Paper Number
ECIS2026-2293
Paper Type
CRP
Abstract
In this study we examine what sociotechnical configuration is effective when implementing AI systems in healthcare. We adopt a configurational perspective using a fuzzy-set qualitative comparative analysis (fsQCA) to uncover how sociotechnical conditions shape implementation outcomes. With this approach, we holistically examine implementation processes instead of looking at isolated factors. We conducted interviews with 18 radiotherapy centers in the Netherlands, focusing on the implementation of auto-segmentation and auto-planning, two AI systems in radiotherapy. The empirical results for auto-segmentation reveal conjunctural pathways, demonstrating that presence of AI capabilities, organizational capacity and technical alignment drives success, while the absence of organizational capacity and technical alignment lead to failure. For auto-planning, the results are less comprehensive, the QCA analysis demonstrated that the lack of external pressure leads to failure. Through this study we advance our understanding of AI implementation by demonstrating that outcomes are nonlinear, asymmetric, and deeply dependent on configuring sociotechnical conditions.
Recommended Citation
Vroegindeweij, Martijn; Heising, Luca Mira; Ou, Carol; Jacobs, Maria; and van Elmpt, Wouter, "Configuring Complexity: Pathways Towards AI Implementation In Radiotherapy" (2026). ECIS 2026 Proceedings. 17.
https://aisel.aisnet.org/ecis2026/is_adopt/is_adopt/17
Configuring Complexity: Pathways Towards AI Implementation In Radiotherapy
In this study we examine what sociotechnical configuration is effective when implementing AI systems in healthcare. We adopt a configurational perspective using a fuzzy-set qualitative comparative analysis (fsQCA) to uncover how sociotechnical conditions shape implementation outcomes. With this approach, we holistically examine implementation processes instead of looking at isolated factors. We conducted interviews with 18 radiotherapy centers in the Netherlands, focusing on the implementation of auto-segmentation and auto-planning, two AI systems in radiotherapy. The empirical results for auto-segmentation reveal conjunctural pathways, demonstrating that presence of AI capabilities, organizational capacity and technical alignment drives success, while the absence of organizational capacity and technical alignment lead to failure. For auto-planning, the results are less comprehensive, the QCA analysis demonstrated that the lack of external pressure leads to failure. Through this study we advance our understanding of AI implementation by demonstrating that outcomes are nonlinear, asymmetric, and deeply dependent on configuring sociotechnical conditions.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.