highest-leverage operational variable. Sites with on-staff cognitive screening completed consent, screening, and draw in a single visit; sites without it scheduled across two visits and saw operational losses at the screening-to-draw transition. Site qualification protocols for cognitive-disorder specimen collection should weight on-staff cognitive-screening capability heavily against general clinical-research throughput.
6. Conclusion
Blood-based Alzheimer's biomarker specimen collection is a memory-clinic-anchored operational frame, not a general-population recruitment frame. Three lead memory-clinic sites with on-staff cognitive screening + a three-site supplementary network for cognitively-normal volume delivered 403 subjects across the cognitive-status spectrum in 13 months, with an 86% lead-site concentration that reflects the structurally narrower site-selection funnel for this disease area. The same operational frame transfers to other plasma-biomarker specimen collection programs in cognitive disorders.