Hidden Geography of Alzheimer's: Site Selection Insights

An estimated 7.4 million Americans age 65 and older are living with Alzheimer's dementia today, a figure that could reach 13.8 million by 2060 ^[1]. The communities carrying the sharpest share of that burden are well-documented: older Black Americans are about twice as likely as older White Americans to develop Alzheimer's or other dementias, and older Hispanic Americans face roughly 1.5 times the risk ^[1]. And yet, across 26 years of Phase III Alzheimer's trials conducted in the United States, White participants constituted nearly 90% of enrolled populations, while Native American participants appeared in fewer than 3% of trials that reported racial data at all ^[2].

That gap is not simply a recruitment problem. It is also, in part, a geography problem. The sites where Alzheimer's trials run are not consistently located where the disease burden is greatest, and the populations who carry the highest risk are not always within practical reach of the clinical research centers running the studies. Understanding why this pattern has persisted, and what it costs operationally and scientifically, is increasingly urgent now that disease-modifying therapies have finally reached patients.

Why Geography Shapes Who Gets Into Alzheimer's Trials

Site selection in Alzheimer's disease research has historically gravitated toward academic medical centers, large urban neurology practices, and established research networks. The logic is operationally sound: these institutions have trained coordinators, neuroimaging infrastructure, experienced principal investigators, and the regulatory systems that complex amyloid-targeting trials demand. They can obtain amyloid PET scans. They can run lumbar punctures for cerebrospinal fluid biomarker confirmation.

What they often cannot do is enroll a representative sample of the population that actually develops Alzheimer's.

A 2023 analysis published in Alzheimer's & Dementia ^[3] found that among 1,237 disease-modifying therapies tested in Alzheimer's and related dementia (ADRD) trials globally, only 11.6% had been studied in emerging economies, including just 2.0% in lower-middle-income countries. This despite the fact that low- and middle-income countries (LMICs) are projected to bear 65% of the global ADRD economic burden by 2050 ^[4]. The same concentration pattern plays out within the United States at a smaller geographic scale: researchers and clinical operations professionals who work across multiple Alzheimer's trial networks have documented that trial infrastructure congregates in established academic neurology hubs, while the communities at highest disease burden often remain structurally disconnected from enrollment pathways. A direct national mapping study of AD trial site distributions in the United States has not been published, but this observation is consistent with the broader literature on trial site concentration in academic centers ^{[5, 6]}.

The rural dimension is equally sharp. A study using national Medicare claims data found that risk-adjusted ADRD diagnostic incidence was actually higher in rural than in metropolitan counties, yet diagnosed prevalence appeared lower, strongly suggesting widespread underdiagnosis in rural areas rather than genuinely lower disease rates ^[5]. A 2025 scoping review published in Alzheimer's & Dementia: Translational Research & Clinical Interventions confirmed the pattern, noting that diagnostic gaps in underserved rural populations can distort prevalence data and confound study findings by making it appear that certain areas have lower dementia rates when the real difference is access to diagnosis ^[6].

If patients are not diagnosed, they will not be referred to trials. And if trials are not located near the communities where patients live, referral alone cannot close the gap.

The Enrollment Numbers Tell a Specific Story

The scale of the mismatch between disease burden and trial access is not abstract. A 2024 USC roundtable convening 40 cross-sector Alzheimer's experts estimated that approximately 51,000 participants were needed to fully populate the 164 Alzheimer's clinical trials then in the active pipeline ^[7]. Against that demand, roughly 11,000 Americans successfully enrolled in clinical studies annually out of approximately 90 million who might be biologically eligible for preclinical, prodromal, or mild Alzheimer's trials ^[7]. Available secondary sources suggest fewer than 1% of eligible U.S. patients are ever referred into clinical research ^[8]; the primary data underlying this estimate come from multiple operational analyses of trial recruitment pipelines reviewed in that roundtable.

The racial and ethnic picture within that enrollment funnel is sharper still. A 2025 preprint posted to medRxiv covering 88 U.S.-based Phase III Alzheimer's trials from 1997 to 2023 found that nearly half (49.3%) did not report participant race or ethnicity at all ^[2]. Among the trials that did report it, median enrollment was 0.9% for Asian or Pacific Islander participants, 4.5% for Black participants (ethnicity unspecified), 5.2% for Hispanic participants, and 0.4% for Native American participants, compared to nearly 90% for White participants ^[2]. Only 4.2% of trials conducted any subgroup analysis by race or ethnicity ^[2]. Because this analysis is a preprint awaiting peer review, these figures should be treated as preliminary, though their directional findings align with earlier peer-reviewed work documenting persistent minority underrepresentation in Alzheimer's research ^[9].

