Year: 2025
Type: review
Abstract: the machine learning (ML) models.Secondly, advancement in computational power enabled complex calculations and model training with improved speed.This was also facilitated by the creation of publicly available critical care databases such as medical information mart for intensive care (MIMIC), with the latest edition being MIMIC-IV.3 This progression over the years is shown in Figure 1. Process o f DeveloPing a Machine learning MoDel f o r outcoMe PreDiction f r o M clinical DataMost basic algorithms of ML use data to analyze and learn.The ML algorithm then uses the results of the learning to make decisions and predict without external human input.ML models use large amounts of data and statistical methods to make real-time predictions using enormous computational power.The steps of development of a typical ML algorithm are shown in Figure 2. introDuction to artificial intelligenceArtificial intelligence (AI) is a broad field of computer science that focuses on developing intelligent machines that can mimic human functions, that is, visual perception, speech recognition, decision making, and language translation.AI is the ability of machines to exhibit intelligent behavior, human thinking, reasoning, and problem solving.AI has the potential to automate critically ill patient monitoring and predict prognosis, thus direct useful utilization of the resources.This has immense application in resource-limited settings such as India. 1 We have explored this broad topic with rapidly emerging applications by conducting this narrative review.This review aims to examine the potential uses of AI in critical care, as well as the benefits and challenges associated with its application and its penetration in India.Glossary of some of the currently used AI-related nomenclature in healthcare is provided in Table 1 for better understanding. history o f artificial intelligence i n critical care MeDicineThe progress from rule-driven management to data-driven healthcare commenced in the 1990s when ICU prognosis models using statistical methods like logistic regression were incorporated, such as acute physiology and chronic health evaluation (APACHE) and simplified acute physiology score (SAPS).They required frequent recalibration due to changing population dynamics and reduction in mortality.This progression was accelerated by modern AI incorporation into critical care starting from 2010. 3 AI in critical care was facilitated by the changing landscape in relevant contributory domains.Primarily, adoption of electronic health records (EHR) made the large amount of usable data available for research and training
Source: Journal of Acute Care
Authors: Ashwin Neelavar Udupa, Sandesh Kumar KJ
Institution: Manipal Hospital
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7118175487