Miguel_2026_Age.Ageing_55_

OBJECTIVE: To estimate the proportion of dementia underdiagnosis in Brazil and assess variations by age, sex and macro-economic regions. METHODS: Diagnosed dementia cases were estimated using Brazilian Unified Health System (SUS) data on prescriptions for acetylcholinesterase inhibitors (AChEIs), as a proxy for recognised Alzheimer's disease (AD) cases. Estimates were adjusted for the proportion of AD among all dementias, treatment coverage and disease stages. Expected dementia prevalence was derived from national estimates. Underdiagnosis was calculated by comparing diagnosed to expected cases, stratified by sex, age group and region. Uncertainty interval (UI) bounds were calculated applying +/-1 SD of the mean Delphi consensus prevalence for each stratum. Estimates were conducted for 2024 and repeated for 2019 and 2022 for pre- and post-COVID-19 pandemic comparison. RESULTS: The overall proportion of dementia underdiagnosis among individuals aged <=60 in 2024 was 84.3% (UI: 81.5-86.4). Proportions were higher among men (86.3%, UI: 83.5-88.2) than women (83.0%, UI: 79.9-85.2). Underdiagnosis peaked at 93.2% (UI: 92.3-93.7) in the 60-64-year age group, declined to 78.9% (UI: 78.1-79.6) among 85-89 years, then rose to 86.2% (UI: 85.7-86.7) among those <=90 years. Proportions were highest in the poorest regions (North, Northeast and Central-West: 93.0%, UI: 91.4-94.1) and lowest in the richest (South and Southeast: 78.3%, UI: 74.3-81.2). Pre-pandemic (2019) underdiagnosis was slightly lower (81.6%, UI: 78.3-84.0) than in 2022 (84.9%, UI: 82.2-86.9) and 2024. CONCLUSION: Dementia underdiagnosis in Brazil is high and unequally distributed, worsening after the COVID-19 pandemic. Expanding diagnostic access is critical to address disparities and improve care.