NCHR Written Comment on HHS National Plan To Address Alzheimer’s Disease

August 14, 2026

Re: Docket (HHS-ASPE-2026-0298) “HHS Request for Comment on the Update to the National Plan To Address Alzheimer’s Disease”

We appreciate the opportunity to comment on the update to the National Plan to Address Alzheimer’s Disease and will focus our comments on several issues related to research, treatment safety, dementia caregivers, and the information patients and care partners need to make informed decisions. The National Center for Health Research (NCHR) is a nonprofit organization that conducts and analyzes research on health and medical issues. We work to ensure that medical treatments, services, and policies are based on strong scientific evidence to protect patients and families.

1. What opportunities or challenges exist in advancing research and development of interventions to prevent or treat AD/ADRD, including translating scientific progress into effective and scalable treatments and care?

Research needed to better predict AD/ADRD

Our ability to predict AD has improved as we’ve learned more about the role of amyloid-beta plaque, APOE3, APOE4, and tau. However, it is still not possible to always accurately predict who will or won’t develop Alzheimer’s disease, even for people who have already been diagnosed with mild cognitive impairment. Rather than focusing on amyloid-beta plaque as a diagnostic tool or focusing only on reducing amyloid-beta plaque as a treatment tool, HHS research, especially at the National Institute on Aging (NIA) and National Institute of Neurological Disorders and Stroke (NINDS), should focus on evaluating the contributions of APOE3, APOE4, and tau for AD/ADRD for screening, diagnosis, and treatment.

However, thanks to CMS’ requirements for registries, there are already key data that NIA and NINDS should be evaluating. To help determine if existing FDA-approved treatments are safe and effective, we strongly support Medicare’s requirement for registries for patients taking anti-amyloid drugs, because the drugs have serious risks as well as evidence of benefits, and patients, families, and healthcare providers need to know which patients are most likely to have benefits that outweigh the risks in the real world. This is essential because not all patients with amyloid-beta plaque have cognitive impairment and some will never develop dementia. In addition, we now know that patients with APOE3 or APOE4 may comprise more than 75% of Alzheimer’s patients but are also most likely to develop the brain swelling and brain bleeding known as ARIA.1 Therefore, APOE3 and APOE4 need to be evaluated in these registries and considered by their healthcare providers. Given new information on the role of tau, that should be evaluated in the registries as well.

Under the CMS Coverage with Evidence Development (CED), patients are eligible for Medicare coverage if they are enrolled in one of four registries2:

  1. CMS Registry: The Anti-Aβ mAb CED Study, Clinicaltrials.gov number: NCT060582343 
  2. ALZ-Net Registry (Alzheimer’s Association): Alzheimer’s National Registry for Treatment and Diagnostics Clinicaltrials.gov number: NCT061702684
  3. Emory University Registry: Georgia Memory Net Anti-Amyloid Monoclonal Antibody Registry Clinicaltrials.gov number: NCT059990845
  4. Beth Israel Deaconess Medical Center Registry: A Prospective Comparative Study Of Monoclonal Antibodies For The Treatment Of Alzheimer’s Disease Clinicaltrials.gov number: NCT059256216

These registries provide an important opportunity to learn how these treatments perform outside of clinical trials and which patients are most likely to benefit or be harmed. However, the usefulness of these registries for advancing research depends on whether sufficiently detailed data are available for independent analysis and comparison. None of these registries have reported results on clinicaltrials.gov and most have not publicly reported data as of August 14, 2026. To provide essential information for patients currently taking the anti-amyloid drugs or considering taking them, it is essential for NIA or NINDS to evaluate the data starting immediately, since ALZ-NET began enrollment in October 2022, the CMS study in July 2023, the BIDMC study in July 2023, and Georgia Memory Net began enrollment in March 2025, thousands of patients have already been enrolled across these studies.

