← Back to Research Papers

Beyond prevalence: A complete catchment eye health system profile using rapid assessment of avoidable blindness, a population-based eye health survey, and the first implementation of the newly develop

Authors: Sabherwal S, McCormick I, Javed M, Ramke J, B R S, Buttan S, Majumdar A, Silwal P, Chandi S, Lakhani B, Tyagi S, Deep U, Jain V, Bastawrous A
Journal: PloS one
mental health psychology open access

Abstract

There are many complex challenges in implementing AI into real-world health care settings [-]. There are technical problems with the tools themselves (eg, poor generalizability, true and false negatives and positives, and reliability). Another, arguably greater, challenge relates to how well an AI system is embedded within the local sociotechnical context of implementation [-]. AI tools do not produce benefit on their own; they must be embedded within an ensemble of knowledge, practices, and procedures that govern the use case for the tool, including the conditions under which it is likely to be effective or not. Previous studies have found that the efficacy of AI tools relied on the flow of information among clinicians, trust and perceptions of utility of the tool, and broader workflow processes (eg, Sandhu et al []). One of the major factors leading to the failed uptake of AI systems and tools is an inadequate understanding of sociotechnical aspects, especially how individuals and organizations adopt new technology []. Despite increasing numbers of studies focusing on this issue [-], there remains a limited understanding of the knowledge, practices, and procedures necessary when implementing AI tools into clinical workplaces. Yet this understanding is critical in terms of maximizing the successful implementation of such tools and planning the time, costs, and resources needed to support successful real-world implementation efforts. Thus, our objective in this study was to describe and understand the contextual challenges of implementing an AI-enabled chatbot app, ChatAI, in a large tertiary care hospital. ChatAI provided quick access to a central repository of clinical and operational information without relying on the intranet (refer to Methods for further description). We examined the feasibility, penetration, acceptability, and sustainability [] of ChatAI.