Review
Medical Informatics
Wael Abdelkader, Tamara Navarro, Rick Parrish, Chris Cotoi, Federico Germini, Alfonso Iorio, R. Brian Haynes, Cynthia Lokker
Summary: Machine learning approaches demonstrate good performance in retrieving high-quality clinical studies, achieving high sensitivity and precision levels. The study shows that supervised machine learning methods have been applied with high-quality standards during model training.
JMIR MEDICAL INFORMATICS
(2021)
Article
Computer Science, Artificial Intelligence
Dario Garigliotti, Krisztian Balog, Katja Hose, Johannes Bjerva
Summary: In this article, the authors propose a method of recommending specific tasks to users based on their search queries. By combining traditional term-based ranking techniques with continuous semantic representations, the proposed method outperforms text-based baselines in recommending tasks such as planning a holiday trip or organizing a party. The study also includes an analysis of features and queries.
NATURAL LANGUAGE ENGINEERING
(2023)
Review
Computer Science, Information Systems
Vannessa Duarte, Sergio Zuniga-Jara, Sergio Contreras
Summary: The study reveals a significant growth in the adoption of machine learning in marketing over the past few years, with increasing specialization in the type of problems solved. A variety of ML methods have been applied to address marketing problems related to consumer behavior, recommender systems, forecasting, marketing segmentation, and text analysis.
Review
Medicine, General & Internal
Luca Ronzio, Federico Cabitza, Alessandro Barbaro, Giuseppe Banfi
Summary: This article presents a systematic literature review on the application of machine learning in laboratory medicine, focusing on studies from 2017 to the present that have utilized machine learning techniques to analyze haematochemical parameters for diagnostic and prognostic purposes. The review aims to address the potential of these techniques in laboratory medicine and the under-utilization of this tool in the past.
Article
Computer Science, Software Engineering
Ziyi Zhou, Huiqun Yu, Guisheng Fan, Zijie Huang, Kang Yang
Summary: Code summarization aims to automatically generate code summaries, and it has recently gained significant research attention. Many recent approaches use neural machine translation techniques, training a Seq2Seq model on a large corpus to handle different code snippets. However, the diversity of codes in practice makes it challenging to learn all the patterns in a single model. This paper proposes a new framework called MLCS, which combines meta-learning and code retrieval to address this issue.
IEEE TRANSACTIONS ON SOFTWARE ENGINEERING
(2023)
Review
Computer Science, Information Systems
Konstantin Piliuk, Sven Tomforde
Summary: Emergency medicine is a popular area for applying artificial intelligence methods, but research in this field is still limited. This study systematically categorizes and investigates the obstacles faced by artificial intelligence applications in emergency medicine, and proposes directions for further research.
INTERNATIONAL JOURNAL OF MEDICAL INFORMATICS
(2023)
Article
Health Care Sciences & Services
Francisco Carrillo-Perez, Juan Carlos Morales, Daniel Castillo-Secilla, Olivier Gevaert, Ignacio Rojas, Luis Javier Herrera
Summary: This study explored fusion of five multi-scale and multi-omic modalities for lung cancer classification using machine learning techniques. The final classification model achieved high scores, indicating that leveraging the multi-scale and multi-omic nature of cancer data can enhance diagnostic performance.
JOURNAL OF PERSONALIZED MEDICINE
(2022)
Article
Computer Science, Interdisciplinary Applications
Nikolaos Stylianou, Ioannis Vlahavas
Summary: Argument Mining is the automatic identification of arguments in text, while Evidence-Based Medicine involves identifying related evidence in medical literature. Combining the two fields enhances performance to create high quality argument graphs annotated with PICO entities, allowing for the identification of arguments, PICO entities, and their relations in medical publications.
JOURNAL OF BIOMEDICAL INFORMATICS
(2021)
Article
Ecology
Nicolas Le Guillarme, Wilfried Thuiller
Summary: Easy access to multi-taxa information in scientific literature is essential for biodiversity understanding. TaxoNERD, a new tool using deep neural network (DNN) models, is proposed to recognize taxon mentions in ecological documents. Leveraging existing pretrained models on biomedical corpora, transfer learning can address the issue of extracting information from ecological texts.
