Pathological Loop in Health: When algorithmic mediation displaces medical treatment

Abstract

Algorithmic mediation in health has transformed how people access, process, and validate information about their clinical condition, modifying their therapeutic decisions before, during, and after the medical consultation. This phenomenon generates a progressive epistemic displacement: the authority of the treating physician gradually yields to search engines, social media, and generative artificial intelligence.

Based on a qualitative ethnographic study conducted at a private clinic in Chile (n = 30), we identified that in 5 cases algorithmic mediation led to documentable rupture of therapeutic adherence. We term this pattern the algorithmic pathological loop in health: a feedback cycle in which the repeated search for information, enhanced by datafication, algorithmization, and platformization, consolidates erroneous representations and deteriorates therapeutic behavior.

Cross-case analysis reveals that the loop is not equivalent to misinformation in the strict sense, but involves heterogeneous mechanisms, including confirmatory search, diagnostic avoidance, platform pseudoexpertise, and family influence, that algorithmic mediation amplifies but does not unilaterally produce. We conclude that interrupting the cycle requires more than correcting erroneous information: it requires reintroducing medical information as an emotionally meaningful idea, empathetically communicated and congruent with the patient's identity.