Background: Drug prescriptions are often recorded in free-text clinical narratives; making this information available in a structured form is important to support many health-related tasks. Although several natural language processing (NLP) methods have been proposed to extract such information, many challenges remain. Objective: This study evaluates the feasibility of using NLP and deep learning approaches for extracting and linking drug names and associated attributes identified in clinical free-text notes and presents an extensive error analysis of different methods. This study initiated with the participation in the 2018 National NLP Clinical Challenges (n2c2) shared task on adverse drug events and medication extraction. Methods: The proposed system (DrugEx) consists of a named entity recognizer (NER) to identify drugs and associated attributes and a relation extraction (RE) method to identify the relations between them. For NER, we explored deep learning-based approaches (ie, bidirectional long-short term memory with conditional random fields [BiLSTM-CRFs]) with various embeddings (ie, word embedding, character embedding [CE], and semantic-feature embedding) to investigate how different embeddings influence the performance. A rule-based method was implemented for RE and compared with a context-aware long-short term memory (LSTM) model. The methods were trained and evaluated using the 2018 n2c2 shared task data. Results: The experiments showed that the best model (BiLSTM-CRFs with pretrained word embeddings [PWE] and CE) achieved lenient micro F-scores of 0.921 for NER, 0.927 for RE, and 0.855 for the end-to-end system. NER, which relies on the pretrained word and semantic embeddings, performed better on most individual entity types, but NER with PWE and CE had the highest classification efficiency among the proposed approaches. Extracting relations using the rule-based method achieved higher accuracy than the context-aware LSTM for most relations. Interestingly, the LSTM model performed notably better in the reason-drug relations, the most challenging relation type. Conclusions: The proposed end-to-end system achieved encouraging results and demonstrated the feasibility of using deep learning methods to extract medication information from free-text data.

Managing large-scale gatherings, such as global festivals, sporting events, and religious congregations, presents substantial challenges in ensuring crowd safety and control. Innovative frameworks are essential to address these complexities effectively. The Integrated Intelligent Crowd Control and Management (IICCM) framework combines cutting-edge technologies, including Computer Vision (CV), Artificial Intelligence (AI), and the Internet of Things (IoT), to enhance participant safety and optimize crowd management. CV enables precise real time identification and tracking, AI analyzes crowd behavior to anticipate risks, and IoT gathers environmental data to improve crowd flow, alleviate congestion, and provide timely assistance. Additionally, the framework facilitates emergency evacuation planning by modeling crowd dynamics and identifying safe, efficient escape routes. Although suitable for diverse events, the Hajj pilgrimage—a uniquely large and dynamic annual gathering—provides a rigorous test case for the IICCM framework. Managing millions of participants from varied cultural and linguistic backgrounds highlights the system’s adaptability and robustness. By effectively addressing Hajj specific challenges, the IICCM framework demonstrates its scalability and applicability to other large-scale events. This research offers valuable insights for decision-makers seeking to implement advanced crowd management technologies.

ObjectivesElectronic Health Records (EHRs) contain a wealth of routinely-collected data that could potentially be used to inform clinical decisions such as the choice between competing treatment regimens. Apart from structured data about diagnoses and biomarkers, these records often include unstructured data such as free-text medication prescriptions. Disjoint toolsets exist for structured and unstructured data, making it difficult to analyse datasets that comprise both structured and unstructured data. Representing free-text items in a structured and standardised format would enable their statistical analysis. The aim of this study was to develop a generic, analytical pipeline that integrates different tools for text analytics and statistical modelling, and to apply it to data from the UK Renal Registry (UKRR) to answer specific clinical and epidemiological questions on kidney transplant immune suppression. ApproachThe UKRR database comprises data from all renal units in England, Wales and Northern Ireland and consists of both structured and semi-structured data. Our workflow starts by using rules to extract medication regimen from free-text prescriptions, which is then automatically combined with structured patient's record. The data is then analysed for patterns of transplant immune suppression prescribing by specific centres across the UK. ResultsWe have developed an analytical pipeline for improving concordance between unstructured and structured medical records, combining new and established text analysis and subsequence analysis tools. Results of the study are underway and will be presented at the conference. ConclusionsWe have developed new framework that integrates tools for text analytics and statistical modelling, to facilitate the analysis of mixed structured and unstructured data. Our analysis of the UKRR data will help to compare immune suppressive treatment regimens, identify best practice, and explore the associations between transplant medication and transplant outcomes.