The three registries that are run by private entities under CMS rules are not required under CMS’s published CED requirements to publicly release or routinely provide de-identified raw data to CMS or any HHS agencies. However, rather than wait for the private entities to analyze and publish their findings, it will be more beneficial to today’s patients, families, and healthcare providers to have more immediate access to findings that have been independently analyzed by unbiased, well-qualified experts who will be required to share all relevant information with the public. Unfortunately, when the registries were approved by CMS, they were only required to publicly report the number of participants who started and completed the study, summary results for primary and secondary outcomes, statistical analyses, and adverse events within 12 months of the study’s primary completion date. CMS’s published CED requirements do not require public release of the underlying participant-level data. That is probably why we could not find any peer-reviewed journal articles reporting clinical outcome results from any of the four registries. Although summary results from the ALZ-Net (Alzheimer’s Association) registry were made public in 2026, the “Real World Longitudinal Data From ALZ-NET: Demographics, Clinical And Safety Outcomes In Participants Receiving Beta Amyloid Targeting Monoclonal Antibodies” presentation included 3,870 enrolled patients, with only 859 patients who initiated anti-amyloid treatment at or after enrollment.7 Moreover, it included Aduhelm patients as well as Leqembi and Kisunla patients in the same analysis, even though Aduhelm has not been on the market since 2024.7 Although ALZ-NET separately reported safety data for Leqembi and Kisunla, it did not separately compare their cognitive or functional outcomes, even though patients and physicians need that information in order to make informed decisions regarding which of these treatments might be more beneficial or more dangerous.

Unfortunately, the ALZ-NET results provide very general information about the impact of the two treatments, stating that the patients on the drugs are relatively stable when comparing patients 1-6 months and 6-12 months prior to treatment to those in the first 1-6 months or 6-12 months after treatment was initiated. Patients are therefore essentially serving as their own control group before and after treatment was initiated, instead of being compared with a concurrent group of similar patients who are not taking these drugs. Since Alzheimer’s is a degenerative disease that varies in how it progresses over time, using patients as their own control group is not scientifically sound as uncontrolled pre/post comparison cannot determine whether the apparent stability was caused by treatment. A control group of never-treated patients would be much more informative if selected for similarities to the treatment patients in terms of age, sex, APOE, and other variables that were evaluated. The ALZ-NET registry apparently includes over a thousand patients who were not taking amyloid-targeting treatments, so they could have selected a control group based on similar traits.7

The 2026 presentation also stated that the ARIA rate was only 12.6%, which is much lower than in pivotal clinical trials, and that “most” were “mild/moderate.” Unfortunately, the ALZ-NET researchers did not define “most” or “mild/moderate” or “severe”. Although they analyzed overall ARIA rates by treatment and APOE4 status, they did not analyze which patients were most likely to develop moderate or severe ARIA.7

In summary, the current publicly available data are based on only one registry, only 858 patients, and does not compare risks and benefits for patients taking the two amyloid-targeting drugs still on the market or for patients with varying patient characteristics, such as age, sex, race and ethnicity, comorbidities, concomitant medications, and other clinical characteristics that could help patients and physicians make more informed treatment decisions.

Meanwhile, it is not publicly reported whether CMS has analyzed the data from the CMS registry for Leqembi and Kisunla, or if the other two registries have done so, since the information has not been made public. Regardless of the sample sizes at this point, those data need to be analyzed and made public on an annual basis starting in 2026.

The following are essential to help translate scientific progress into better treatment decisions for patients and provide important evidence about how the risks and benefits of anti-amyloid drugs differ across patients and treatments:

  1. APOE3, APOE4, and tau data should be added as a requirement in all four registries, including both current and newly enrolled patients. If not already specified in the Informed Consent documents, this would require permission from all patients.  However, if the information is not made public for individual patients, most patients would be willing to comply if the importance of the information for understanding treatment benefits and risks is clearly explained.
  2. For each of the four registries, NINDS should conduct intramural research to independently compare the risks and benefits of the different drugs used, as well as conduct subgroup analyses by genetic information, amyloid information, cognitive function, and relevant demographic information. For the three non-CMS registries, de-identified raw data will be required to be provided to NINDS on an annual basis.
  3. The CMS, Emory, and Beth Israel registries are scheduled to end in 2028. As noted above, the data should be evaluated and reported annually starting in 2026, and the registries should be extended if the data suggest the need for longer-term follow-up.

