On the use of logistics data to anticipate drugs shortages through data mining

Abstract

Drugs shortages have become a major issue affecting healthcare systems worldwide. With the growing use of electronic records, data analytics techniques have been identified as a major asset to provide support in managing pharmaceutical supply chains. However, applications addressing drugs shortages, which arise from various complex factors, are still expected. Therefore, this paper investigates whether the use of logistics data at the national level can provide insights into drugs supply disruptions issues. Basic features were defined from a dataset of 1.281.545 electronic data interchange records between French hospitals and drugs distributors in 2021. Additionally, anomaly detection and correlation analysis were performed to (i) detect supply disruptions of four products in 2021; and (ii) assess whether the features defined can be used in future time series predictive models. Findings highlight a promising opportunity to harness these data through more advanced analytics techniques to anticipate and manage drugs shortages at local, national, and European levels.