Purpose: This work provides an overview of academic articles on the application of artificial intelligence (AI) in healthcare. It delves into the innovation process, encompassing a two-stage trajectory of exploration and development followed by dissemination and adoption. To illuminate the transition from the first to the second stage, we use prospect theory (PT) to offer insights into the effects of risk and uncertainty on individual decision-making, which potentially lead to partially irrational choices. The primary objective is to discern whether clinical decision support systems (CDSSs) can serve as effective means of “cognitive debiasing”, thus countering the perceived risks. Design/methodology/approach: This study presents a comprehensive systematic literature review (SLR) of the adoption of clinical decision support systems (CDSSs) in healthcare. We selected English articles dated 2013–2023 from Scopus, Web of Science and PubMed, found using keywords such as “Artificial Intelligence,” “Healthcare” and “CDSS.” A bibliometric analysis was conducted to evaluate literature productivity and its impact on this topic. Findings: Of 322 articles, 113 met the eligibility criteria. These pointed to a widespread reluctance among physicians to adopt AI systems, primarily due to trust-related issues. Although our systematic literature review underscores the positive effects of AI in healthcare, it barely addresses the associated risks. Research limitations/implications: This study has certain limitations, including potential concerns regarding generalizability, biases in the literature review and reliance on theoretical frameworks that lack empirical evidence. Originality/value: The uniqueness of this study lies in its examination of healthcare professionals’ perceptions of the risks associated with implementing AI systems. Moreover, it addresses liability issues involving a range of stakeholders, including algorithm developers, Internet of Things (IoT) manufacturers, communication systems and cybersecurity providers.
Mind the gap: unveiling the advantages and challenges of artificial intelligence in the healthcare ecosystem
Simona Curiello;Enrica Iannuzzi;Dirk Meissner;Claudio Nigro
2025-01-01
Abstract
Purpose: This work provides an overview of academic articles on the application of artificial intelligence (AI) in healthcare. It delves into the innovation process, encompassing a two-stage trajectory of exploration and development followed by dissemination and adoption. To illuminate the transition from the first to the second stage, we use prospect theory (PT) to offer insights into the effects of risk and uncertainty on individual decision-making, which potentially lead to partially irrational choices. The primary objective is to discern whether clinical decision support systems (CDSSs) can serve as effective means of “cognitive debiasing”, thus countering the perceived risks. Design/methodology/approach: This study presents a comprehensive systematic literature review (SLR) of the adoption of clinical decision support systems (CDSSs) in healthcare. We selected English articles dated 2013–2023 from Scopus, Web of Science and PubMed, found using keywords such as “Artificial Intelligence,” “Healthcare” and “CDSS.” A bibliometric analysis was conducted to evaluate literature productivity and its impact on this topic. Findings: Of 322 articles, 113 met the eligibility criteria. These pointed to a widespread reluctance among physicians to adopt AI systems, primarily due to trust-related issues. Although our systematic literature review underscores the positive effects of AI in healthcare, it barely addresses the associated risks. Research limitations/implications: This study has certain limitations, including potential concerns regarding generalizability, biases in the literature review and reliance on theoretical frameworks that lack empirical evidence. Originality/value: The uniqueness of this study lies in its examination of healthcare professionals’ perceptions of the risks associated with implementing AI systems. Moreover, it addresses liability issues involving a range of stakeholders, including algorithm developers, Internet of Things (IoT) manufacturers, communication systems and cybersecurity providers.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
