125 AI-based risk predictions: Evaluation of pilot DRAAI 5 of PU risk relying on a logistic regression wit an L2 penalty and spline transformation applied to a limited set of predictors. The rapid adoption of generative AI by the public, driven by advancements from industry, has outpaced academia in some respects. Nonetheless, the number of academic publications on AI has tripled from 88,000 in 2010 and 240,000 in 2022 [20]. However, as Van de Sande, Van Genderen [21] highlighted in a systematic review of AI-applications in intensive care units, most AI models remain confined to the testing and prototyping phase; with only a small fraction being evaluated in realworld clinical practice. Incorporating the principle of implementation science early in the development of AI has the potential to improve its adoption and integration into clinical workflows. Despite this potential, few studies have focused on using implementation science to facilitate the deployment of AI in healthcare settings [22, 23]. To our knowledge, conducting a comprehensive mixed-method process evaluation of the pilot implementation of an AI-based risk prediction model like DRAAI is a novel contribution to this field. The Study Aims This process evaluation aimed to evaluate acceptability and feasibility of the Decubitus Risk prediction Alert based on Artificial Intelligence (DRAAI) prediction model among nurses. The secondary aim was to evaluate the feasibility of the implementation plan by quantifying the implementation efforts and analyzing nurses’ responses to these efforts. The third aim was to examine fidelity to the risk predictions, which was defined as the degree to which nursed adhered to DRAAI’s risk alerts. Additionally, the timeliness of the follow-up actions based on these risk predictions was also assessed. Methods Design This process evaluation of a pilot implementation study, employing both qualitative and quantitative methods, was conducted according to the principles of Moore, Audrey [24]. Study Setting Three general wards at Erasmus MC, a tertiary university hospital in the Netherlands, were sequentially included in the study between June 2023 and September 2023.
RkJQdWJsaXNoZXIy MTk4NDMw