This is not primarily a failure of patient interest. JAMA Network Open published a cross-sectional analysis of 5,945 screening participants in a preclinical Alzheimer's trial finding that Black, Hispanic, and Asian participants were more likely to be recruited from local efforts than White participants, who arrived through more distributed national campaigns, and were also more likely to fail screening after the first visit, even after adjusting for demographics ^[9]. Two additional data points from conference abstracts, where results are preliminary, reinforce this picture: evidence from the ENVISION aducanumab confirmatory trial reported a 90% screen failure rate among African American and Hispanic participants versus 75% in non-Hispanic White participants ^[10], and an enrollment pipeline analysis covering 490 prescreened participants across two active AD trials found only 1.6% were ultimately enrolled, with 38% of prescreened candidates failing for medical conditions or exclusionary medications, of which nearly half were not safety-critical ^[11].

Each amyloid PET scan resulting in a screen failure carries a direct financial cost. A vendor-authored conference abstract from SiteRx described a 64% amyloid PET pass rate from patients identified through its data-driven pre-screening approach, compared to a 23% pass rate from conventional sources, with estimated PET scan cost savings of approximately $16,200 per patient ^[12]. This evidence is preliminary and vendor-reported; independent peer-reviewed validation has not been published. The directional finding, that upstream patient pre-selection quality affects PET confirmation rates, is consistent with the emerging blood-based biomarker prescreening literature.

What Geographic Misalignment Costs a Trial

Geography of site placement can materially influence these numbers. Sites serving populations with higher rates of undertreated cardiovascular comorbidities, which are more prevalent in certain demographic groups, will encounter more exclusionary comorbidity failures. Sites that rely on the same academic referral networks used for decades will draw from the same narrow population. The relationship between site placement choices and enrollment economics is plausible and operationally observed, even if its precise magnitude has not been isolated in peer-reviewed controlled comparisons.

Rural Alzheimer's: A Missed Site Opportunity

The rural gap in Alzheimer's trial access deserves specific attention because it is both large and systematically underdocumented. A peer-reviewed geospatial study published in the Journal of Transport & Health in 2026, analyzing 422,735 ADRD patients in Maryland using Healthcare Cost and Utilization Project (HCUP) data, applied Getis-Ord G* hot spot analysis and geographically weighted regression to identify ZIP code-level areas where ADRD mortality rates were disproportionately high relative to diagnosed case counts ^[13]. Eastern and southern Maryland showed zones with high mortality per ADRD patient surrounded by similarly elevated-rate zones, while central Maryland, with greater hospital concentration, showed lower death rates per patient ^[13]. The authors concluded that underserved regions with limited hospital accessibility exhibited high ADRD mortality rates despite low diagnosis rates, consistent with substantial underdiagnosis rather than lower underlying disease burden ^[13]. This is a Maryland-specific analysis; whether the same spatial patterns hold nationally requires study at broader geographic scales.

At the national level, Medicare claims data covering roughly 170 million person-years from 2008 to 2015 found that rural county residents diagnosed with ADRD experienced shorter survival compared to metropolitan county residents, a finding the researchers attributed to later-stage diagnosis rather than different underlying biology ^[5]. State-level NP practice authority compounds the diagnostic gap: a 2024 analysis found that states with restricted NP practice authority showed lower ADRD prevalence in Medicare claims data particularly in rural communities, while states granting NPs full independent practice authority showed a smaller rural-urban diagnostic gap ^[14].

For site selection teams, rural areas are not conventionally attractive. They lack dense referral networks and immediate imaging access. But they hold large populations of older adults with measurable disease burden, low rates of prior trial exposure, and a genuine unmet need for research access. Whether the infrastructure gap is an obstacle or an opportunity depends heavily on how well a sponsor can characterize the local patient population before site commitment. Critically, rural site strategies often require structural adaptations that purely academic network approaches do not: satellite phlebotomy workflows, partnerships with mobile imaging providers or regional hospitals, transportation support programs, and longer lead times for site qualification. These are solvable operational problems, but they require planning that begins at the feasibility stage, not after activation.

