Which Dementia? Many Diagnoses Never Name the Type

“Dementia” is an umbrella term, but in routine health records the most common type recorded is often no type at all. Of almost a million death certificates in England between 2013 and 2023 that recorded dementia, half named it only as unspecified dementia^[1]; among people with a dementia diagnosis in California’s Medicare fee-for-service data in 2015, the figure was 59.6%^[2].

Knowing which disease is causing someone’s dementia is not a technicality. It shapes what a person and their family are told to expect, which medicines are offered and which are best avoided, and what help will be needed as things change. This article sets out what registers, insurance claims, death records and population studies in several countries show about which types people are diagnosed with, how often the type is left unnamed, who is most often left without one, and why that matters.

Alzheimer’s disease is the most common cause of dementia, but rarely the whole story

The World Health Organization describes dementia as the result of many different diseases and injuries that damage the brain. Alzheimer’s disease is the most common, and WHO says it may contribute to 60–70% of cases; other forms include vascular dementia, dementia with Lewy bodies and the group of diseases behind frontotemporal dementia^[5]. WHO adds a caveat that runs through all of the evidence below: “The boundaries between different forms of dementia are not always clear and mixed forms often co-exist”^[5].

Population studies, where researchers assess everyone in a sample rather than relying on who has been diagnosed, give a similar picture. China’s national survey of 46,011 adults aged 60 and over, carried out in 2015–2018 across 12 provinces, estimated that 15.07 million older people in China had dementia: 9.83 million with Alzheimer’s disease, 3.92 million with vascular dementia and 1.32 million with other dementias^[6]. That is roughly two-thirds Alzheimer’s disease and a quarter vascular dementia.

Clinical records show what doctors wrote down rather than what the brain contains. The sources in the table are not directly comparable: they differ in country, year, age and setting.

In the national US Medicare study a person could carry several dementia codes over three years, so shares add to more than 100%; Lewy body dementia appeared for 5.4% and frontotemporal dementia for 1.0%^[7].

Brain studies after death show why single labels can mislead. In a community-based study of 428 older people (average age at death 91) who, in life, would have been considered for anti-amyloid treatment eligibility assessment, 94% had Alzheimer’s disease changes in the brain, but only 26% had those changes without another major pathology; 68% had Alzheimer’s changes plus at least one other^[8]. A study of 919 people with dementia autopsied through US Alzheimer’s Disease Centers in 2005–2010 found that the sensitivity of a clinical diagnosis of Alzheimer’s disease ranged from 70.9% to 87.3%, and its specificity from 44.3% to 70.8%, depending on the thresholds used^[9]. Mixed disease is common, and even specialists do not always name the cause correctly.

In routine records, “unspecified” is often the most common diagnosis

In US Medicare claims for California in 2015, 8.2% of just over 3 million beneficiaries had a dementia diagnosis, and unspecified dementia was the single most common category, at 59.6%, ahead of Alzheimer’s disease at 23.2%^[2].

England’s death certificates show the same pattern at the end of life. Between 2013 and 2023, 987,719 certificates recorded dementia as a cause of death; unspecified dementia was the most commonly recorded type, on 497,416 (50%), followed by Alzheimer’s disease (26%) and vascular dementia (19%)^[1]. A death certificate is not a diagnostic record and may omit a type known in life, so this measures how often the type travels with a person to the end, not how often it was never diagnosed.

Sweden offers a useful contrast. Its national quality registry, SveDem, was set up to improve diagnosis and care and records the type for every person entered. Of 28,722 people registered in 2007–2012, 23.5% still had unspecified dementia^[3]. The registry’s authors note that some people are too frail for a full assessment, or present real diagnostic difficulty, and receive an unspecified diagnosis pending specialist referral, and that the unspecified group “probably includes a number of AD and VaD patients”^[3].

An unspecified label is not always a failure: for some very frail people, a detailed work-up may offer little. But when an unnamed diagnosis is the norm, many people are probably not being offered the assessment that would tell them what they are living with.

Lewy body dementia is probably the type missed most often; frontotemporal dementia clusters at younger ages

For dementia with Lewy bodies, earlier studies had reported anything from zero to 22.8% of all dementia cases^[10]. A 2014 systematic review found that it accounted for 4.2% of diagnosed dementia in the community but 7.5% in secondary (specialist) care and concluded that the true figure is likely to be much higher “because DLB diagnoses are often missed”^[10]. A 2026 meta-analysis of population-based studies estimated a pooled prevalence of 352 per 100,000 people aged 65 and over, with very wide variation between studies, and said that clinically diagnosed dementia with Lewy bodies was uncommon, “likely reflecting underdiagnosis and diagnostic insensitivity”^[11].

