The development of a scoring and ranking strategy for a patient-tailored adverse drug reaction prediction in polypharmacy

Abstract

only few applications are currently dealing with personalized adverse drug reactions (ADRs) prediction in case of polypharmacy. the study aimed to develop a patient-tailored ADR web application, considering characteristics from 734 drugs and relevant patient related factors. The application was designed in python using a scoring and ranking system based on frequency and severity, computed for each ADR and expressed through an online platform. A neural networks algorithm was used for predicting the severity of ADRs. the application inputs are: age, gender, drugs, relevant pathologies. The outputs are: an overall severity profile (hospitalization and mortality risk), a stratified risk on specific ADR groups and a sorted list of the most important ADRs depending on frequency and severity. The Severity prediction model validation resulted in 79.7-85.1% Area Under the Receiver Operating Characteristic Curve Score, which lies in the good cutoff of 75-90%. The program offers a complex view regarding the ADR profile of a given patient and could be used by the physician and clinical pharmacist during patient safety monitoring, for a coherent therapy choice or medication adjustment, due to the good therapy coverage and the inclusion of relevant patient comorbidities. Currently, adverse drug reactions (ADRs) represent major worldwide health issues, due to their negative impact on the quality of life and outcome of patients receiving multiple drug therapy. It is believed that almost 10% of the in-patients usually develop at least one ADR during hospitalization, while almost 50% of all ADRs appear prior to the admission 1-5. The major risk factors which specifically influence the development of ADRs are age, gender, ethnicity, pregnancy, multiple pathology, polypharmacy, drug dosage and frequency of administration, genetic polymorphism or special conditions (renal impairment, hepatic impairment) 6-9. The complex combination of these factors leads to the need of implementing precision medicine in case of a drug therapy for a certain patient 10. Therefore, the consideration of a patient-tailored ADR profile becomes of an utmost importance when choosing the best drug combination and during the patient safety monitoring. Several standardized databases containing valuable information regarding ADRs can be found, which could aid in extracting the information for such purposes. Among them, PROTECT ADR database lists all the ADRs and additional information derived from the Summaries of Product Characteristics of a high number of centrally authorized products by the European Medicines Agency 11. A prediction of the ADRs which are likely to appear for a certain patient considered pre-clinical data and spontaneous reports and was described by Ngufor and Wojtusiak 12 , which took into account the specific drug information and known ADRs from DrugBank and Side Effect Resource Database (SIDER), as well as drug-event signals and demographics from Food and Drug Administration Adverse Event Reporting System (FAERS) and MedEffect. In terms of ...

Patient ADE Risk Prediction through Hierarchical Time-Aware Neural Network Using Claim Codes

“…[25] proposed a hybrid model to predict the outcomes of ADEs, based on patients demographic data, such as age and gender, and drug-taken information, such as the route of the drug intake and whether the adverse reaction subsided when drug in-take was terminated. [26] developed a system that takes patient demographics, drugs, relevant diseases in pathology as input, and outputs ADE risk outcome assessment.…”