Artificial Intelligence And Machine Learning In Healthcare: Application And Challenges

Abstract

Introduction: The use of AI in healthcare offers many opportunities that support decision making. The objective of this paper is to give an insight of artificial intelligence’s main methods and applications in the field of healthcare. As well as to discuss its limitations and challenges.

Methods: A search was conducted on PubMed database.  Results were reviewed for articles published during the 5 last years.  Search terms included “artificial intelligence”, “machine learning”, and “healthcare".

Result:  Machine learning was the method of AI most used in healthcare. It was used especially for prediction (36.9 %), to make a diagnosis (18 %), or for monitoring (9 %). The machine learning’s algorithms most used were the random forest (29.6%), the regression logistic (27.8%), neural network (27.8%). The machine learning was used in 15.6% of case to improve the health system management. It was used in 12.5% of cases to describe and to try to understand the population’s psychosocial behaviours. In the clinical practice, it was mostly used in infectiology (15.6%).

Conclusion: AI promises a lot of opportunities in healthcare. However, challenges about ethical implications and the adoption of this technology in daily clinical practice must be arise.