Reviewed by the Help Dementia Editorial Team — our editors review every article for accuracy against guidance from the National Institute on Aging, the Alzheimer’s Association, and peer-reviewed sources.
Drug companies measure Alzheimer’s treatment success using standardized tests that assess cognitive function, behavioral changes, and functional decline—though these measures don’t always translate to meaningful improvement in daily life. The most common approach combines cognitive tests (such as the Alzheimer’s Disease Assessment Scale-Cognitive), functional scales (like Activities of Daily Living assessments), and increasingly, biological markers found in blood or cerebrospinal fluid that reflect the underlying disease process. For example, when Eli Lilly’s lecanemab showed a 35% slowing of cognitive decline over 18 months in clinical trials, that meant patients on the drug declined by roughly 25% less on the cognitive scale compared to placebo—a statistically significant but clinically modest difference that some patients experienced as meaningful while others noticed little change in their daily lives.
The shift in how companies measure success has accelerated over the past five years, moving away from relying solely on cognitive test scores toward incorporating biomarkers and real-world functional outcomes. This reflects growing recognition that a patient scoring two points higher on a 70-point cognitive test might not feel substantially better, even if the difference is statistically significant. Regulatory agencies like the FDA now expect companies to demonstrate not just cognitive slowing but some measurable impact on how patients function—whether they can still manage finances, remember appointments, or maintain independence with daily tasks.
Table of Contents
- What Cognitive Tests Measure and Their Limitations
- Biomarkers and Their Growing Role in Treatment Measurement
- Functional Outcome Measures and Real-World Impact
- Clinical Trial Design and How It Shapes Success Metrics
- The Amyloid Imaging Dilemma and What It Actually Shows
- Biomarker-Based Eligibility and the Controversy Around Who Qualifies
- Caregiver Burden and Caregiver-Reported Outcomes
What Cognitive Tests Measure and Their Limitations
The Alzheimer’s Disease Assessment Scale-Cognitive subscale (ADAS-Cog) has been the gold standard for nearly 40 years, testing memory, language, and reasoning through tasks like asking a patient to recall words after a delay or identify objects by touch. Scores range from 0 to 70, with higher scores indicating worse cognition; a 4-point improvement is generally considered clinically meaningful, though even a 3-point difference can be statistically significant in large trials. Another common measure, the Mini-Cog, takes just 3 minutes and checks for memory loss and executive function by asking patients to recall three words and draw a clock—much simpler than ADAS-Cog but less sensitive to subtle changes in mild cognitive impairment.
The limitation of these tests is that they don’t measure what patients and families actually care about. A patient might score 2 points better on the ADAS-Cog but still struggle to recognize grandchildren or manage a checkbook. These cognitive scales also have a ceiling effect in very early disease (where subtle decline is hard to detect) and a floor effect in advanced Alzheimer’s (where scores bottom out and improvement becomes unmeasurable). Additionally, cognitive tests are prone to practice effects—patients improve simply from taking the test multiple times—and they can be affected by depression, anxiety, medication side effects, and testing environment, making it difficult to isolate the true drug effect from other variables.
Biomarkers and Their Growing Role in Treatment Measurement
biomarkers—measurable biological indicators of disease—have become central to how companies now measure Alzheimer’s treatment success, because they directly reflect what’s happening in the brain before cognitive symptoms become obvious. These include amyloid-beta and tau proteins that accumulate in the brain, phosphorylated tau variants that appear in blood tests, and neuroinflammatory markers like plasma phospho-tau181 and phospho-tau217. When Eli Lilly’s donanemab showed positive results in 2023, the company pointed not just to cognitive score improvements but also to reductions in these biomarkers—a signal that the drug was actually removing amyloid plaques from the brain, visible on PET imaging.
The advantage of biomarkers is that they provide objective, measurable evidence of biological change independent of subjective cognitive testing. A patient’s blood phospho-tau level either goes down or it doesn’t; there’s less room for interpretation than “did you feel better?” However, biomarkers come with a significant caveat: clearing amyloid from the brain doesn’t guarantee the patient will feel better or function better, and it sometimes causes amyloid-related imaging abnormalities (ARIA)—swelling or microhemorrhages in the brain visible on MRI. In the lecanemab trial, about 17% of treated patients showed these asymptomatic brain changes (compared to 9% on placebo), raising questions about whether a measurable reduction in amyloid is worth the risk of silent brain injury. Additionally, many biomarkers require expensive blood tests or spinal taps that patients may not have access to, making them impractical for routine monitoring in real-world clinical practice.
Functional Outcome Measures and Real-World Impact
Beyond cognition, drug companies now measure the Activities of Daily Living (ADL) scale, which asks whether patients can bathe, dress, use the toilet, eat, and manage household tasks independently. This is what families actually care about—can Mom still live at home, or does she need round-the-clock supervision? Some trials also measure instrumental activities of daily living (IADL), such as shopping, managing finances, or taking medications on time. The Disability Assessment for Dementia (DAD) scale is another tool that combines both personal care and instrumental activities, rated by caregivers based on their observations rather than the patient’s self-report.
