Protocol for a Scoping Review of Artificial Intelligence Applications in Palliative Care and Hospice: Investigating Data Diversity and Model Robustness (Preprint)

Abstract

<sec> <title>BACKGROUND</title> Artificial Intelligence (AI) has made significant strides in medical fields, including palliative care where its predictive models are increasingly utilized for estimating patient survival. However, the transition from proof-of-concept studies to actual clinical applications, including decision support and triage tools, remains limited. This gap is partly due to AI's propensity for reinforcing existing biases and concerns about its robustness, such as its consistent reliability and adaptability in varied clinical settings. In the context of end-of-life care, where personalized and sensitive patient care is paramount, these challenges become even more critical. Therefore, to effectively bridge AI from theoretical concepts to practical, ethically sound applications in palliative care, in-depth research and the establishment of robust ethical frameworks are indispensable. </sec> <sec> <title>OBJECTIVE</title> Our objective is to conduct a scoping review of AI applications in palliative care and hospice environments. This review will focus on data diversity, the robustness of AI models, and the ethical guidelines that govern their use. We aim to critically analyze how these AI models are reported and evaluated, identify and mitigate potential biases, and provide a detailed perspective on AI's role in enhancing end-of-life care. </sec> <sec> <title>METHODS</title> We are following a six-step methodology developed by Arksey and O'Malley for our scoping review, utilizing six distinct databases. Selected studies must be peer-reviewed and readily accessible online, with a clear emphasis on goals or outcomes pertinent to palliative care or hospice settings. The screening process will be conducted independently by at least two reviewers, and will adhere to the PRISMA-ScR guidelines. Our approach includes organizing data in a structured format and conducting a thematic analysis to uncover technical, societal, and ethical considerations. </sec> <sec> <title>RESULTS</title> A search conducted in November 2023 across several electronic databases resulted in 4614 studies, with 1083 being duplicates. We plan to compile and summarize these findings in a scoping review publication, targeted for completion by spring 2024. </sec> <sec> <title>CONCLUSIONS</title> The review's findings are expected to offer interdisciplinary insights into the application of AI in palliative care and hospice settings. These insights will be instrumental in guiding future research and in developing AI applications that are both ethically sound and effective for palliative care and hospice use. The results will be published in peer-reviewed journals and shared with stakeholders through workshops and webinars, ensuring wide dissemination and impact. </sec> <sec> <title>CLINICALTRIAL</title> N/A </sec>

Ähnliche Arbeiten