Yes, screening tools can and do cause overdiagnosis. A screening test identifies disease in asymptomatic people, but it cannot always distinguish between pathology that will cause harm and pathology that never will. When a screening tool detects a biological finding—say, amyloid plaques on a PET scan, or a cognitive score slightly below normal—it often cannot predict whether that finding will progress to symptoms, remain stable for years, or resolve entirely. This creates a cascade: the person receives a diagnosis they might never have needed, begins monitoring or treatment based on that label, and experiences years of anxiety tied to a condition that may never affect their daily life. A patient undergoing screening for mild cognitive impairment might score in the impaired range on a test like the Montreal Cognitive Assessment, triggering a diagnosis, referrals, and prescriptions—yet never develop dementia and die from unrelated causes with their cognitive function intact.
The risk of overdiagnosis increases when screening tools are applied to large populations without clear symptoms, when the condition being screened progresses slowly or unpredictably, and when the test cannot distinguish between people who will deteriorate and those who will not. Screening is not the same as diagnosis. A diagnosis should reflect the presence of disease with a reasonable expectation of harm or dysfunction. Screening tools, however, often flag biological or neuropsychological abnormalities without that certainty. The gap between what a test detects and what will actually cause illness is where overdiagnosis lives.
Table of Contents
- Why Do Screening Tools Create Overdiagnosis Risk?
- The Silent Burden of Preclinical Labels
- Specific Screening Tools and Their Overdiagnosis Patterns
- Evaluating Screening Results—What Questions to Ask
- Regulatory and Professional Variability in Screening
- Amyloid and Tau Biomarkers—Overdiagnosis at the Preclinical Level
- Cognitive Aging Versus Pathological Decline—The Critical Distinction
- Frequently Asked Questions
Why Do Screening Tools Create Overdiagnosis Risk?
Screening tools are designed to be sensitive—to cast a wide net and catch as many cases as possible before symptoms appear. Sensitivity and specificity represent a trade-off. A screening test that is highly sensitive catches many people with the condition but also generates false positives, flagging people without the condition or without clinically meaningful disease. The Montreal Cognitive Assessment, for example, has a sensitivity around 90% for detecting mild cognitive impairment—meaning it catches 90 out of 100 people who have been diagnosed with MCI—but its specificity is lower when applied to diverse populations, producing false positives especially in people with lower education, non-native language speakers, or those with depression or sleep disorders that mimic cognitive decline. A person flagged by the test may not have a progressive disease at all.
Furthermore, screening tools detect prevalence, not incidence. They capture existing abnormalities at one moment in time, not the likelihood of future disease. A brain amyloid PET scan in a cognitively normal 65-year-old might reveal amyloid plaques—a hallmark of Alzheimer’s pathology—yet 30% of cognitively normal amyloid-positive people never develop cognitive decline in their remaining lifespan. Screening identified a biological finding, but not a disease that would cause harm. The person may then be labeled as having “preclinical Alzheimer’s disease,” enrolled in monitoring protocols, or offered preventive medications, all based on a finding that may never progress to symptoms.
The Silent Burden of Preclinical Labels
When screening detects early or asymptomatic disease—pathology without symptoms—it creates a diagnostic category that sits in diagnostic limbo. These findings lack clear natural history data. Unlike symptomatic Alzheimer’s disease, where progression is documented, preclinical amyloid positivity or mild cognitive impairment may stabilize or resolve, yet we cannot predict which individuals will deteriorate. The burden of this ambiguity falls on the screened person. They carry a diagnosis without symptoms, experience worry about future decline, and may make life decisions—career changes, retirement timing, family planning—based on a label that may never manifest as disease. One limitation of screening in cognitive aging is that normal aging itself involves cognitive changes. Processing speed slows, some memory becomes less efficient, and retrieval takes longer. Screening tools often use age-adjusted norms, but these norms are population averages; many normal older adults score below the cutoff simply due to normal variation in aging, not pathology.
A 75-year-old scoring in the low range on memory tests may be within normal aging variation, yet be flagged for cognitive impairment. The psychological impact of overdiagnosis from screening is substantial and often underestimated. A person told they have mild cognitive impairment or are at high risk for dementia reports increased anxiety, depression, and hypervigilance about memory. They become more likely to interpret normal forgetting as a sign of decline. They may avoid cognitively demanding activities, fearing they will accelerate decline, which ironically can worsen cognitive function through disuse. There is also the issue of test-retest variability. Cognitive screening scores fluctuate based on mood, sleep, medication side effects, and test-related anxiety. A single low score may not reflect stable cognitive impairment but rather a temporary state. Repeating screening tests within a short window sometimes shows improvement back to normal ranges—a finding inconsistent with progressive disease but consistent with noise in the measurement.