The LMIC Problem and What It Reveals About Site Selection Logic

At the global level, the same calculus that drives site selection toward established academic centers in high-income countries systematically excludes the regions where Alzheimer's is growing fastest.

The 2023 Alzheimer's & Dementia analysis ^[3] found that North America and Western Europe accounted for the overwhelming majority of ADRD trial sites globally. Sub-Saharan Africa, Southeast Asia, and Latin America, regions where dementia prevalence is rising sharply and where genetic and environmental risk factor profiles differ meaningfully from Northern European populations, hosted a tiny fraction of registered trial activity. Identified barriers included insufficient research infrastructure, low industry presence, regulatory obstacles, and limited access to neuroimaging ^[3].

A 2022 Lancet eClinicalMedicine analysis estimated the global economic burden of ADRD at $2.8 trillion in 2019, projected to reach $16.9 trillion by 2050, with LMICs accounting for 65% of that future burden ^[4]. Trials that do not enroll from these populations produce efficacy and safety data of limited generalizability to the patients who will ultimately carry the largest share of the disease.

This is both an equity and a scientific argument. A 2015 analysis of multi-national Alzheimer's trials published in Alzheimer's Research & Therapy found that disease progression and its measurement differed across geographic regions within the same trial, with population heterogeneity at baseline translating into real differences in outcome measurement ^[15]. When site selection consistently skews toward demographically narrow populations, sponsors may not discover these differences until post-approval.

Biomarker Eligibility as a Geographic Filter

One mechanism through which site selection quietly shapes enrollment is the interaction between amyloid eligibility criteria and population-level biomarker variation.

A 2024 peer-reviewed study published in Alzheimer's & Dementia by Molina-Henry et al. examined plasma biomarker screening data from 4,905 participants across 75 AHEAD 3-45 study sites ^[16]. The researchers found that people who identified as Black or Hispanic were less likely to meet the plasma amyloid eligibility thresholds required to continue in screening, even though they face higher risk for Alzheimer's and related dementias. Among non-Hispanic Black participants, 24.7% were plasma-eligible, compared to higher rates among non-Hispanic White participants ^[16]. The authors noted that this difference in biomarker eligibility may reflect underlying variation in amyloid accumulation patterns across populations at this early preclinical stage, and raised questions about whether eligibility thresholds developed largely in studies with predominantly White participants adequately capture disease biology across all groups ^[16]. These are active research questions; the mechanisms driving group-level differences in amyloid accumulation remain incompletely understood and warrant further study before definitive conclusions about threshold design are drawn.

For site selection strategy, the operational implication is clear regardless of mechanism: sites serving predominantly minority populations will encounter higher biomarker screen failure rates, and sponsors who do not account for this during feasibility analysis will underestimate screen failure rates and overestimate enrollment velocity. This is a solvable planning problem if it is anticipated, and a costly surprise if it is not.

Blood-based biomarker prescreening has emerged as a partial mitigation. A Phase 2 trial incorporating plasma pTau181 and PrecivityAD assays into the prescreening pipeline for PROGRESS-AD reported reducing the screen failure rate for amyloid confirmation to 9.9% ^[17]. This is conference-level evidence from a single phase 2 study and should be treated as preliminary. Deploying blood-based prescreening assays at sites serving diverse populations may improve the proportion of minority participants who reach randomization, but larger studies are needed to confirm this in practice.

Regulatory and Documentation Considerations

The regulatory context for trial diversity in Alzheimer's research involves two distinct frameworks that are sometimes conflated.

The Food and Drug Omnibus Reform Act of 2022 (FDORA), Section 3601, created the statutory framework for Diversity Action Plans (DAPs), requiring sponsors of pivotal clinical studies to submit DAPs to the Secretary of Health and Human Services ^[23]. The statute ties applicability to studies whose enrollment begins 180 days after publication of the required final DAP implementation guidance. The June 2024 FDA draft DAP implementation guidance, which was intended to satisfy that statutory requirement, was removed from the FDA website in January 2025 following a presidential executive order on DEI programs ^[19]. As of publication, no final DAP implementation guidance has been issued; the statutory mandate remains in force, but sponsors should distinguish the FDORA statutory framework from currently operative FDA guidance, and verify current DAP submission obligations with regulatory counsel rather than relying on this article alone ^[23].

