Electronic health records (EHR) used in dementia research in 2026 systematically capture some patients and miss others, based on where and how often someone sees a doctor. People with frequent outpatient care, commercial insurance, and higher neighborhood income show up reliably in these databases; those with more hospital stays, lower income, or Black race are documented less completely, according to Maserejian et al.
in Alzheimer's & Dementia: DADM. This gap matters because EHR data increasingly drives decisions about who gets flagged for cognitive screening, who enters clinical trials, and which disparities get studied at all. A family member or caregiver trying to understand why research findings don't always match their own experience can trace much of the mismatch to this uneven starting point.
Medical information disclaimer: This article is for general educational purposes only and does not provide medical advice, diagnosis, or treatment. Always consult a physician or other qualified health professional about symptoms, medications, tests, or treatment decisions.
Table of Contents
- Why So Many Dementia Patients Have No Cognitive Data on Record
- What Published Dementia Studies Reveal About Race and Ethnicity Reporting
- How Diagnosis Timing Differs by Race in Claims-Based Research
- What a Large 2026 EHR Study Found About Nursing Home Risk
- Who Falls Completely Outside Current Dementia Data Systems
- Frequently Asked Questions
Why So Many Dementia Patients Have No Cognitive Data on Record
In a study of 111,125 dementia patients and 30,203 Alzheimer's patients drawn from US electronic health records, only 11% and 24% respectively had any cognitive measure documented in the five years before their diagnosis, according to Maserejian et al. That means researchers trying to study early cognitive decline using EHRs are working with a small, non-random slice of patients who happened to get tested. The same study found that lacking a documented cognitive measure was not random. Black race, older age, non-commercial insurance, lower neighborhood income, more inpatient stays, and fewer outpatient visits were each independently linked to missing data, per the same Maserejian et al.
findings. People who rely on emergency rooms or hospital admissions rather than regular primary care checkups are the ones least likely to have a cognitive test on file before a diagnosis is ever made. For a reader, this means a loved one's pattern of care, not just their health, shapes whether they would ever appear in this kind of research. Someone who sees the same primary care doctor every year is far more likely to have a documented baseline than someone whose care happens mostly through hospital visits.
What Published Dementia Studies Reveal About Race and Ethnicity Reporting
A scoping review of 97 population-based dementia risk-factor cohort studies found that only 40 of them, or 41%, even reported participants' race or ethnicity, according to Krishnan et al. in BMJ Open. Fewer still, just 29 studies or 30%, confirmed they included racial or ethnic minority groups at all.
The same review found that poor demographic reporting often made it impossible to assess selection bias in a study's results. Among 337 publications examining multiethnic cohorts, only 14, or 4.2%, offered actual evidence explaining what drives racial or ethnic disparities in dementia outcomes, rather than simply adjusting statistical models for them, per Krishnan et al. That distinction matters: adjusting for race in a model tells a reader a difference exists, not why it exists. For families looking at dementia risk research, this means a majority of the foundational studies cannot tell them whether findings apply equally across racial or ethnic groups, because the studies never reported who was actually included.
How Diagnosis Timing Differs by Race in Claims-Based Research
A Health and Retirement Study analysis linked to Medicare claims found that only 24% of non-Hispanic Black participants were diagnosed with dementia within 36 months of disease onset, compared with 31% of non-Hispanic White participants, according to an analysis cited via AHCJ Health Journalism. That gap means EHR and claims-based research built on diagnosis dates captures Black patients later in their disease course than White patients, on average.
A University of Wisconsin–Madison researcher has pointed to one reason why: people at highest dementia risk, including racial and ethnic minorities and those from lower socioeconomic backgrounds, tend to have less access to primary care and specialized memory clinics, according to coverage from news.wisc.edu. Without that access, there is no visit to generate the record a later study would need. The same researcher proposed that mining the free-text clinical notes inside EHRs, rather than relying only on structured diagnosis codes, could help identify underrepresented cases that current methods miss entirely, per the news.wisc.edu reporting.
What a Large 2026 EHR Study Found About Nursing Home Risk
A 2026 study spanning two large US health systems from 1994 to 2022 followed 29,262 adults diagnosed with Alzheimer's disease who were not in a nursing home at the time of diagnosis and had at least two years of prior EHR data, according to coverage of the Communications Medicine study via Urban Health Today. This cohort design, requiring two years of prior records, means patients with thinner EHR histories were excluded from the study before it even began.
Within that cohort, 49% were admitted to a nursing home during follow-up. After statistical adjustment, women had a significantly higher risk of nursing home admission than men, with a hazard ratio of 1.061, while admission risk did not differ significantly by race or ethnicity, per the same Urban Health Today report. The absence of a racial difference here stands in contrast to the diagnosis-timing gaps found in claims data, a reminder that different EHR-based measures can tell different parts of the same story.
Who Falls Completely Outside Current Dementia Data Systems
Some people with dementia are missed not because their records are incomplete, but because no system currently counts them in a standardized way. Johns Hopkins researcher Mary Louise Pomeroy received a 2025 Impact Collaboratory Real World Data Scholars award to build an EHR-based measure identifying "unrepresented" nursing home residents with dementia, meaning those who lack a surrogate decision-maker or advance directive, according to the IMPACT Collaboratory. This group currently has no validated, scalable way to be counted in national data at all.
A separate pattern runs in the opposite direction. A London study of linked primary-care and specialist dementia records found evidence of under-documentation of dementia diagnoses in both settings, with specialist-provider data likely under-representing patients who have the most complex needs, according to the PMC-indexed study. That is the reverse of the usual selection bias seen in EHR cohorts, where sicker, more hospitalized patients are typically overrepresented rather than missed. For a reader trying to judge how representative any single dementia study is, both ends of this spectrum matter: patients without decision-making support on one side, and patients too complex for specialist records to fully capture on the other.
Frequently Asked Questions
Does having regular doctor visits make someone more likely to be included in dementia research?
Yes. Patients with more outpatient visits and fewer hospital stays were more likely to have a documented cognitive measure before diagnosis, according to Maserejian et al.
Do most dementia studies report the race or ethnicity of participants?
No. Only 41% of 97 population-based dementia risk-factor studies reported race or ethnicity, per Krishnan et al.
Are nursing home residents with dementia fully counted in national data?
Not all of them. Residents without a surrogate decision-maker or advance directive currently have no validated way to be identified in national data, according to the IMPACT Collaboratory
You Might Also Like
- UAS-CLEAR Dementia Care Research Data Priorities in 2026: Who Was Studied and Who Was Not for Adults With Memory Concerns
- Dementia Subtype Data Gaps Identified by NIA in 2026: Who Was Studied and Who Was Not for Family Caregivers
- Home-Based Dementia Care Research Priorities in 2026: Cost and Access Questions for Adult Children
