Algorithmic Choreography: Redefining Human and Non-Human Relations in Digital Health.

This article contributes to the growing debate on algorithms in digital health, which has expanded alongside the integration of algorithmic technologies into diverse healthcare practices and contexts. Although much of the existing literature adopts a deterministic perspective, focusing on the impact of algorithms on society, we propose a different approach. Drawing on the concept of algorithmic choreography, we show how humans and algorithmic technologies "gear together" in shared performances. Based on an ethnographic study in a nursing home for people with dementia, we examine a telemonitoring algorithmic system designed to prevent adverse events and how it enters into relations with the humans inhabiting this healthcare organization. The system gives rise to two distinct choreographies: night monitoring, in which staff use it for real-time observation of residents' movements and quantified decision-making, where algorithmically generated indexes support longitudinal assessments of residents' health status. We argue that in both choreographies, humans remain central rather than passive recipients of innovation. Care professionals give or withhold agency from algorithms, sustain their functioning and engage in collective improvisations through which human-nonhuman interactions are continually reshaped.