2. What barriers or challenges affect early detection and timely diagnosis of AD/ADRD?

An important challenge for early detection and diagnosis is that early cognitive impairment or biomarkers associated with AD/ADRD do not always predict whether an individual will develop cognitive impairment or dementia.

Our ability to predict AD has improved as we’ve learned more about the role of amyloid-beta plaque, APOE3, APOE4, and tau. However, they cannot reliably predict whether or when an individual will or won’t develop Alzheimer’s disease or other types of dementia in the next 5, 10, or even 15 years, even for people already diagnosed with mild cognitive impairment, because impairment may have other, reversible medical or psychosocial causes.

As biomarker testing becomes more widely used, patients and physicians need better evidence about what those results mean for an individual patient’s likelihood of developing dementia, rate of cognitive decline, and potential benefits and risks of treatment.

4. What are the most significant gaps in dementia care, services, and supports for people living with AD/ADRD and their families, caregivers and care partners, including caregiver well-being?

Protecting the Lives of Patients with ADRD

A major gap in dementia care is ensuring that agitation and behavioral symptoms are managed in ways that protect patients while also addressing the very real challenges these symptoms create for families, caregivers, and care partners.

A recent HHS IG report found that dementia patients are often given atypical antipsychotic medications such as Seroquel, Rexulti, Risperdal, and Zyprexa to calm agitation and behavior problems, even though there are numerous well-established serious risks, including sudden death, and recent research indicating that these drugs increase the risk of heart failure, pneumonia, and other very serious diseases.8,9 Most atypical antipsychotics have a boxed warning that they increase the risk of death in older adults with dementia-related psychosis, but are used anyway.11 In contrast, the FDA approved Rexulti for dementia patients with agitation in 2023, despite these concerns.

We understand that agitation and other behavior problems can be harmful to the patients and their caregivers (paid or unpaid), but it is also well known that these drugs can be used to overly sedate these patients and can be deadly. We have worked with numerous family members who found their loved ones no longer able to function because of these drugs that were prescribed in nursing homes and other inpatient facilities against the wishes of the family members. CMS and other experts have attempted to reduce the inappropriate use of these drugs, especially in nursing homes, because of their risks. Nevertheless, atypical antipsychotics, benzodiazepines, and other similar central nervous system drugs continue to be inappropriately used by approximately 25% of older patients with dementia.10

The HHS IG report is based on data prior to the FDA approval of Rexulti for nursing home patients with agitation due to dementia.  This unfortunate FDA approval decision concerns us and will inevitably make its use more prevalent in nursing homes as well as for dementia patients living with family members. Our concerns are consistent with HHS concerns about SSRI antidepressants, which clearly pale in comparison to the known danger of atypical antipsychotics for vulnerable dementia patients.

More research and stronger oversight are needed to ensure that dementia patients receive treatments with benefits that outweigh their risks and that caregivers have access to safer and effective strategies for managing agitation and behavioral symptoms.

We therefore recommend:

  • HHS should review available data on the last 5 years use of atypical antipsychotics for dementia patients, focusing on whether risks outweigh the benefits;
  • HHS should conduct or fund new research on Rexulti’s use with dementia patients, evaluating the increased use in nursing homes and family residences and the risks and benefits to patients, including fatalities;
  • HHS should evaluate evidence-based research to determine how the inappropriate use of these drugs can be reduced, especially in nursing homes that receive CMS funding. 

6. What health outcomes and care goals are most meaningful to people living with AD/ADRD and their families, caregivers, and care partners, and how should these be measured and incorporated into care over time?

