Artificial Intelligence as a Clinical Decision Support System in Hospitals

Abstract

Introduction:

The rapid advancement of AI technologies, particularly in machine learning and deep learning, comes at a critical juncture where studies estimate that preventable medical errors rank as the third leading cause of death in the world. This systematic review aimed to provide a comprehensive understanding of the current status and potential of AI as a Clinical Decision Support System (CDSS) in hospital settings.

Methods:

To conduct the systematic review, a comprehensive search strategy was implemented, utilizing electronic databases such as PubMed, Embase, and Scopus. The search, focused on keywords like "Artificial Intelligence," "Clinical Decision Support System," and "Hospital," employed Boolean operators for precision. The inclusion criteria for studies were specified to focus on AI in Clinical Decision Support Systems within hospital settings, with outcomes related to clinical decision-making or patient care, and published in English. After an initial search yielding 37 clinical trials, 22 unique records were retained following duplicate removal. Two reviewers independently screened titles and abstracts, with full-text evaluations and final study selection achieved through consensus. A standardized data extraction form was used to gather study characteristics, participant demographics, AI technologies, outcomes, and key findings.

Results:

Eight clinical trials, ranging in sample sizes from 170 to over 2,500 participants, were included in the systematic review, offering nuanced insights into the application of Artificial Intelligence (AI) as a Clinical Decision Support System (CDSS) across diverse healthcare settings [9-16]. The trials involved various health professions, with 53% focusing on physicians, 31% on nurses, and 16% encompassing a combination of healthcare professionals, showcasing the broadapplicability of AI in multidisciplinary teams. The multifaceted interventions, predominantly centered on diagnostic support, underscore the versatility of AI applications in clinical decision-making, addressing diverse healthcare challenges. The trials consistently reported a substantial reduction in diagnostic errors (33%) and medication errors (25%), along with a statistically significant 20% decrease in adverse events associated with medical errors, suggesting that AI as a CDSS holds significant potential for enhancing the accuracy and safety of clinical decision-making processes.

Conclusions:

The results of this systematic review consistently demonstrate a substantial reduction in medical errors with the implementation of Artificial Intelligence (AI) as a Clinical Decision Support System (CDSS). Across diverse healthcare settings, the trials reported an average reduction of a third in diagnostic errors, a quarter in medication errors, and a statistically significant decrease in adverse events associated with medical errors.