Digital anthropology as an alternative to "big data" analysis for decision support. The case of health care

In health care, the European digital transition mandates the creation of a shared, federated data space. With each interaction with the health care system, patients and caregivers leave a trail of data, whose analysis informs the design and provision of services. While the technology to do this is new, the idea is coherent with the 20th century trend of governance as a superimposition on reality of an administrative order, evoked in James Scott’s “Seeing Like a State”. As a result, care service design optimizes not for individual patients, but for statistical abstractions (“the average diabetic”). This flies in the face of the inherent complexity of human physiology and of the relationships of care, and introduces distortions and epistemological errors. The latter generalize beyond health care, to fields such as environmental or inequality-reducing policies.

We argue that digital ethnographic methods could mitigate the problem by offering a methodological approach capable of embracing diversity. Recent advances allow for analysis of human conversations at scale as data. We illustrate the point with data from a study on community provision of health and social care, and propose an argument for the epistemological superiority of digital anthropology over big data analysis.