Historical profile will tell? A deep learning-based multi-level embedding framework for adverse drug event detection and extraction

Xia, Long

doi:10.1016/j.dss.2022.113832

Cited by 12 publications

(8 citation statements)

References 66 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The main advantages noted were that social media data included unexpected or new adverse events [35-43, 49, 51, 53, 54, 57, 60, 64, 67, 71-73, 80, 84, 86-90, 92, 98, 101, 104, 105] (24 studies) and that adverse events could be identified earlier [35, 60, 71, 72, 79, 86-88, 92, 93, 98, 101] (9 studies) in social media as compared to those reported in spontaneous reporting systems [35,71,72,76,79,93] , search query logs from search engines [35], drug safety communications [101] and scientific literature [76,[86][87][88]. In contrast, 3 out of the 60 studies suggested that routine surveillance of social media would not aid in earlier identification of ADE signals [24,50,95] , while one stated it will not be useful to confirm previously identified safety signals [45] and another one that certain social media platforms (such as online health forums) may be timelier in signal detection while others (Twitter) will not [35].…”

Section: Results Of Comparisonmentioning

confidence: 99%

“…In contrast, studies utilizing advanced NLP and ML techniques, such as deep learning, have demonstrated superior performance in ADE recognition, including rare and previously unknown ADEs. For instance, Xia 2022 [101] developed a historical awareness multilevel framework that leverages transfer learning from prior review embeddings and utilizes BERTbased sentence and word embeddings with an attention mechanism. This approach achieved state-ofthe-art performance with an impressive F-1 score of 0.944.…”

Section: Discussionmentioning

confidence: 99%

“…In several studies it was observed that the frequency of drug mentions in social media varied depending on the specific drug [24,50,101,105]. It was reported that drugs ranked in the top 100 by sales generated more posts compared to other drugs.…”

Section: Discussionmentioning

confidence: 99%

See 2 more Smart Citations

The value of social media analysis for adverse events detection and pharmacovigilance: a Scoping Review (Preprint)

Golder,

O'Connor,

Wang

et al. 2024

JMIR Public Health and Surveillance

View full text Add to dashboard Cite

Background: Adverse drug events pose a significant public health burden leading to hospitalisation, disability, and death. Even those adverse events categorised as non-serious can severely impact on patient quality of life, adherence, and persistence. Monitoring medication safety is challenging. Online patient reports on social media may be a useful supplementary source of realworld data. Despite the growth of sophisticated techniques for identifying adverse events using social media data, a consensus has not been reached as to the value of social media in relation to more traditional data sources.Objective: To evaluate and characterise the utility of social media analysis in adverse drug event detection and pharmacovigilance as compared to other data sources (such as spontaneous reporting systems and the clinical literature). Methods:In this scoping review, we searched 11 bibliographical databases and Google Scholar, followed by handsearching and forward and backward citation searching. Each records was screened by two independent reviewers at both title/abstract stage and the full-text screening stage. Studies were included if they 1) used any type of social media (such as Twitter or patient forums) to detect any adverse event associated with any drug medication and 2) compared the results ascertained from social media to any other data source. Study information was collated using a piloted data extraction sheet . Data were extracted on 1) the adverse events and drugs searched for and included, 2) the methods used (such as machine learning), 3) social media data source, 4) volume of data analysed, 5) limitations of the methodology, 6) availability of data and code, 7) comparison data source and comparison methods, 8) results, including the volume of adverse events, and how adverse events found compared to other data sources in their seriousness, frequencies, and expectedness/novelty (new vs. known knowledge) and 9) conclusions.Results: Of the 6538 unique records screened, 73 publications representing 60 studies and a wide variety of methods met our inclusion criteria. The most common social media platforms used were Twitter and online health forums. The most common comparator data source was spontaneous reporting systems, although other comparisons were made such as with scientific literature and product labels. Although similar patterns of adverse event reporting tended to be identified, the frequencies were lower in social media. Social media data were found useful in identifying new or unexpected adverse events and adverse events in a more timely manner.Conclusions: There is a large body of research comparing adverse events from social media to other sources. Most studies advocated the use of social media as an adjunct to traditional data sources. Some indicated the value of social media in understanding patient perspectives such as the impact of adverse events, which could be better explored.

show abstract

Section: Results Of Comparisonmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

The value of social media analysis for adverse events detection and pharmacovigilance: a Scoping Review (Preprint)

