Almost Ready to Go Live: Turf Wars, Infrastructural Barriers, and Unpredictability in Clinical AI-based Predictions of Sepsis and Organ Failure
Year: 2026
Type: conference-abstract
Abstract: This paper reports findings from ongoing research on the development of an AI-based application for identify-ing patients at risk of sepsis and organ failure. The study responds to the call to identify and understand the challenges of clinical AI integration. The study is based on interviews with stakeholders, meeting observations, and document research to unfold “the imagined future, unstable present, and structured past” (Vepřek and Carlson, 2026). The project has been on the verge of “going live” as a trial implementation in clinical practice at three different hospital departments for several years. Retrospectively, we follow the project team’s challeng-ing and winding work to reach this critical stage of the project: gaining access to real-time data from the region-al EHR, developing the algorithm, obtaining needed devices, fine-tuning the volume of notifications, engaging the clinicians, and overcoming uncertainties. The analysis unfolds this process, and the important roles that project members, private vendors, the healthcare organization, and science play in bringing the AI solution into being. Drawing on an infrastructure perspective, we conceptualize these so-called barriers as ‘infrastruc-tural frictions’ (e.g., Edwards, 2019) to start discussing how we can understand organizational turf wars, insti-tutional boundaries, and infrastructural ownership as trading zones that necessitate new hybrid roles and co-operations.
Language: English
Subfield: Management Information Systems
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7201933446