Background: In the last 10 years, the PResymptomatic EValuation of Experimental or Novel Treatments for Alzheimer Disease (PREVENT-AD) research program has collected both biological and behavioral data longitudinally, using well-recognized biomarkers, to better understand and prevent Alzheimer's disease (AD). Our cohort is composed of older individuals with a family history of AD, but who were themselves cognitively unimpaired when enrolled in the study (aged 63 years old ± 5, at entry). To move AD research further, we recently shared this rich dataset with the global research community.
Method:We created a data sharing model including two levels of access based on data sensitivity and risk of potential re-identification, framed with a series of usage terms.Preparation of the datasets included selection of the variables to share, quality controls, outlier analysis, de-identification, and creation of detailed data dictionaries for proper re-usability. Each research participant was retrospectively re-consented.Result: More than 90% of participants (349 out of 386) agreed to openly share their data. Repositories are accessible openly at https://openpreventad.loris.ca, to qualified researchers at https://registeredpreventad.loris.ca, and through the unified interface of the Canadian Open Neuroscience Platform. The shared data collected from 2012 to 2020 (Table 1) includes longitudinal multimodal magnetic resonance imaging, gene variants, neuropsychological assessments, neurosensory evaluations, subjective cognitive decline information, as well as multiple behavioral data (e.g. sleep quality, neuropsychiatric factors, personality traits, etc). Amyloid and tau measurements from cerebrospinal fluid (n=106; longitudinal) and positron emission tomography (n = 130) are also available on subsamples of participants. To date, more than 250 users have accessed the PREVENT-AD datasets.
Conclusion:Creation of open datasets from sensitive human data requires technical, human, and financial investments. The challenge remains important as we must con-