Separately, on December 15, 2025, FDA finalized its guidance on "Enhancing Participation in Clinical Trials: Eligibility Criteria, Enrollment Practices, and Trial Designs" ^[18]. This guidance is not the final DAP implementation document; it addresses how sponsors should design eligibility criteria and enrollment strategies to include representative populations. Like most FDA guidance, it is not legally binding but reflects the agency's current expectations for well-conducted trials. It covers demographic characteristics (age, sex, race, ethnicity, location of residency) and non-demographic ones (comorbidities, organ dysfunction, disability) and recommends that sponsors eliminate or substantially modify overly restrictive Phase 2 exclusion criteria for subsequent Phase 3 studies ^[18].

For Alzheimer's trials specifically, both frameworks point in the same direction: a trial that enrolls predominantly from a narrow demographic and geographic band will face questions about generalizability at review, and those questions become more pointed when the approved therapy is intended for populations the trial did not adequately study. Regardless of the current status of DAP final guidance, the scientific rationale for representative enrollment has not changed.

A Data-Driven Approach to Alzheimer's Site Selection

The National Institute on Aging's Alzheimer's Clinical Trials Consortium (ACTC), a cooperative agreement between NIA and a consortium led by USC, Harvard-affiliated hospitals, and Mayo Clinic, established infrastructure at 35 sites across 24 states with an explicit mandate to improve diversity and accelerate startup for Alzheimer's trials ^[20]. As part of this infrastructure, the ACTC developed a centralized prescreening database (DART) within the AHEAD 3-45 study, collecting zip code, recruitment source, race, ethnicity, and prescreening eligibility status data to systematically understand where participants were entering and exiting the enrollment funnel before study activation ^[21].

That kind of geographic and demographic prescreening intelligence, applied before a site is activated rather than after, is the shift that operational teams increasingly need to make. Trade coverage of a Novartis pilot program presented at DPHARM 2024 described an AI-driven diversity-recruitment scoring model in which principal investigators identified as high-diversity recruiters enrolled 2.7 times more targeted Black or African American patients than peers in a large U.S. trial ^[22]. This is conference presentation coverage and has not been independently peer-reviewed, but the general principle, that PI diversity performance history is measurable and predictive of future diverse enrollment, is consistent with the broader literature on site selection variables.

The variable driving the Novartis difference was not simply geographic proximity to a diverse neighborhood, but documented PI history of proactive minority outreach. That kind of signal is visible in prior trial data and is the type of insight that structured feasibility analysis can surface before site commitments are made.

What Sponsors Should Assess Before Site Activation

For Alzheimer's trials specifically, a structured pre-activation feasibility assessment should go beyond PI publication lists and existing sponsor relationships. The table below outlines the variables that most directly predict enrollment performance in a high-screen-failure, biomarker-confirmation indication.

How Kitsa Supports Geographic Intelligence in Site Selection

KScout, Kitsa's site selection and research site intelligence product, is designed to support this kind of structured feasibility analysis. Rather than relying on historical sponsor relationships or proximity to academic centers as proxies for site quality, KScout is built to surface the demographic, geographic, and operational factors that predict whether a site can enroll the right patients for a given protocol. For Alzheimer's trials, that means layering disease burden data, prior PI performance history, imaging access, and catchment area demographics to identify where enrollment potential and population representativeness align, and where geographic access for underserved communities is built into selection criteria from the outset.

