Vigilant Algorithms and the Contagion of Risk in Automated Opioid Risk Assessment

Area of Science:

• Health Informatics
• Sociology of Health and Illness
• Public Health Policy

Background:

• U.S. health systems utilize prescription monitoring interfaces with algorithmic risk scores to guide opioid decisions amid the opioid crisis.
• The impact of these tools on clinical encounters, professional authority, and accountability remains under-examined.

Purpose of the Study:

• To examine how the NarxCare risk-scoring system shapes clinical practice, professional authority, and accountability in everyday healthcare.
• To understand the implications of algorithmic automation in managing the opioid crisis.

Main Methods:

• A multi-sited ethnographic study.
• Methods included clinician interviews, interface walkthroughs, documentary analysis, and clinic observation.

Main Results:

• NarxCare presents scores as authoritative but unexaminable, collapsing diverse risks into a single number.
• Physicians become risk subjects, with their careers linked to patient scores, illustrating algorithmic automation recasting prescribers as governable agents.
• Clinicians respond by refusing, anticipating, or domesticating scores, but cannot fully evade them, indicating reconfigured rather than resolved opioid risk.

Conclusions:

• Algorithmic scoring systems like NarxCare reconfigure opioid risk management rather than resolving it.
• These dynamics have significant implications as algorithmic infrastructures expand across healthcare and scoring enters new domains.