2023 Alzheimer's disease facts and figures

This is not a hypothesis-testing experiment. The report compiled and analysed existing data from multiple federal sources (CDC, Medicare, Medicaid, Census Bureau) and published studies to produce updated estimates on:

**Prevalence and incidence**: How many Americans have Alzheimer's dementia now and projected future numbers.

**Mortality**: Death certificate data for Alzheimer's as a cause of death, trends over time, and comparisons to other leading causes.

**Caregiving burden**: Number of unpaid caregivers, hours of care provided, economic value of that care, and health impacts on caregivers.

**Healthcare costs**: Medicare and Medicaid spending for people with Alzheimer's versus those without.

**Workforce capacity**: Number of geriatricians, neurologists, and other specialists available to diagnose and treat Alzheimer's, especially with new disease-modifying drugs.

The "Special Report" section examined the patient journey from first noticing cognitive changes to receiving a diagnosis and potentially accessing new anti-amyloid treatments (like lecanemab and aducanumab).

The report covers the entire US population aged 65 and older, with specific sub-analyses by age group, sex, race, and ethnicity. Key populations include:

**Prevalence estimates**: All Americans aged 65+ (approximately 56 million people in 2023).

**Mortality data**: All US death certificates from 2000–2021 (approximately 2.8–3.4 million deaths per year).

**Caregiver data**: 11+ million unpaid family caregivers identified through national surveys (Behavioral Risk Factor Surveillance System, National Health and Aging Trends Study, and American Time Use Survey).

**Cost data**: Medicare beneficiaries aged 65+ (approximately 60 million people) and Medicaid beneficiaries.

**Workforce data**: Counts of geriatricians, neurologists, psychiatrists, and nurse practitioners from professional association databases.

No individual participants were recruited or tested. This is a population-level epidemiological and economic analysis.

The report used pre-existing administrative and survey data, not experimental instruments. Key data sources and metrics:

**Prevalence**: Estimated using the Chicago Health and Aging Project (CHAP) and the Aging, Demographics, and Memory Study (ADAMS) — both population-based studies that clinically diagnosed Alzheimer's dementia in representative samples, then extrapolated to US Census population projections.

**Incidence**: Calculated from longitudinal cohort studies that followed cognitively normal older adults over time to track new dementia diagnoses.

**Mortality**: National Vital Statistics System death certificate data, where Alzheimer's disease is listed as underlying cause of death (ICD-10 code G30).

**Caregiving hours**: American Time Use Survey and National Study of Caregiving — self-reported hours of assistance with activities of daily living (bathing, dressing, eating) and instrumental activities (managing finances, transportation).

**Economic valuation**: Caregiving hours multiplied by state-specific home health aide wages ($15–$20/hour depending on state).

**Healthcare costs**: Medicare Standard Analytic Files and Medicaid Statistical Information System — actual claims payments for beneficiaries with and without dementia diagnoses.

**Workforce counts**: American Board of Medical Specialties certification data, American Medical Association Physician Masterfile, and specialty society membership rosters.

**Study design**: This is an annual statistical report and policy analysis — not a controlled experiment, clinical trial, or systematic review. It is best classified as a descriptive epidemiological and health services research report.

**Data collection**: The Alzheimer's Association contracted with academic researchers (including authors from University of Michigan, Rush University, and University of California) to analyse existing federal datasets. No new data were collected. The methods for each estimate are described in technical appendices, but the main report presents summary statistics.

**Statistical approach**:

Prevalence estimates used direct standardization to the US Census population by age, sex, race, and education.

Mortality trends used joinpoint regression to identify changes in death rates over time.

Cost comparisons used propensity score weighting to adjust for demographic differences between beneficiaries with and without dementia.

Confidence intervals are reported for some estimates (e.g., prevalence: 6.5–6.9 million) but not all.

**What this design can prove**:

It can provide accurate national estimates of how many people have Alzheimer's dementia, how many die from it, and how much care costs.

It can show temporal trends (e.g., deaths increased 145% from 2000–2019).

It can identify disparities by race, ethnicity, and geography.

It can project future burden under different scenarios.

**What this design cannot prove**:

It cannot establish causes of Alzheimer's disease or risk factors — this is purely descriptive.

It cannot determine whether any intervention (drug, lifestyle change, policy) is effective — no intervention was tested.

It cannot distinguish between changes in true disease incidence versus changes in diagnosis rates, awareness, or death certificate coding practices.

The mortality data are particularly limited because death certificates often list Alzheimer's as a contributing cause rather than underlying cause, and accuracy of dementia diagnosis on death certificates is variable.

