Abstract
Whether redesigning a wall mount or implementing artificial intelligence (AI) in a clinical setting, the hardest part of innovation is never the technology itself but instead everything already in place. The technical component of this project aims to develop a mediating component between the existing brackets in the University of Virginia’s (UVA) Health System and the new dispensers to make them compatible with the old brackets. Although it seems easy to uninstall the old mounts and install the new mounts, drilling into the walls can cause respiratory infections and is too costly and time-inefficient. The sociotechnical component of this project aims to analyze the discourse surrounding artificial intelligence in healthcare due to its increasing prevalence. Specifically, this project investigates how different stakeholders interpret AI in healthcare differently and what these differences reveal about the values, risks, and priorities at stake as AI continues to expand. Although both topics seem disparate, both address how new technologies integrate into and reshape existing systems. Both situations reveal that innovation rarely replaces older systems outright but instead requires careful analysis, adaptation, and mediation to produce sustainable and ethical change.
UVA Health treats over 29,000 patients annually and relies on friction-reducing sheets called blue tubes to safely move patients with mobility restrictions. The current blue tube dispensers are being discontinued and the new dispensers do not slot onto the old mounts. Installing the new mounts would require drilling into the walls which can generate airborne dust that contains harmful silica particulates and pathogens, posing serious risks to patients. With over 100 dispensers needing replacement, the associated costs of installation and construction make direct installation economically and functionally infeasible. The proposed solution is a cost-effective mediating component that connects the new dispenser to the existing wall mounts through a 3D-printed component that underwent rigorous tolerance checks and finite element analysis (FEA) in Autodesk Fusion and will undergo field testing to validate its safety, functionality, and usability.
Currently, a prototype of the mediating component is being 3D-printed. FEAs were run on the prototype in Fusion to ascertain it will sustain the weight of a full dispenser and gravity and to simulate the force of the dispenser being hit by intravenous poles and hospital beds, a common occurrence as noted by the client. Field testing will be tested shortly to ensure the mediating component is compatible and fulfills the client’s demands. Afterwards, the mediating component will either be manufactured in-house or outsourced to meet UVA Health’s needs.
Despite its apparent novelty, AI has been part of healthcare since the 1970s and is now widespread, leading to different perspectives from its stakeholders on its role, risks, and benefits. This study investigates how these stakeholder groups understand AI in healthcare and what values underlie their respective outlooks. Using interpretative flexibility as a sociotechnical framework, a comparative literature review will organize sources into four different groups based on the various stakeholders and compare them across three different dimensions: how each defines AI’s role, what concerns or benefits they emphasize, and what underlying values drive their opinions.
What this study found was that each stakeholder constructs a fundamentally different meaning of medical AI based on their role: physicians see it as an efficient tool, patients fear depersonalization and what happens to their data, researchers worry about algorithmic bias and inequity, and policymakers seek to regulate it. Through the lens of interpretative flexibility, these diverging interpretations confirm that the central conflict around medical AI is not if it works, but whose values will define how it is used moving forward.