For people living with AD/ADRD and their families, evidence from clinical trials and observational studies indicates that outcomes meaningful to patients are not always measured by changes in biomarkers and extend beyond performance on cognitive tests. Functional outcomes, such as the ability to maintain independence in activities of daily living, communicate effectively, and remain safely in the community, are strongly associated with quality of life and caregiver burden. In addition, clinically important outcomes include serious adverse events, hospitalizations, transition to institutional care, and mortality, all of which are routinely used in dementia research and comparative effectiveness studies. The impact of disease progression and treatment on caregiver health, stress, and time burden is also well documented in the literature as a key determinant of overall disease impact.

HHS should support the consistent use of validated, patient- and caregiver-reported outcome measures in both research and clinical care, and encourage their integration into longitudinal studies, clinical trials, and real-world evidence generation to inform treatment evaluation and shared decision-making.

8. What workforce or infrastructure challenges most affect:

a.AD/ADRD research and intervention development
b. risk reduction activities
c. public health AD/ADRD initiatives
d. delivery of healthcare and long-term care
e. support for families, caregivers, and care partners?

AD/ADRD care requires a workforce with patience, appropriate training, and sufficient time and resources to provide individualized care. This is important in all settings but is particularly important in nursing homes and other long-term care settings, where inadequate staffing and limited access to clinicians trained in dementia care can make it more difficult to manage behavioral symptoms safely and appropriately. HHS should examine workforce shortages, training needs, and barriers to providing evidence-based dementia care across healthcare and long-term care settings. Recent deportations of healthcare professionals and nonprofessionals not born in the U.S. are already causing greater staff shortages and jeopardizing the quality of care and are likely to also increase the cost of care. Greater support is also needed for family caregivers, including access to practical education, respite services, and other resources that can help them safely care for people living with dementia.

Respectfully submitted,
National Center for Health Research
Washington, D.C.

References:

  1. https://www.the-scientist.com/one-gene-influences-75-percent-of-alzheimer-s-disease-cases-73934?utm_campaign=5750943-TS_News%20Alerts_2025&utm_medium=email&_hsenc=p2ANqtz-_CKoqYx4U8Hb3XKTLDTAsF6EY-aFa33bPsT_p8Yv5VjEhK0xgx5UBob5wos_Z_ixj0Iis1_EVpJT0noRE2qU7OG7jecA&_hsmi=400855007&utm_content=400855007&utm_source=hs_email?context=pdf&id=73934#_=_
  2. https://www.cms.gov/medicare/coverage/coverage-evidence-development/monoclonal-antibodies-directed-against-amyloid-treatment-alzheimers-disease-ad
  3. https://clinicaltrials.gov/study/NCT06058234?term=NCT06058234&rank=1
  4. https://clinicaltrials.gov/study/NCT06170268?term=NCT06170268&rank=1
  5. https://clinicaltrials.gov/study/NCT05999084?term=NCT05999084%20&rank=1
  6. https://clinicaltrials.gov/study/NCT05925621?term=NCT05925621&rank=1
  7. https://www.alz-net.org/system/files/media/file/2026-05/20260320_alz-net-adpd-presentation_final_watermark.pdf
  8. Rogowska, M., Thornton, M., Creese, B., Velayudhan, L., Aarsland, D., Ballard, C., Tsamakis, K., Stewart, R., & Mueller, C. (2023). Implications of Adverse Outcomes Associated with Antipsychotics in Older Patients with Dementia: A 2011-2022 Update. Drugs & aging, 40(1), 21–32. https://doi.org/10.1007/s40266-022-00992-5
  9. Mok, P. L., Carr, M. J., Guthrie, B., Morales, D. R., Sheikh, A., Elliott, R. A., … & Ashcroft, D. M. (2024). Multiple adverse outcomes associated with antipsychotic use in people with dementia: population based matched cohort study. bmj, 385.
  10. Yang, A. W., Leng, M., Ly, D. P., Tseng, C. H., Sarkisian, C., Damberg, C. L., … & Mafi, J. N. (2026). Prescribing patterns of potentially inappropriate CNS-active medications in older adults. JAMA, 335(6), 544-546.
  11. https://www.accessdata.fda.gov/drugsatfda_docs/label/2026/205422s017lbl.pdf