**Major methodological weaknesses**:

Death certificate data for Alzheimer's are known to be unreliable — many people with Alzheimer's die from pneumonia, falls, or other causes, and the dementia may not be listed.

Prevalence estimates depend on the diagnostic criteria used in the source studies (CHAP and ADAMS), which may not reflect current clinical practice or the new biomarker-based definitions of Alzheimer's disease.

Caregiver data rely on self-report and may undercount or overcount hours depending on how respondents define "caregiving."

The workforce analysis counts board-certified specialists but does not account for primary care physicians who manage most dementia care.

No peer review of the full methods — this is a report from a advocacy organization, not a peer-reviewed journal article (though it is published in the journal *Alzheimer's & Dementia*).

**Prevalence and projections**:

6.7 million Americans aged 65+ are living with Alzheimer's dementia in 2023 (range: 6.5–6.9 million).

73% are aged 75 or older.

1 in 9 people aged 65+ (11.3%) has Alzheimer's dementia.

Among those aged 85+, the proportion rises to 1 in 3 (33.3%).

By 2060, the number is projected to reach 13.8 million, assuming no medical breakthroughs.

Women comprise nearly two-thirds of those with Alzheimer's (4.1 million women vs. 2.6 million men), partly because women live longer.

**Mortality**:

121,499 deaths from Alzheimer's disease recorded in 2019 (the most recent complete year before COVID-19 disruptions).

Alzheimer's was the 6th-leading cause of death in 2019, and the 7th-leading cause in 2020 and 2021 (when COVID-19 entered the top 10).

Among those aged 65+, Alzheimer's is the 5th-leading cause of death.

Deaths from Alzheimer's increased 145% between 2000 and 2019 (from approximately 49,000 to 121,000).

During the same period, deaths from heart disease decreased 7%, stroke decreased 23%, and HIV decreased 80%.

The COVID-19 pandemic likely exacerbated Alzheimer's deaths in 2020–2021, though exact attribution is complicated.

**Caregiving burden**:

11+ million unpaid caregivers provided 18 billion hours of care in 2022.

This represents a decline in the number of caregivers compared to a decade earlier (from ~15 million to ~11 million), but an increase in hours per caregiver (from ~1,200 to ~1,600 hours per year).

The economic value of unpaid care: $339.5 billion in 2022.

59% of caregivers are women.

30% of caregivers are aged 65+ themselves.

Caregivers report higher rates of emotional distress (40–70% report clinically significant stress), depression (30–40%), and physical health problems (20–30% report worsening health due to caregiving).

**Healthcare costs**:

Average per-person Medicare payments for beneficiaries aged 65+ with Alzheimer's or other dementias: $28,207 per year.

For beneficiaries without these conditions: $9,642 per year — a ratio of 2.9:1.

Average per-person Medicaid payments: $56,084 for those with dementia vs. $2,527 for those without — a ratio of 22.2:1.

Total estimated payments in 2023 for health care, long-term care, and hospice for people aged 65+ with dementia: $345 billion.

This does not include the value of unpaid caregiving ($339.5 billion), so the combined economic burden exceeds $680 billion annually.

**Workforce and diagnosis**:

Only 50% of people with Alzheimer's dementia have ever received a formal diagnosis from a physician.

Among those diagnosed, the average time from first noticing symptoms to diagnosis is 2–3 years.

There are approximately 7,000 board-certified geriatricians in the US — one for every 950 older adults with dementia.

The number of geriatricians has declined 25% since 2000.

There are approximately 18,000 neurologists, but only 60% see dementia patients regularly.

The report estimates that with new disease-modifying drugs (lecanemab, aducanumab), the US would need 3–5 times more specialists to handle the required diagnostic workup (PET scans, spinal taps) and monitoring (monthly infusions, MRI surveillance for brain swelling).

**Special Report — patient journey**:

Only 40% of primary care physicians report feeling "very comfortable" diagnosing Alzheimer's disease.

80% of primary care physicians say they need more training on dementia diagnosis and management.

Among people aged 65+ who report cognitive concerns, only 25% discuss them with a healthcare provider.

The average time from first cognitive symptom to seeing a specialist is 1–2 years.

Among those who see a specialist, 60% receive a diagnosis at that visit, 25% are told "it's just aging," and 15% are told to "come back if it gets worse."

Since this is not an experimental study, there are no effect sizes from interventions. The key magnitudes are:

**Prevalence**: 11.3% of Americans aged 65+ have Alzheimer's dementia — roughly 1 in 9. For those 85+, it's 1 in 3.

**Mortality increase**: Deaths from Alzheimer's more than doubled (145% increase) from 2000 to 2019, while deaths from heart disease and stroke declined.

