AI-Assisted Decision-Making in Long-Term Care: Qualitative Study on Prerequisites for Responsible Innovation (Preprint)

• DOI: 10.2196/preprints.55962
• Publication Date: 2024-01-03T20:48:31Z
• Authors: Dirk R M Lukkien; Nathalie E Stolwijk; Sima Ipakchian As...; Bob M Hofstede; Henk Herman Nap; Wouter P C Boon; Alexander Peine; Ellen H M Moors; Mirella M N Minkman
• Keywords: Thematic Analysis; Brainstorming; Preprint

<sec> <title>BACKGROUND</title> Although the use of artificial intelligence (AI)–based technologies, such as AI-based decision support systems (AI-DSSs), can help sustain and improve quality efficiency care, their deployment creates ethical social challenges. In recent years, a growing prevalence high-level guidelines frameworks for responsible AI innovation has been observed. However, few studies have specified embedding AI-DSSs, in specific contexts, nursing process long-term care (LTC) older adults. </sec> <title>OBJECTIVE</title> Prerequisites AI-assisted decision-making practice were explored from perspectives nurses other professional stakeholders LTC. <title>METHODS</title> Semistructured interviews conducted with 24 professionals Dutch LTC, including nurses, coordinators, data specialists, centralists. A total 2 imaginary scenarios about AI-DSSs developed beforehand used to enable participants articulate expectations regarding opportunities risks decision-making. addition, 6 principles probing themes evoke further consideration associated using Furthermore, asked brainstorm possible strategies actions design, implementation, address or mitigate these risks. thematic analysis was performed identify prerequisites this area. <title>RESULTS</title> The stance on is not matter purely positive negative but rather nuanced interplay elements that lead weighed perception Both identified relation early identification needs, guidance devising strategies, shared decision-making, workload work experience caregivers. To optimally balance seven categories identified: (1) regular deliberation collection; (2) balanced proactive nature AI-DSSs; (3) incremental advancements aligned trust experience; (4) customization all user groups, clients caregivers; (5) measures counteract bias narrow perspectives; (6) human-centric learning loops; (7) routinization AI-DSSs. <title>CONCLUSIONS</title> could turn into drawbacks depending shaping design Therefore, we recommend considering balancing act. Moreover, interrelatedness prerequisites, call various actors, developers users cohesively different factors important practice.