A critical review of the applications of artificial intelligence in healthcare

Abstract

This paper primarily focused on the critical evaluation of six papers related to AI applications in healthcare. Four of these papers discussed AI applications for predictive analytics. The remaining two discussed AI applications in healthcare in general and the opportunities and challenges of using AI in healthcare. In the first paper, Ramirez (2024) describes the methodology adopted to reach conclusions. But no details are given. Some references were listed at the end, but not cited in the paper. The second paper of Shaheen (2021) qualitatively reviews the topic but uses very few references. The review is only a generalised discussion without any recommendations to address the challenges or future research. The third paper is a good research paper. Lin et al. (2017) developed a Bayesian multitask learning (BMTL) to enhance risk profiling of chronic patients. The model was validated through experiments. All the required details have been given with tables and figures as required. In the fourth paper, a qualitative review is done by Lee and Yoon (2021). They approach the topic from an industry perspective. Some real-world examples of AI applications provide strength to this paper. However, the paper is mostly textual with only one diagram of the historical progress of AI technology. Amarasingham et al. (2014) discuss the considerations and challenges of implementing AI for predictive analytics. In this fifth paper, descriptions are good, but there is only one table of challenges and actions. More tables and figures would have enhanced the quality of the paper. The sixth paper of Nithya and Ilango (2017) wasted a lot of paper space in presenting basic ideas about ML methods. There are several papers which describe these basic details. Discussions focusing on predictive analytics of cardiovascular disease, diabetes, hepatitis and cancer were good. Adequate citations are also provided. The detailed description of ML methods lowers the quality of the paper. Two reviews on the topic were used to examine what others say. Mostly, these papers concur with the points highlighted in the six papers reviewed.