Antoine SAAB, Ph.D in Medical Informatics, Quality and Development Manager, LHG-UMC
Use of Artificial Intelligence to improve the performance of clinical Adverse Events detection tools in healthcare institutions
Patient safety is considered a top public health priority. According to the World Health Organization, studies have shown that on average one in ten patients is subject to an adverse event (AE) during their hospitalization in high-income countries. The estimate for low- and middle-income countries even suggests that up to one in four patients are affected. Despite their high preventability, AEs remain one of the leading causes of morbidity and mortality, leading to excess hospitalization costs, longer hospital stays and higher rates of readmissions.
Despite a positive momentum in the last twenty years, substantial opportunities for improvement remain. One of the main reasons why the high incidence of AEs persists is related to the lack of reliable and effective tools for the routine measurement and monitoring of clinical AEs, the availability of which is in itself important to identify the evolution of the incidences of different categories of AEs and the impact of the improvement actions deployed.
Currently, only a small percentage of healthcare institutions possess tools capable of such measurements of AEs, and for those who have deployed such systems, the overall detection performance remains low.
The work that will be presented concerns the use of artificial intelligence (AI) techniques (in particular rules-based and machine learning) to improve the performance of current tools for measuring AEs and the cost of their implementation.
This approach was also materialized through the development of two monitoring tools already implemented routinely in a university hospital center.
