Barriers to Artificial Intelligence Adoption in Healthcare Management: A Systematic Review

Abstract

Purpose: Humanity has been able to claim the supremacy over other creations through the intelligence it has been granted. Mimicking this intelligence by computers is artificial intelligence (AI), the oldest branches of computer science (Russel & Norvig, 2010). By the very nature of a lifeless artifact mimicking human intelligence and often exceeding it in that regard, artificial intelligence has been of great interest to many for several decades. Numerous scientists have worked laboriously towards bringing artificial intelligence into healthcare. In 2017, US healthcare spending reached $3.5 trillion (Centers for Medicare & Medicaid Services, 2018). AI has several applications in healthcare with the ability to diagnose quickly and reliably, provide data driven prognosis, some power to predict future health based on current trends, discover new drugs, and reduce costs. The present work studies the barriers to adoption of artificial intelligence in the healthcare industry. Methodology: A systematic review of qualitative and quantitative research examined the existing evidence regarding artificial intelligence in healthcare sector. A configurative analysis and thematic synthesis were conducted of mixed methods case studies. Research limitations: This cross-sectional research may be biased towards only those cases published recently in the literature. Not all studies pertinent to the topic may have been included in this study. Practical implications: A few implications for artificial intelligence researchers, healthcare management, and one for regulatory agencies are presented considering the barriers and demands observed in this study. Management implications include testing artificial intelligence products and training staff to use them, creating vision and plans for valuable yet vulnerable data in healthcare facilities, collaborating on or with data, modeling, and policymaking, and experimentation avenues. Researchers are given some goals in light of this study to direct their research including bringing transparency to results, accounting for ethical, legal, bias, and coexistence issues, engaging healthcare professionals in their artificial intelligence research, and some ideas for applications that can be delivered today with the existing technology for which no barriers were observed. A regulatory agency implication is to find ways to test and rate artificially intelligent systems for healthcare use. Originality/value: Modern factors were considered that may impact decisions about adopting artificial intelligence in healthcare settings.