A critical limitation is that functional decline in early-stage Alzheimer’s is often subtle and slow, making it difficult to detect meaningful change within the timeframe of a typical 18-month trial. A patient might maintain ADL function for years even as cognition declines, so a drug that shows cognitive slowing may have no measurable impact on ADL scores during the trial period—not because it’s ineffective, but because the outcome measure isn’t sensitive enough at that stage of disease. This is why some companies have started looking at caregiver burden and quality of life as secondary measures, though these are harder to quantify and more subject to placebo effects than objective cognitive scores.
Clinical Trial Design and How It Shapes Success Metrics
Drug companies typically conduct Phase 2 and Phase 3 trials in populations at specific disease stages: preclinical (no symptoms but positive biomarkers), mild cognitive impairment, or mild dementia. This matters enormously because a drug that slows decline in mild cognitive impairment may show no benefit in mild dementia, or vice versa. For example, aducanumab showed some cognitive benefit in preclinical and mild cognitive impairment stages in early trials but ultimately failed to replicate that benefit in a larger Phase 3 trial, leading to a controversial FDA accelerated approval that was later withdrawn—a cautionary tale about how trial design and population selection directly influence whether a drug appears to work.
Trial design also influences success metrics through choices about primary versus secondary outcomes. A company might choose a cognitive test as the primary outcome (which makes regulatory approval easier to achieve) while relegating functional measures to secondary outcomes, because cognitive tests have less variability and require fewer patients to show statistical significance. This creates a perverse incentive: companies design trials to measure what’s easiest to show improvement in, not necessarily what matters most to patients. The FDA has increasingly pushed back on this, requiring companies to demonstrate clinically meaningful functional benefit, but the bar for “meaningful” remains contested and often unmet in real-world experience.
The Amyloid Imaging Dilemma and What It Actually Shows
PET imaging that visualizes amyloid plaques in the brain is now a standard part of measuring whether a drug is working biologically. Companies and the FDA accept amyloid reduction as proof of mechanism—that the drug is doing what it’s supposed to do on a molecular level. However, amyloid reduction doesn’t always correlate with cognitive improvement. Some patients clear massive amounts of amyloid and show minimal cognitive benefit, while others show modest amyloid reduction but functional improvements their families notice.
This disconnect between biomarker change and symptom change is one of the thorniest problems in Alzheimer’s drug development. There’s also the issue of amyloid-related imaging abnormalities (ARIA), which appear as microhemorrhages (ARIA-H) or brain microinfarcts and edema (ARIA-E) on MRI scans. Lecanemab trials documented these in 17% of treated patients; some were asymptomatic, others caused headaches, confusion, or vision changes. Companies and regulators accept these as acceptable trade-offs if cognitive benefit is shown, but this requires patients to undergo regular MRI screening to detect asymptomatic changes—an expensive, time-consuming burden that many real-world patients may skip or be unable to afford.
Biomarker-Based Eligibility and the Controversy Around Who Qualifies
Most modern Alzheimer’s drugs require documented amyloid positivity (either on PET or in blood biomarkers) before patients are eligible for treatment. This means a 70-year-old with cognitive complaints and a negative amyloid PET or blood test won’t qualify for lecanemab or donanemab, even if their cognitive decline is measurable and functional.
This creates a measurement problem: it excludes people whose Alzheimer’s-like symptoms might be caused by other factors (tau, neuroinflammation, vascular disease) that the current drug portfolio doesn’t target, making it harder to measure whether these drugs are truly specific to amyloid pathology or whether they work for other reasons in other populations. The biomarker requirement also reflects how drug company success is now defined by precision medicine—targeting people most likely to respond—rather than treating everyone with cognitive decline. This is more efficient for clinical trials and approval, but it narrows the populations in which companies measure and report efficacy, potentially overestimating real-world benefit for the broader population of Alzheimer’s patients, many of whom may have mixed pathology (amyloid plus tau plus vascular disease).
Caregiver Burden and Caregiver-Reported Outcomes
In recent years, drug companies have begun measuring caregiver burden—the stress, depression, and physical health impacts on the person caring for the patient. The Neuropsychiatric Inventory (NPI) captures behavioral and mood changes (agitation, depression, anxiety, hallucinations) as reported by caregivers, which is particularly relevant because behavioral symptoms often drive institutionalization more than cognitive decline alone. Some newer trials include the Zarit Caregiver Burden Interview, which directly measures how strained or depressed the caregiver feels.
The practical value of these measures is that they reflect something families experience directly—whether the behavioral chaos has gotten worse, whether the patient is easier or harder to manage. However, caregiver-reported outcomes are subjective and vulnerable to placebo effects; a caregiver who knows the patient is on a new drug may report improvement even if objective measures show none. This creates a measurement challenge: is the benefit real, or is it the caregiver’s expectation and hope? Companies often rely on blinded outcome assessors (trained staff who rate patients without knowing treatment group) to reduce bias, but for caregiver burden, blinding is impossible—the caregiver knows which treatment the patient received.
- —