US claims data show what that gap can look like. In a 15% random sample of Medicare fee-for-service beneficiaries in 2017–2020, 0.5% of 4,261,782 people had a code for Lewy body dementia, while 1.6% had a different dementia code alongside at least one core feature of Lewy body disease, such as parkinsonism or hallucinations^[12]. Not all of these people will have Lewy body disease, but the authors concluded it is “likely underdiagnosed overall and in certain ethnoracial groups”^[12].

Frontotemporal dementia is uncommon overall but much more prominent at younger ages. A 2016 systematic review of 26 studies found that it made up an average of 2.7% of dementia in prevalence studies including people aged 65 and over, but 10.2% in studies restricted to people under 65^[13]. A 2025 meta-analysis of 32 population-based studies estimated a crude prevalence of 9.17 per 100,000 people of all ages and judged frontotemporal dementia comparable in frequency to dementia with Lewy bodies^[14]. Both reviews found estimates varied widely. For dementia that begins before 65, see our study on young-onset dementia.

Whether the type gets named depends on who makes the diagnosis and where a person lives

In both the US and Sweden, specialist assessment goes with more specific diagnoses. In a US study following 226,604 Medicare beneficiaries first diagnosed with dementia in 2008–09, 85% were diagnosed by a doctor who was not a dementia specialist. Only 22% saw a dementia specialist within a year of diagnosis, and 36% within five years. Unspecified dementia was recorded for 33.2% of people diagnosed by non-specialists, against 21.6% of those diagnosed by dementia specialists^[15]. Use of specialist care was particularly low among Hispanic and Asian beneficiaries^[15]. In Sweden, 31.6% of SveDem registrations came from primary care units, and the registry’s authors identify a higher share of unspecified dementia among people diagnosed in primary care as “a clear difference” from specialist care^[3].

Where someone lives matters too. In England, the share of dementia death certificates naming only unspecified dementia rose from 47.5% in the least deprived fifth of areas to 52.9% in the most deprived^[1]. An earlier analysis of death records in England and Wales for 2001–2017 found people in the least deprived tenth of areas had 0.78 times the odds of an unspecified diagnosis compared with those in the most deprived tenth, and read this as a sign of “poorer access to specialist diagnosis” in deprived areas^[16]. In the US, Lewy body dementia was least often diagnosed in Black, Asian and Pacific Islander, and American Indian and Alaska Native beneficiaries^[12].

These are associations in routine data, not proof of why the gaps exist; age, other illnesses and certification practice will play a part. But they point the same way: people who are poorer, from minority ethnic groups or seen only outside specialist services are less likely to have a specific type recorded. We look at these gaps more widely in our studies on deprivation and dementia and ethnicity and dementia.

Why the type matters: medicines, safety, planning and new treatments

Clinical guidance treats the type of dementia as something that changes care. England’s NICE guideline NG97 asks specialist services to diagnose a subtype where possible once reversible causes are ruled out, to use validated criteria for each type, and to use further tests only where knowing the subtype “would change management”^[17]. It also says people should be told at diagnosis “what their dementia subtype is and the changes to expect as the condition progresses”^[17].

The same guideline shows how different the types are in practice. It recommends offering donepezil or rivastigmine to people with mild to moderate dementia with Lewy bodies; says these medicines should be considered for vascular dementia only where Alzheimer’s disease, Parkinson’s disease dementia or dementia with Lewy bodies is also suspected; and advises against offering them or memantine to people with frontotemporal dementia^[17]. It warns that for people with dementia with Lewy bodies or Parkinson’s disease dementia, antipsychotics can worsen movement symptoms and “in some cases cause severe antipsychotic sensitivity reactions”^[17]. A person whose Lewy body disease is recorded only as unspecified dementia may lose that protection. See our study on antipsychotic prescribing for how often these medicines are used.

The type also predicts the care a person will need. In California’s Medicare data, Lewy body dementia was the costliest type to Medicare on average, followed by vascular dementia; the higher costs in Lewy body dementia were explained in part by falls, urinary incontinence or infection, depression, anxiety, dehydration and delirium^[2]. Services can plan for these when they know to expect them.

