We screen for cancer before symptoms appear. Cognitive decline should be no different.

Editor’s note: The Alzheimer’s Association marks Sept. 21 as World Alzheimer’s Day, a global effort to raise awareness and research money, and to challenge the stigma around Alzheimer’s disease and other forms of dementia. Learn more here.

The first cognitive test most patients ever take comes after someone notices something is wrong.

By then, clinicians aren’t measuring healthy cognition but rather decline. Without knowing the patient’s baseline, it is difficult to know how much function has already been lost or how quickly it has changed.

This initial test will be the first data point in a patient’s cognitive evaluation, collected at a stage in life and in health that is no longer a healthy baseline. If the physician erroneously marks this point of decline as a healthy baseline, it becomes harder to track further cognitive decline and intervene earlier.

Practices on value-based or risk-bearing contracts can then face downstream issues, including potential emergency room (ER) visits, extra imaging and specialist referrals, all driven by further decline.

We would never wait until someone felt a breast lump before considering a mammogram, yet that is effectively how we approach cognitive health today.

Today’s testing is built around population averages

Widely adopted population-normed screening tools like the Montreal Cognitive Assessment (MoCA) lie at the heart of this challenge, and the same design flaw cuts both ways. Patients with higher than average cognitive functioning can often “outsmart” these tests, scoring within a normal range even as their cognition has already begun to decline.

At the other end, a single fixed cutoff score applied to every patient does not perform equally well across populations. A 2018 study in Alzheimer’s & Dementia: Diagnosis, Assessment & Disease Monitoring found the optimal MoCA cutoff for detecting mild cognitive impairment (MCI) was 25 among non-Hispanic White patients, 24 among Hispanic patients, and 23 among non-Hispanic Black patients, with a similar pattern for dementia-detection cutoffs (19 versus 16), showing that a single fixed score performs unevenly across racial and ethnic groups.

Both problems trace back to the same root cause: A score built around a population average, rather than an individual's own baseline, will always miss some patients in one direction or the other. Because reimbursement rewards these same quick, standardized screeners over more individualized testing, the blind spot recurs at every visit instead of being corrected once identified.

Instead, we need to start investing in personal baselines, driven by serial testing over the years.

Population norms, diagnostic blind spots and downstream costs

When building a normative database for population screening tools, researchers unknowingly include “healthy” people who may have preclinical, asymptomatic stages of neurodegenerative disease, unnoticed by loved ones. This lowers the average score and increases the variance in the “normal” group, so a patient tested in the clinic may register a normal score when initial cognitive decline has already been missed.

Populations also change over time: education standards and lifestyles shift, so a screening test normed on a sample from the 1990s or 2000s may not align well with patients in the 2020s. A physician unaware of a patient’s educational background may not be able to accurately interpret a population-normed score.

A missed or delayed diagnosis carries a measurable cost to the system, not just the individual patient or practice. A 2023 study in Alzheimer’s & Dementia found that people later diagnosed with dementia already had higher rates of ER visits, hospitalizations, skilled nursing use and home health use in the six years before diagnosis, compared with matched peers without dementia over the same span.

They also exhibited higher total health care costs. The cost burden accrues before anyone notices cognitive change or suggests a diagnosis. For many practices, but especially those operating on a value-based or risk-bearing model, this increased utilization and poorer clinical outcomes can be detrimental.

Navigating the reimbursement landscape

Current approaches to reimbursement entrench a preference for quick population-based screening over comprehensive neuropsychological batteries. Current codes allow physicians to bundle quick screening into routine visits, saving administrative time and simplifying the process.

Requesting comprehensive, ongoing testing takes longer and requires additional codes, creating friction for health care professionals and encouraging quick screening as a triage tool, even though relying on it drives the diagnostic misses and delays already discussed.

Learning from preventive approaches to cancer

To begin changing behavior in our approach to cognitive testing, we should turn to another staple of the primary care physician’s tool kit: preventive approaches to cancer. Colorectal, breast, and prostate cancer all have something cognitive health lacks –– an age-based or risk-based screening protocol that starts before any symptom appears.

The United States Preventive Services Task Force (USPSTF) recommends colorectal cancer screening for all adults starting at age 45 and breast cancer screening starting at age 40. For prostate cancer, USPSTF supports shared, individualized decision-making about prostate-specific antigen (PSA) screening for men aged 55 to 69, while specialty guidelines go further. The 2026 American Urological Association/Society for Urologic Oncology guideline recommends regular screening every two to four years from age 50, a baseline PSA as early as 45 and screening from 40 to 45 for men at increased risk.

None of these protocols waits for a family member to notice a lump or a change in bowel habits; they are triggered by age and risk factors, rather than by symptoms. Cognitive testing, by contrast, is still overwhelmingly symptom triggered, leading us to the same blind spots described above.

Cancer screening also builds in the same serial-comparison logic cognitive care needs: comparing a patient’s own results over time versus measuring them once against a population average. This is also why cancer screenings receive preventive care reimbursement support, as population-level evidence has justified the development of a proactive protocol for them. Cognitive health has comparable evidence demonstrating clinical and financial value for early detection, yet the reimbursement structure has not changed.

Similar to our approach to cancer prevention, we need to establish a cognitive baseline earlier, before the onset of aging-related diseases. This initial baseline then needs to be supported by regular screening to detect the first signs of decline before disease becomes symptomatic.

This approach benefits both patients and clinics. Earlier diagnosis enables intervention while there is still time to change a person’s trajectory, meaning fewer crisis-driven decisions, ER visits and specialist referrals, as well as less spent on imaging. Long-term data also protect physicians operating in value-based or risk-bearing models, as a documented trajectory over time makes it easier to demonstrate proactive interventions and justify clinical outcomes.

Practical steps for physicians

To realize this preventive model, we must make baseline testing standard for higher-risk patients, including those over age 50 with a family history of dementia or who raise subjective concerns, rather than waiting for a triggering event to order the first test. Because a full battery runs several hours, most primary care practices will need a referral pathway to a neuropsychologist or memory clinic for the initial baseline, reserving quick screeners for interim monitoring between batteries rather than as the baseline itself.

We must also align our retesting cadence with clinical guidance: roughly every two to five years for cognitively healthy patients (depending on personal risk factors) and annually for those with identified MCI or other concerns. We can then use the resulting trajectory as the basis for medical necessity documentation and proactive management under value-based or risk-bearing contracts.

Finally, we must raise the reimbursement gap through professional societies. Current codes still reward one-time screening over baseline-and-track approaches, and that is a policy lever physicians can push on directly.

We already know how to practice preventive medicine. We establish baselines, monitor change over time and intervene before disease becomes harder to alter. Cognitive health deserves the same standard.

A patient’s first cognitive assessment should not happen after a worried family member notices something is wrong. It should happen years earlier, when clinicians still have the opportunity to identify change, intervene and alter the patient’s trajectory.

Julie Chen, M.D., is chief medical officer of Radence, a longevity medical practice. She is a fellowship-trained integrative internal medicine physician with extensive institutional research credentials (FDA, National Institutes of Health, National Cancer Institute, Mount Sinai and others). She is a member of the Buck Institute’s President’s Circle, dedicated to advancing research in aging and longevity.