Golder,

O'Connor,

Wang

et al. 2024

JMIR Public Health and Surveillance

View full text Add to dashboard Cite

show abstract

“…The need for Business Intelligence systems in hospital management has been highlighted by different authors, but there are few examples in the context of drug management ( González-Pérez et al, 2022 ) and none of them considers both structured and unstructured data from different sources as the proposed architecture does. Existing studies focus on processing structured data obtained from the hospital ERP ( González-Pérez et al, 2022 ; Sajogo, Teoh & Lebedevs, 2023 ), sometimes using AI ( Galli et al, 2021 ; Xia, 2022 ; Raza et al, 2022 ), or unstructured data from social networks ( Li et al, 2019 ; Hajjami, Berrada & Harti, 2020 ).…”

Section: Discussionmentioning

confidence: 99%

A computer architecture based on disruptive information technologies for drug management in hospitals

Chalmeta

Ruíz

Soriano-Irigaray

2023

PeerJ Computer Science

View full text Add to dashboard Cite

The drug management currently carried out in hospitals is inadequate due to several factors, such as processes carried out manually, the lack of visibility of the hospital supply chain, the lack of standardized identification of medicines, inefficient stock management, an inability to follow the traceability of medicines, and poor data exploitation. Disruptive information technologies could be used to develop and implement a drug management system in hospitals that is innovative in all its phases and allows these problems to be overcome. However, there are no examples in the literature that show how these technologies can be used and combined for efficient drug management in hospitals. To help solve this research gap in the literature, this article proposes a computer architecture for the whole drug management process in hospitals that uses and combines different disruptive computer technologies such as blockchain, radio frequency identification (RFID), quick response code (QR), Internet of Things (IoT), artificial intelligence and big data, for data capture, data storage and data exploitation throughout the whole drug management process, from the moment the drug enters the hospital until it is dispensed and eliminated.

show abstract

“…Therefore, the study of new media events mainly includes the study of network public opinion and crisis communication and the study of new media event early warning [8]. The new media event early warning construction method is mainly studied from the aspects of semantic analysis and opinion mining of network information text [9]. Literature [10] achieved better research results in the construction of sentiment corpus and singlegrained viewpoint mining algorithm, which is of guiding significance to the construction of new media early warning model; Literature [11] analyzes the eight contents affecting network public opinion security for the network public opinion security monitoring and early warning of colleges and universities; Literature [12] analyzes eight contents affecting network public opinion security for the imperfection of public opinion early warning method in the establishment of network public opinion crisis early warning model In the establishment of network public opinion crisis early warning model, three sub-models of flexible public opinion mining, viewpoint evolution and network public opinion crisis early warning are established; Literature [13] utilizes the hierarchical analysis method, and constructs the early warning index system of network public opinion emergencies with three types of factors, namely, warning sources, warning signs and warning situations, by distributing questionnaires to the experts; Literature [14] starts from analyzing the current situation of network public opinion crisis early warning research findings, grasps the network public opinion early warning research and practice as a whole, and gives the corresponding strategies.…”

Section: Introductionmentioning

confidence: 99%

Design of New Media Event Warning Method Based on K-means and Seasonal Optimization Algorithm

Gao,

Zheng

2024

ICST Transactions on Scalable Information Systems

View full text Add to dashboard Cite

INTRODUCTION: Timely and effective early warning of new media events not only provides academic value to the study of new media events, but also can play a positive role in promoting the resolution of public opinion. OBJECTIVES: Aiming at the current research on early warning of new media events, there are problems such as the theoretical research is not in-depth and the early warning model is not comprehensive. METHOD: In this paper, K-means and seasonal optimization algorithm are used to construct new media event early warning method. Firstly, by analyzing the construction process of new media event early warning system, extracting text feature vector and carrying out text feature dimensionality reduction; then, combining with the random forest algorithm, the new media event early warning method based on intelligent optimization algorithm optimizing K-means clustering algorithm is proposed; finally, the validity and superiority of the proposed method is verified through the analysis of simulation experiments. RESULTS: The method developed in this paper improves the accuracy, time performance of new media event warning techniques. CONCLUSION: Addresses the lack of comprehensiveness of current approaches to early warning of new media events.

show abstract

Historical profile will tell? A deep learning-based multi-level embedding framework for adverse drug event detection and extraction

Cited by 12 publications

References 66 publications

The value of social media analysis for adverse events detection and pharmacovigilance: a Scoping Review (Preprint)

The value of social media analysis for adverse events detection and pharmacovigilance: a Scoping Review (Preprint)

A computer architecture based on disruptive information technologies for drug management in hospitals

Design of New Media Event Warning Method Based on K-means and Seasonal Optimization Algorithm

Contact Info

Product

Resources

About