The newest treatments sharpen the point. Anti-amyloid antibodies such as lecanemab were approved in the US for the early stages of “biomarker-confirmed symptomatic AD”^[4]: a specific, test-backed diagnosis of the kind many people never receive. Even among people who have one, few qualify. In the population-based Mayo Clinic Study of Aging in the US, only 19 of 237 people (8%) with mild cognitive impairment or mild dementia and raised amyloid met the lecanemab trial’s criteria, with other chronic conditions and brain-scan findings the common reasons for exclusion^[4]. In a Swedish specialist memory clinic, 6% of 410 patients met the appropriate-use criteria for lecanemab and 6% those for donanemab (the groups partly overlapping), falling to 3% once people with two copies of the APOE ε4 gene were excluded in line with European guidelines^[18].

And because mixed disease is common, clearing amyloid treats only part of the picture: in the autopsy study described above, tangles, other protein changes (LATE) and blood-vessel disease were all linked to faster decline among people who would have been considered for treatment^[8]. For the evidence on these medicines, see our study on new dementia treatments.

What the evidence asks of us

People living with dementia, and their families, should be able to know what they are living with, and guidance already says so. The evidence asks that this becomes routine rather than a matter of luck.

• Make naming the type the default, not the exception. NICE NG97 asks for a subtype diagnosis where possible and for people to be told at diagnosis what their subtype is and what to expect^[17]. Commissioners and health systems could audit how often diagnoses stay unspecified, as Sweden’s registry already makes possible^[3].
• Give people diagnosed outside specialist services a route to a specific diagnosis. In the US, most people are diagnosed by non-specialists, and their diagnoses are more often left unspecified^[15]. The study’s authors call professional development for non-specialists “urgent”^[15]; clear referral routes matter too.
• Close the gaps by deprivation and ethnicity. Unspecified diagnoses are more common in England’s most deprived areas^[1][16], and Lewy body dementia is diagnosed least often in several US minority groups^[12]. Services should monitor who receives a specific diagnosis, not just who receives any diagnosis.
• Look actively for Lewy body and frontotemporal dementia. Lewy body dementia appears to be under-recognised^[10][11], and both change which medicines are offered or avoided^[17].
• Plan new-treatment services around mixed disease. In the studies above, few people met the criteria for anti-amyloid treatment^[4][18], and many have more than one disease in the brain^[8]. Better diagnostic services should benefit everyone, not only the few who qualify for a drug.

Language: this piece uses person-first language ("people living with dementia", "people with dementia"), as most dementia organisations and people with lived experience prefer, and uses clinical terms such as "Alzheimer's disease" or "mild cognitive impairment" only where a study, dataset or diagnosis refers to that specific condition.

Why we're publishing this

Which type of dementia someone has goes a long way towards answering what comes next. We brought this evidence together so that dementia charities, carers’ groups and the organisations that support people with Lewy body, frontotemporal and vascular dementia can show, with sources, how often people are left with only the word “dementia”, and who is most likely to be left there. The case belongs to them; our part is to set out what the data show. This article is part of our Dementia & Health research library.

How to read this data

The figures come from very different sources. Population studies, such as China’s national survey, assess a sample directly and estimate how common each type is, but rely on clinical criteria applied by researchers. Registers (Sweden’s SveDem), insurance claims (US Medicare) and death certificates (England) record what clinicians wrote down: they measure diagnostic practice as much as disease. Claims data can give one person several codes over time, so shares can add to more than 100%. Autopsy studies show what is in the brain but come from research volunteers who are not typical of everyone with dementia. None of these sources is directly comparable with another, and the years differ; the Swedish registry figures cover 2007–2012 and recording may have changed since. Shares derived from counts (such as “roughly two-thirds” for Alzheimer’s disease in China) are our own arithmetic from the published figures. The evidence is thinnest for low- and middle-income countries, where few studies report types at all.

Use this data

Free to cite with attribution to Understanding Aging. Charities, carers' groups, campaigners and journalists are welcome to reuse these figures and charts. For the underlying figures as a spreadsheet, get in touch.

Important: This article summarises published research and statistics for general information. It is not medical, care or financial advice and cannot be used to diagnose, treat or predict anything about an individual. If you have concerns about your own or someone else's memory or health, please speak to a GP or another qualified health professional.

