Can the prevalence of Alzheimer’s disease and other forms of dementia be estimated using medical-administrative databases? A comparison with data from population-based cohorts.

Introduction: Data from the National Inter-Regime Health Insurance Information System (Sniiram) represent a potential tool for the epidemiological surveillance of Alzheimer’s disease and other dementias (ADOD). The objective of this study was to compare the prevalence of ADD treated based on Sniiram data with the results of French population-based cohort studies where possible, or, failing that, European studies. Methods: In Sniiram, individuals taking a specific AD/D medication, classified as having a long-term illness (ALD), or hospitalized for AD/D were identified among those aged 40 and older, enrolled in the general health insurance scheme, and alive as of December 31, 2014. Crude prevalence rates and rates standardized by age and sex for the French population were calculated. Among those aged 65 and older, age- and sex-specific rates were compared to estimates obtained from the French Paquid cohort by calculating prevalence ratios (PR) using Poisson regressions. Among those under 65, in the absence of French data, age-specific rates were compared to data from other European studies. Results: A total of 524,770 prevalent cases of MAAD were identified at the end of 2014 (standardized rate = 2.2%, 2.9% among women and 1.5% among men). After age 70, prevalence estimates obtained from Sniiram are on average 1.6 times lower than those expected based on the Paquid study (8.1% versus 12.1%). Before age 70, the estimates obtained from Sniiram are close to those of the Rotterdam study (RR 55–70 years = 0.7, 95% CI = 0.5–1.1), the only study selected for our comparisons in this age group. Conclusion: Prevalence data for early-onset AD obtained from the Sniiram study should be interpreted with caution and must account for age. While these data underestimate the prevalence of early-onset AD after age 70, this study highlights their potential for prevalence estimates prior to that age. The number of cases of early-onset AD can be estimated at approximately 34,500 in France (2% of people aged 40–65). Direct validation of these data remains essential.

Author(s): Carcaillon-Bentata L, Quintin C, Moutengou E, Boussac Zarebska M, Moisan F, Ha C, Elbaz A

Publishing year: 2016

Pages: 459-67

Weekly Epidemiological Bulletin, 2016, n° 28-29, p. 459-67

In relation to

Our latest news

news

2026 “Sexual Behavior” Survey (ERAS) for men who have sex with men

news

Hervé Maisonneuve has been appointed scientific integrity officer for a...

Visuel illustratif

news

Public Health France 2026 Barometer: Launch of the Survey