Key Takeaways

• An estimated 7.4 million Americans age 65 and older are living with Alzheimer's dementia in 2026, with Black and Hispanic older adults facing approximately 2x and 1.5x the risk of White Americans respectively, yet these groups account for a small fraction of clinical trial enrollment ^{[1, 2]}.
• A 2025 medRxiv preprint covering 88 U.S.-based Phase III Alzheimer's trials from 1997 to 2023 found that nearly half did not report participant race or ethnicity, and fewer than 5% conducted any subgroup analysis by demographic group ^[2]. These findings are preprint-stage and await peer review, but align directionally with prior peer-reviewed evidence ^[9].
• Rural communities show higher Alzheimer's disease burden in cognitive assessment data than their claims-based diagnosed prevalence rates reflect, because diagnostic access is structurally limited; geospatial analysis of Maryland found high ADRD mortality zones in areas with low hospital accessibility ^{[5, 6, 13]}.
• Only 11.6% of disease-modifying therapy trials in Alzheimer's and related dementias have been conducted in emerging economies, despite LMICs projected to bear 65% of the global economic burden of ADRD by 2050 ^{[3, 4]}.
• Plasma amyloid eligibility data from 4,905 AHEAD 3-45 study participants found that Black and Hispanic participants were less likely to meet amyloid thresholds for trial continuation, a pattern that may reflect population differences in amyloid accumulation and raises questions about threshold generalizability ^[16].
• FDORA Section 3601 created the statutory framework for Diversity Action Plans, but applicability is tied to final DAP implementation guidance that has not yet been issued; the January 2025 removal of FDA's draft DAP guidance did not repeal the statute. The FDA's December 2025 final guidance on clinical trial participation addresses eligibility and enrollment design and is the operative agency guidance on those topics ^{[18, 19, 23]}.
• Data-driven site selection that accounts for PI recruitment history, catchment demographics, biomarker eligibility risk, and community health infrastructure offers a more systematic alternative to academic-relationship-based site placement, though peer-reviewed validation of specific AI-driven approaches in Alzheimer's site selection remains limited ^[22].