Specific Screening Tools and Their Overdiagnosis Patterns
The Mini-Cog is widely used in primary care to screen for cognitive impairment. It combines a 3-item recall task with clock drawing and takes 3 minutes. Its simplicity makes it attractive for population screening, but this brevity is also its limitation. It captures global cognition coarsely and cannot detect mild or circumscribed cognitive changes. A person with isolated memory loss or executive dysfunction may score normally on the Mini-Cog, leading to false reassurance, or conversely, a person with normal cognition and depression might score poorly, leading to false alarm. The sensitivity and specificity vary dramatically by population studied, education level, and how investigators defined the disease they were screening for. When sensitivity and specificity come from different populations than yours, the test’s performance in your setting is unpredictable.
Neuropsychological batteries—comprehensive, multi-hour testing by specialists—are more detailed but also prone to overdiagnosis through a different mechanism: multiple comparisons and regression to the mean. A neuropsych battery may include 15 to 20 cognitive tests. If normal variation is assumed and a cutoff of 1.5 standard deviations below the mean is used for impairment, statistical probability guarantees that some test scores will fall into that range purely by chance, not by true impairment. A person with normal overall cognition will typically have some scattered low scores just due to individual strengths and weaknesses. Neuropsychologists are trained to integrate results and distinguish genuine impairment from statistical noise, but the process is subjective. Two equally qualified neuropsychologists may reach different conclusions from the same test battery. The diagnosis becomes partly dependent on the examiner, not solely on objective pathology.
Evaluating Screening Results—What Questions to Ask
When a screening test suggests cognitive impairment, it is essential to understand what the result actually means. Ask your clinician: Is this test validated in my age group and population? A screening tool validated in a university neurology clinic may not perform the same way in a primary care office or in a different ethnic or education background. Does a positive result mean I have disease, or does it mean I need further evaluation? Screening is not diagnosis. A positive screening should prompt comprehensive evaluation, not immediate diagnosis or treatment. What is the test’s positive predictive value in my risk category? The PPV tells you, of all people who screen positive, how many actually have the condition. A test with high overall sensitivity can still have low PPV if the condition is rare in your population. If 100 people screen positive, how many will actually have progressive disease? This is the question that distinguishes screening from overdiagnosis.
Request repeated testing if the initial result is borderline or ambiguous. A single cognitive screen, especially if you were tired, stressed, or not feeling well that day, is not a stable measurement of your cognitive status. Cognitive screening combined with functional history is more informative than testing alone. Can you manage finances, medication, and daily living independently? Have people who know you noticed cognitive decline? An abnormal cognitive test in the absence of functional decline is overdiagnosis—identifying impairment that has no real-world consequence. Clinicians should compare current results to any prior cognitive testing you may have had. Stable or slowly changing cognitive test scores over years is reassuring, even if absolute scores are low. Progressive decline over months is concerning. Without prior baseline, a single score is difficult to interpret.
Regulatory and Professional Variability in Screening
The clinical utility of cognitive screening varies by professional guidelines. The U.S. Preventive Services Task Force (USPSTF) does not recommend routine cognitive screening in asymptomatic older adults due to insufficient evidence that screening improves health outcomes. However, other organizations, including some dementia societies and pharmaceutical companies, promote cognitive screening as a standard practice. This discrepancy reflects genuine uncertainty about whether finding cognitive impairment early, in the absence of symptoms, leads to benefit. Early diagnosis may allow time for planning and family communication, but it does not currently slow or stop Alzheimer’s disease in preclinical stages with high confidence. A person screened and labeled with mild cognitive impairment experiences potential harms—worry, medicalization, unnecessary treatment—with unclear offsetting benefits. Different screening cutoffs produce different rates of case-finding. Lower cutoffs increase sensitivity and catch more people but also increase false positives.
Higher cutoffs reduce false positives but miss true cases. The choice of cutoff is often arbitrary or driven by research definition, not by clinical utility. A clinic might use a Montreal Cognitive Assessment cutoff of 26, another might use 25. This single-point difference could shift a person from “normal” to “impaired,” yet neither cutoff is objectively correct. The cutoff chosen determines prevalence, costs, and how many people are diagnosed. There is also professional variability in interpretation. Some clinicians integrate cognitive screening into broader clinical assessment; others rely heavily on test scores. Some clinicians counsel patients thoroughly about limitations and next steps; others present screening results as definitive diagnosis. The experience of overdiagnosis depends partly on how results are communicated.
