Big data and the study of dementia: Epistemological promises and pitfalls

Abstract

This paper examines the potential contribution of big data analytical methods to our understanding of dementia and the best means of caring for those afflicted with dementia. Using Ian Hacking's history of statistics as a basis, the paper examines the three fundamental premises of big data: large datasets, a high tolerance of error and a focus on correlation rather than causation. Big data both evokes and departs from the epistemological assumptions that gave rise to statistical analysis, and its innovations merit both optimism and caution. While big data offers important means of reframing questions of dementia and dementia care along lines more conducive to the needs of patients and caregivers, the enthusiastic adoption of big data in business circles threatens to distort the practices into cost‐saving measures that reflect false efficiencies rather than genuine improvements.