Who decides what is healthy? Algorithmic classification and medical normativity

Area of Science:

• Medical Sociology
• Science and Technology Studies
• Bioethics

Background:

• Algorithmic classification systems are increasingly used in healthcare.
• This raises critical questions about the definition and enforcement of medical norms.
• Existing approaches to algorithmic fairness often overlook social and ethical dimensions.

Purpose of the Study:

• To conceptualize algorithmic normativity as a multidimensional phenomenon.
• To link structural inequities with diagnostic identity fragmentation.
• To propose a human-centric governance framework for algorithmic medicine.

Main Methods:

• Conceptual analysis drawing from medical sociology, STS, and bioethics.
• Examination of training data selection, loss function design, and deployment contexts.
• Integration of transparency, plural normativities, and patient participation.

Main Results:

• Algorithmic normativity operates through multiple dimensions of system development and deployment.
• Algorithmic systems actively construct medical norms, reflecting and reproducing social inequities.
• A novel link is established between population-level inequity and diagnostic identity fragmentation.

Conclusions:

• Algorithmic medicine's normative frameworks are socially constructed, not objectively derived.
• Purely technical solutions for algorithmic fairness are insufficient.
• Democratic engagement is essential to address the ethical, social, and political aspects of algorithmic health standards.