Search

Drug labeling contains an ‘INDICATIONS AND USAGE’ that provides vital information to support clinical decision making and regulatory management. Effective extraction of drug indication information from free-text based resources could facilitate drug repositioning projects and help collect real-world evidence in support of secondary use of approved medicines. To enable AI-powered language models for the extraction of drug indication information, we used manual reading and curation to develop a Drug Indication Classification and Encyclopedia (DICE) based on FDA approved human prescription drug labeling. A DICE scheme with 7,231 sentences categorized into five classes (indications, contradictions, side effects, usage instructions, and clinical observations) was developed. To further elucidate the utility of the DICE, we developed nine different AI-based classifiers for the prediction of indications based on the developed DICE to comprehensively assess their performance. We found that the transformer-based language models yielded an average MCC of 0.887, outperforming the word embedding-based Bidirectional long short-term memory (BiLSTM) models (0.862) with a 2.82% improvement on the test set. The best classifiers were also used to extract drug indication information in DrugBank and achieved a high enrichment rate (>0.930) for this task. We found that domain-specific training could provide more explainable models without performance sacrifices and better generalization for external validation datasets. Altogether, the proposed DICE could be a standard resource for the development and evaluation of task-specific AI-powered, natural language processing (NLP) models.

Introduction The US Food and Drug Administration (FDA) priority review process applies to a drug that is considered a significant improvement over the available alternatives. The European Medicines Agency (EMA) accelerated approval applies to a product that is of major public health interest. This study assessed differences in the characteristics of priority review new molecular entities and new therapeutic biologic products approved by the FDA and the EMA. Methods This study includes regulatory information on drug applications, approvals, indications, and orphan designations of all priority review drugs approved by the FDA and the EMA in the period 1999-2011. Descriptive statistics, t-tests, and chi-squared and Wilcoxon tests were performed. Results Overall, 100 FDA priority review new molecular entities and new therapeutic biologics were approved by both agencies; 87.0%of the products were first approved by the FDA. The average FDA review time (9.2 ± 8.4months) was significantly lower than the EMA average review time (14.6 ± 4.0months) (p<0.0001). The FDA and the EMA granted orphan designation to 43.0% and 33.0%, respectively, of the applications. There were differences in the administration route (1.0%of all products), dosage (8.0%), strength (23%), posology (51.0%), indications (30.0%), restrictions of use (52.0%), limitations of use (19.0%), and outcomes limitations (28.0%) approved by both regulatory agencies. Conclusion Significant differences exist in the characteristics of the priority review drugs approved by the FDA and the EMA. Harmonization of the US and European regulatory frameworks may facilitate timely approval of pharmaceutical products. [ABSTRACT FROM AUTHOR]

Highlights: [•] Drugs and adverse reactions (ADRs) were clustered by a biclustering approach. [•] ADR mechanisms were inferred for the clusters with drug indications. [•] Some ADR cases were attributed to the patient's physiological backgrounds. [Copyright &y& Elsevier]