Amyloid and Tau Biomarkers—Overdiagnosis at the Preclinical Level
Blood biomarkers for Alzheimer’s pathology—phosphorylated tau and phosphorylated tau-181, amyloid-beta ratios, and neurofilament light—can now be measured and are increasingly used to identify preclinical Alzheimer’s disease in cognitively normal people. The appeal is clear: find pathology before symptoms, perhaps intervene early. However, biomarker positivity carries significant overdiagnosis risk. Many cognitively normal amyloid-positive people will never develop cognitive decline.
Tau positivity adds information about brain degeneration, but tau-positive cognitively normal people still have heterogeneous outcomes; some decline rapidly, others never decline. Lecanemab, an amyloid-lowering monoclonal antibody approved for early symptomatic Alzheimer’s disease (mild cognitive impairment or mild dementia with confirmed amyloid pathology), showed a slowing of cognitive decline by 27% over 18 months—a modest effect. Extrapolating this to asymptomatic biomarker-positive people is unproven. Giving treatment with cognitive and infusion-related side effects to people who may never develop symptoms is a form of overtreatment driven by overdiagnosis.
Cognitive Aging Versus Pathological Decline—The Critical Distinction
Normal aging involves measurable cognitive changes. Processing speed declines with age due to changes in white matter and neural timing. Episodic memory—memory for specific events—becomes less efficient, though semantic memory and crystallized intelligence often remain stable or improve. These changes are present in the majority of cognitively healthy older adults and do not inevitably progress to dementia. The distinction between normal aging and pathological cognitive decline is critical but difficult to make from screening tests alone. A person whose memory complaints reflect normal aging but whose screening test score falls in the low range receives a label of mild cognitive impairment, medicalizing a normal process. True pathological decline progresses despite maintained cognitive reserve and continues to worsen over months and years.
Normal aging variation is stable or changes very slowly. Cognitive reserve—the brain’s ability to tolerate pathology without manifesting symptoms—also contributes to heterogeneous outcomes after screening. Two people with identical amyloid plaques or identical cognitive test scores may have very different functional trajectories. A person with higher education, ongoing cognitive engagement, and stronger social networks may tolerate more pathology before showing decline. Screening cannot measure cognitive reserve, yet it is a powerful predictor of real-world outcomes. A cognitively normal person with high cognitive reserve who screens positive for amyloid has a different prognosis than a person with low cognitive reserve with the same biomarker finding. Screening results, presented without context about reserve and normal aging variation, obscure this crucial variation in individual trajectories.
Frequently Asked Questions
Does a positive cognitive screening test mean I have dementia?
No. A positive screening indicates you should receive further evaluation, not that you have dementia. Cognitive screening is a first step, not a diagnosis. Many people with positive screens have no cognitive impairment or have cognitive changes from normal aging, depression, or sleep problems.
If my screening test is normal, am I certain not to develop dementia?
A normal screening test does not guarantee you won’t develop dementia in the future. Cognitive decline is gradual, and screening captures a single moment in time. Normal screening results are reassuring, but they don’t predict future risk with certainty.
Should asymptomatic older adults routinely undergo cognitive screening?
Current major guidelines, including the U.S. Preventive Services Task Force, do not recommend routine screening in asymptomatic people. Screening may be appropriate if you have concerns about memory or thinking, a family history of dementia, or other risk factors. Discuss with your clinician whether screening is right for you.
What should I do if my screening test shows mild cognitive impairment?
Ask your clinician to clarify whether the result reflects genuine cognitive impairment or normal variation. Request functional assessment—are you managing finances, medication, and daily tasks independently? Ask about test-retest reliability and whether repeating the test would be helpful. Request baseline comparison if you have prior testing.
Are blood biomarkers for Alzheimer’s more accurate than cognitive screening?
Blood biomarkers detect Alzheimer’s pathology (amyloid and tau) with good accuracy, but detecting pathology is not the same as diagnosing disease. Many cognitively normal people with Alzheimer’s pathology never develop cognitive decline. Biomarker results should be interpreted alongside cognitive testing and functional assessment, not in isolation.
What is the difference between normal aging and mild cognitive impairment?
Normal aging involves gradual slowing of processing speed and some changes in memory retrieval. Mild cognitive impairment is more pronounced decline that goes beyond normal aging but does not interfere with daily function. The boundary is sometimes subjective, which is why evaluation by a specialist is important if screening results are ambiguous.