References

1. [1]Alzheimer's Association. "2026 Alzheimer's Disease Facts and Figures." Alzheimer's & Dementia. 2026;22:e71345. https://doi.org/10.1002/alz.71345
2. [2]Vu T, et al. "Racial and Ethnic Reporting and Representation in Phase III Alzheimer's Disease Clinical Trials in the US, 1997-2023." medRxiv preprint, May 2025. [Preprint; not yet peer-reviewed] https://doi.org/10.1101/2025.05.03.25326933
3. [3]Llibre-Guerra JJ, et al. "A Call for Clinical Trial Globalization in Alzheimer's Disease and Related Dementia." Alzheimer's & Dementia. February 2023. https://doi.org/10.1002/alz.12995
4. [4]Bartsch P, et al. "Global and Regional Projections of the Economic Burden of Alzheimer's Disease and Related Dementias from 2019 to 2050: A Value of Statistical Life Approach." eClinicalMedicine (The Lancet). July 2022. https://doi.org/10.1016/j.eclinm.2022.101580
5. [5]Gilmore-Bykovskyi A, et al. "Rural-Urban Differences in Diagnostic Incidence and Prevalence of Alzheimer's Disease and Related Dementias." PubMed Central, 2021. https://pmc.ncbi.nlm.nih.gov/articles/PMC8277695/
6. [6]Kramer DB, et al. "Rural-Urban Disparities of Alzheimer's Disease and Related Dementias: A Scoping Review." Alzheimer's & Dementia: Translational Research & Clinical Interventions. February 2025. https://doi.org/10.1002/trc2.70047
7. [7]Boehme AK, et al. "Advancing the Science of Recruitment for Alzheimer's Clinical Trials: Challenges and Opportunities." Journal of Prevention of Alzheimer's Disease. 2025. [Review article summarizing a 2024 USC cross-sector roundtable; enrollment pipeline figures reflect expert consensus estimates from 40 participants; primary sources include Rafii et al. and U.S. Alzheimer's Clinical Trials Landscape analyses] https://pmc.ncbi.nlm.nih.gov/articles/PMC12321627/
8. [8]ACRP. "Alzheimer's Trials: Challenges, Determination, and the People Behind Progress." ACRP, September 2025. [Trade synthesis; primary sources cited include Mitchell et al. 2024 and Clement et al. 2019 for the 10-27% eligibility estimate] https://acrpnet.org/2025/09/02/alzheimers-trials-challenges-determination-and-the-people-behind-progress
9. [9]Manly JJ, et al. "Disparities by Race and Ethnicity Among Adults Recruited for a Preclinical Alzheimer Disease Trial." JAMA Network Open. 2021;4(7):e2114605. https://pmc.ncbi.nlm.nih.gov/articles/PMC8261604/
10. [10]Showell S, Murphy J, Dent G, Dawson K. "Advancing Inclusivity for African American and Hispanic Populations in Alzheimer's Disease Trials: Understanding Screen Failures from the Aducanumab Confirmatory Trial." Alzheimer's & Dementia, conference abstract, January 2025. [Conference-level evidence; findings are preliminary] https://pmc.ncbi.nlm.nih.gov/articles/PMC11713872/
11. [11]Doody RS, et al. "The Current State of the Screening and Enrollment Pipeline for Clinical Research Engagement in the Era of Disease Modifying Therapies for Alzheimer's Disease." Alzheimer's & Dementia, conference abstract, 2025. [Single-site experience reported at conference; findings should not be generalized without caution] https://pmc.ncbi.nlm.nih.gov/articles/PMC12741756/
12. [12]SiteRx. "SiteRx AI-Enabled Clinical Intelligence Platform Delivers Superior Amyloid PET Conversion in Early Alzheimer's Disease Trial." Alzheimer's & Dementia, conference abstract, January 2025. [Vendor-authored conference abstract; independent peer-reviewed validation not published] https://pmc.ncbi.nlm.nih.gov/articles/PMC11713331/
13. [13]Namadi SS, et al. "Access to Hospitals for People with Alzheimer's Disease and Related Dementias." Journal of Transport & Health. 2026. https://doi.org/10.1016/j.jth.2026.102304
14. [14]Ferdows N. "Rural-Urban Disparities in Alzheimer's Disease and Related Dementia: The Role of State-Level Nurse Practitioner Policies." Alzheimer's & Dementia, conference abstract (Gerontological Society of America annual meeting supplement), December 2024. [Conference abstract; directional findings consistent with broader Medicare claims literature] https://pmc.ncbi.nlm.nih.gov/articles/PMC12724791/
15. [15]Cummings JL, et al. "Alzheimer's Disease Progression by Geographical Region in a Clinical Trial Setting." Alzheimer's Research & Therapy. 2015. https://pmc.ncbi.nlm.nih.gov/articles/PMC4481070/
16. [16]Molina-Henry DP, Raman R, Liu A, et al. "Racial and Ethnic Differences in Plasma Biomarker Eligibility for a Preclinical Alzheimer's Disease Trial." Alzheimer's & Dementia. 2024;20:3827-3838. https://doi.org/10.1002/alz.13803
17. [17]Esquivel RN, Mathur S, Parker C, et al. "Blood-Based Biomarkers Enrich for Amyloid Positive Participants, Reducing Patient Burden in PROGRESS-AD." Alzheimer's & Dementia, conference abstract, December 2025. [Phase 2 trial; conference-level evidence; findings are preliminary] https://pmc.ncbi.nlm.nih.gov/articles/PMC12741176/
18. [18]U.S. Food and Drug Administration. "Enhancing Participation in Clinical Trials: Eligibility Criteria, Enrollment Practices, and Trial Designs." Guidance for Industry, December 15, 2025. [Final guidance superseding the November 2020 draft on the same subject; separate from FDORA's statutory Diversity Action Plan requirements under Pub. L. 117-328] https://www.fda.gov/regulatory-information/search-fda-guidance-documents/enhancing-participation-clinical-trials-eligibility-criteria-enrollment-practices-and-trial-designs
19. [19]AJMC. "FDA Quietly Removes Draft Guidance on Diversity in Clinical Trials Following Executive Order on DEI." AJMC, January 2025. https://www.ajmc.com/view/fda-quietly-removes-draft-guidance-on-diversity-in-clinical-trials-following-executive-order-on-dei
20. [20]National Institute on Aging. "Alzheimer's Clinical Trials Consortium (ACTC)." NIA, NIH. https://www.nia.nih.gov/research/dn/alzheimers-clinical-trials-consortium-actc
21. [21]Grill JD, et al. "Centralizing Prescreening Data Collection to Inform Data-Driven Approaches to Clinical Trial Recruitment." Contemporary Clinical Trials, 2023. https://pmc.ncbi.nlm.nih.gov/articles/PMC10152012/
22. [22]Clinical Trial Vanguard. "How Novartis Is Using AI for Clinical Trial Feasibility and Site Selection." December 2024. [Trade/conference coverage; not peer-reviewed; independent validation not published] https://www.clinicaltrialvanguard.com/conference-coverage/how-novartis-is-using-ai-for-clinical-trial-feasibility-and-site-selection/
23. [23]Food and Drug Omnibus Reform Act of 2022, Pub. L. No. 117-328, Section 3601, Consolidated Appropriations Act, 2023 (Dec. 29, 2022). [Statutory text; Section 3601 mandates Diversity Action Plan submission for pivotal clinical studies of drugs and devices] https://www.congress.gov/117/bills/hr2617/BILLS-117hr2617enr.pdf