Analysis of adverse drug events as a way to improve cancer patient care.

PURPOSE

To define the signals that a new artificial intelligence (AI) system must emit to improve adverse drug events (ADEs) management in oral antineoplastic agents (OAA).

METHODS

A multidisciplinary group of experts in patient safety was set up to define what signals the new AI system must emit to improve ADEs management in OAAs. The baseline data for the new AI system were generated through an observational and ambispective study carried out in a university hospital. All patients who met the inclusion criteria were selected consecutively every working day for 6 months. The ADEs were collected by interview and by the review of health records. The ADEs were categorised according to how they could be detected: patient, analysis, examination.

RESULTS

The group defined what signals the AI system must emit to improve ADEs management in OAAs: a signal to educate the patient when the possible ADEs were categorised as patient, a signal as a reminder to request a blood test or a microbiological culture when the possible ADEs were categorised as analysis, and a signal as a reminder for the necessity of a clinical examination when the possible ADEs were categorised as examination. A total of 1652 ADEs were reported in the interviews (ADE-interview) with the pharmacist, and doctors noted 1989 ADEs in the health record (ADE-HR). The most frequent ADEs were identified in the patient category.

CONCLUSION

This study opens a new way for better management of ADEs and is the first step in the development of a future technology, which will improve the quality of life of patients.