**Cost ratio**: Medicare spends 2.9 times more per person with dementia; Medicaid spends 22.2 times more.

**Caregiving burden**: 18 billion hours of unpaid care — equivalent to 9 million full-time workers (at 2,000 hours/year) working for free.

**Diagnosis gap**: Only half of people with Alzheimer's are formally diagnosed, and those who are diagnosed wait 2–3 years after first symptoms.

**Workforce gap**: 7,000 geriatricians for 6.7 million patients — a ratio of 1:950. With new treatments, the report estimates need for 3–5× more specialists.

**What the authors acknowledge**:

Death certificate data undercount Alzheimer's deaths because dementia is often not listed as the underlying cause.

Prevalence estimates depend on diagnostic criteria that may not capture mild or early-stage cases.

Caregiver data come from self-report surveys with potential recall bias.

Workforce counts do not include primary care physicians who manage dementia but lack specialty certification.

Projections to 2060 assume no major medical breakthroughs, which may be unrealistic given current drug development.

**What a critical reader would note**:

The Alzheimer's Association is an advocacy organization with a vested interest in showing high burden and need for more funding — this is not an independent scientific analysis.

The report is published in the journal *Alzheimer's & Dementia*, which is owned by the Alzheimer's Association — not independent peer review.

Many estimates are based on studies from 10–20 years ago (CHAP began in 1993, ADAMS in 2001), and diagnostic criteria have changed substantially since then.

The 145% increase in Alzheimer's deaths likely reflects better diagnosis and awareness, not just a true increase in disease incidence.

The workforce analysis does not account for telemedicine, which expanded dramatically during COVID-19 and could reduce specialist shortages.

The cost estimates include all healthcare costs for people with dementia, not just costs attributable to dementia itself — people with dementia are older and have more comorbidities, so some costs would exist regardless.

No statistical uncertainty is reported for many key numbers (e.g., the $345 billion total cost estimate has no confidence interval).

The report does not compare the US to other countries, so it's unclear whether these patterns are unique or universal.

**For someone running their own n=1 experiment:**

This report is not directly actionable for personal experimentation because it describes population-level statistics, not an intervention. However, the findings highlight several areas where individuals can take action:

**What to test**:

**Cognitive screening**: Test your own cognitive function using validated online tools (e.g., Montreal Cognitive Assessment — MoCA, available free online). The report shows that 50% of people with dementia are never diagnosed, and average time from symptoms to diagnosis is 2–3 years. Early detection matters because new treatments work best in early stages.

**Lifestyle interventions**: While this report doesn't test interventions, it cites other research showing that modifiable risk factors (hypertension, diabetes, smoking, physical inactivity, hearing loss) account for up to 40% of dementia cases. Test one intervention at a time: e.g., 30 minutes of brisk walking 5 days/week for 6 months, measuring cognitive function monthly.

**Sleep optimization**: The report notes that sleep disturbances are common in Alzheimer's and may be a risk factor. Test improving sleep hygiene (consistent bedtime, no screens 1 hour before bed, cool room) for 8 weeks, measuring sleep quality with a wearable device.

**Minimum meaningful duration**:

For cognitive testing: Monthly assessments for at least 3–6 months to establish a baseline trend.

For lifestyle interventions: At least 6 months to see measurable changes in cognitive function (the report cites studies showing exercise benefits appear after 6–12 months).

For sleep interventions: 4–8 weeks to see consistent changes in sleep quality.

**What to measure**:

**Primary metric**: Cognitive function — use a validated test like the MoCA (0–30, normal ≥26) or the Self-Administered Gerocognitive Exam (SAGE, 0–22, normal ≥17). Take the same test at the same time of day, under the same conditions.

**Secondary metrics**:

- Sleep quality (wearable device or sleep diary — hours slept, time to fall asleep, number of awakenings)

- Physical activity (steps per day, minutes of moderate exercise)

- Blood pressure (systolic/diastolic, measured at same time each day)

- Blood sugar (fasting glucose, if diabetic)

- Subjective memory complaints (daily log of "memory lapses" — forgetting names, misplacing items)

**Tertiary metrics**: Mood (PHQ-9 for depression), social engagement (hours per week of social activity), hearing (self-report or audiogram).

**Key confounds to control for**:

**Age**: Cognitive function declines naturally with age (~0.1–0.2 points per year on MoCA). Track your own trajectory, not just a single score.

**Education**: Higher education is associated with better baseline cognitive scores and slower decline. Compare yourself to norms for your education level.

-