The main two goals of my PharmacoInformatics research line are: 1) to improve pharmacotherapy outcomes in patients with multimorbidity, and 2) to improve the detection and prevention of adverse drug events.

Since patients with multimorbidity are usually excluded from randomized controlled trials, I reuse routinely-collected data from Electronic Health Records (EHRs) to address many existing knowledge gaps regarding the safety and effectiveness of drugs and their combinations. At present, my focus is on Intensive Care patients and patients with chronic kidney disease.

Currently, the gold standard for detecting Adverse Drug Events (ADEs) involves a manual patient chart review conducted by multiple medical experts. However, this procedure is resource-intensive and has a limited reproducibility, making it impractical for continuous patient safety monitoring. In this regard, I also leverage routinely-collected data from EHRs to develop and validate automated tools for ADE detection.

To analyze routinely-collected data, I employ various techniques ranging from conventional statistics to machine learning and natural language processing. Specifically, I utilize methods for causal inference that enable personalized estimations of treatment effects and side-effects.

To implement real-world pharmacotherapy insights at the point-of-care, I study different implementation strategies, including computerized decision support systems, prediction and causal models, dashboards, and performance feedback reports.

PharmacoInformatics, Clinical Pharmacology, Multimorbidity, Adverse Drug Events, Reuse of routinely-collected patient data