Sources

1. Zabihi S, Jackson M, Sampson EL, Banerjee S, Kenten C, Cooper C. Association of comorbidities and socioeconomic deprivation among people who died from dementia in England between 2013 and 2023: analysis of death certificates. Age and Ageing, 2026. europepmc.org/article/MED/41482962
2. Chen Y, Wilson L, Kornak J, et al. The costs of dementia subtypes to California Medicare fee-for-service, 2015. Alzheimer’s & Dementia, 2019. europepmc.org/article/MED/31175026
3. Religa D, Fereshtehnejad SM, Cermakova P, et al. SveDem, the Swedish Dementia Registry: a tool for improving the quality of diagnostics, treatment and care of dementia patients in clinical practice. PLoS One, 2015. europepmc.org/article/MED/25695768
4. Pittock RR, Aakre JA, Castillo AM, et al. Eligibility for anti-amyloid treatment in a population-based study of cognitive aging. Neurology, 2023. europepmc.org/article/MED/37586881
5. World Health Organization. Dementia (fact sheet). WHO, 2026. who.int/news-room/fact-sheets/detail/dementia
6. Jia L, Du Y, Chu L, et al. Prevalence, risk factors, and management of dementia and mild cognitive impairment in adults aged 60 years or older in China: a cross-sectional study. The Lancet Public Health, 2020. europepmc.org/article/MED/33271079
7. Goodman RA, Lochner KA, Thambisetty M, Wingo TS, Posner SF, Ling SM. Prevalence of dementia subtypes in United States Medicare fee-for-service beneficiaries, 2011–2013. Alzheimer’s & Dementia, 2017. europepmc.org/article/MED/27172148
8. Kapasi A, James BD, Yu L, et al. Mixed pathologies and cognitive outcomes in persons considered for anti-amyloid treatment eligibility assessment: a community-based study. Neurology, 2025. europepmc.org/article/MED/40825161
9. Beach TG, Monsell SE, Phillips LE, Kukull W. Accuracy of the clinical diagnosis of Alzheimer disease at National Institute on Aging Alzheimer Disease Centers, 2005–2010. Journal of Neuropathology and Experimental Neurology, 2012. europepmc.org/article/MED/22437338
10. Vann Jones SA, O’Brien JT. The prevalence and incidence of dementia with Lewy bodies: a systematic review of population and clinical studies. Psychological Medicine, 2014. europepmc.org/article/MED/23521899
11. Urso D, Giannoni-Luza S, Giannelli T, Brayne C, Ray N, Logroscino G. Incidence and prevalence of dementia with Lewy bodies: a systematic review and meta-analysis. JAMA Neurology, 2026. europepmc.org/article/MED/42113545
12. Chen WH, Guo J, Levy SA, Britton CB, Armstrong MJ. Symptom diagnosis codes informing under-diagnosis of Lewy body dementia in Medicare beneficiaries from different ethnoracial backgrounds. Journal of Alzheimer’s Disease, 2026. europepmc.org/article/MED/41460603
13. Hogan DB, Jetté N, Fiest KM, et al. The prevalence and incidence of frontotemporal dementia: a systematic review. Canadian Journal of Neurological Sciences, 2016. europepmc.org/article/MED/27307130
14. Urso D, Giannoni-Luza S, Brayne C, Ray N, Logroscino G. Incidence and prevalence of frontotemporal dementia: a systematic review and meta-analysis. JAMA Neurology, 2025. europepmc.org/article/MED/40920400
15. Drabo EF, Barthold D, Joyce G, Ferido P, Chang Chui H, Zissimopoulos J. Longitudinal analysis of dementia diagnosis and specialty care among racially diverse Medicare beneficiaries. Alzheimer’s & Dementia, 2019. europepmc.org/article/MED/31494079
16. Jitlal M, Amirthalingam GNK, Karania T, et al. The influence of socioeconomic deprivation on dementia mortality, age at death, and quality of diagnosis: a nationwide death records study in England and Wales 2001–2017. Journal of Alzheimer’s Disease, 2021. europepmc.org/article/MED/33780372
17. National Institute for Health and Care Excellence. Dementia: assessment, management and support for people living with dementia and their carers (NG97), Recommendations. NICE, 2018. nice.org.uk/guidance/ng97/chapter/Recommendations
18. Rosenberg A, Solomon A, Bonnard A, et al. Preparing for the implementation of anti-amyloid therapies in Europe: assessing real-world eligibility for lecanemab and donanemab in a Swedish memory clinic. The Journal of Prevention of Alzheimer’s Disease, 2026. europepmc.org/article/MED/41544372