METHODS IN ECOLOGY AND EVOLUTION
(2022)
Article
Computer Science, Software Engineering
Carlos Hernandez-Olivan, Jose R. Beltran
Summary: In this article, the musicaiz library, an object-oriented tool for analyzing, generating, and evaluating symbolic music, is presented. The library's submodules allow users to create symbolic music data, analyze it, encode MIDI data for deep learning models, modify existing music, and evaluate music generation systems. The library's evaluation submodule builds on previous work to objectively measure and reproduce the results of music generation models. The library is publicly available online and community contributions and feedback are encouraged.
Review
Oncology
Melissa Estevez, Corey M. Benedum, Chengsheng Jiang, Aaron B. Cohen, Sharang Phadke, Somnath Sarkar, Selen Bozkurt
Summary: This study presents an evaluation framework to assist model developers, data users, and other stakeholders in assessing the quality and applicability of data extracted using machine learning techniques from patient documents. This framework can facilitate effective utilization of this data in research.
Article
Agriculture, Multidisciplinary
Andrea Loddo, Mauro Loddo, Cecilia Di Ruberto
Summary: This paper focuses on using deep learning techniques to classify plant seeds datasets, achieving accuracy values of 95.65% and 97.47% for two datasets. It also successfully addresses the retrieval problem with deep learning approach. The results are considered as an excellent starting point for developing a complete seeds recognition, classification, and retrieval system to support agriculture and botany fields.
COMPUTERS AND ELECTRONICS IN AGRICULTURE
(2021)
Article
Computer Science, Artificial Intelligence
Bolin Zhang, Jiangbo Qian
Summary: This paper introduces a new unsupervised deep hashing method based on autoencoder, named AUCH, which can simultaneously learn feature representations, hashing functions, and cluster assignments, unifying unsupervised clustering and retrieval tasks into a single learning model, achieving competitive experimental results.
APPLIED INTELLIGENCE
(2021)
Article
Environmental Sciences
Qianqian Yang, Qiangqiang Yuan, Meng Gao, Tongwen Li
Summary: Remote sensing of air pollution is crucial for air quality management and health risk assessment. This study proposes a physics informed multi-task deep neural network (phyMTDNN) for the joint estimation of six main air pollutants. The proposed model utilizes the relationships among the pollutants to design the network structure and achieve simultaneous estimation of the pollutants.
SCIENCE OF THE TOTAL ENVIRONMENT
(2023)
Review
Health Care Sciences & Services
Yousef Khamis Ahmed Baqraf, Pantea Keikhosrokiani, Manal Al-Rawashdeh
Summary: This article provides a systematic review of the recent developments and current state of evaluating online health information quality, and identifies the direction of future research.
Article
Health Care Sciences & Services
Tamara Lotfi, Anisa Hajizadeh, Lorenzo Moja, Elie A. Akl, Thomas Piggott, Tamara Kredo, Miranda W. Langendam, Alfonso Iorio, Miloslav Klugar, Jitka Klugarova, Ignacio Neumann, Wojtek Wiercioch, Grigorios Leontiadis, Lawrence Mbuagbaw, Alexis F. Turgeon, Joerg Meerpohl, Adrienne Stevens, Jan Brozek, Nancy Santesso, Kevin Pottie, Omar Dewidar, Signe A. Flottorp, Justine Karpusheff, Zuleika Saz-Parkinson, Maria X. Rojas, Elena Parmelli, Derek K. Chu, Peter Tugwell, Vivian Welch, Marc T. Avey, Romina Brignardello-Petersen, Joseph L. Mathew, Zachary Munn, Robby Nieuwlaat, Nathan Ford, Amir Qaseem, Lisa M. Askie, Holger J. Schunemann
Summary: This study proposes a taxonomy and framework to identify and present actionable statements in guidelines. By reviewing case studies and testing the framework using COVID-19 guidelines, the study distinguishes five types of actionable statements. The results suggest that the framework can help guideline developers create actionable statements with clear intent, avoid informal recommendations, and improve understanding and implementation by users.
JOURNAL OF CLINICAL EPIDEMIOLOGY
(2022)
Article
Public, Environmental & Occupational Health
Thomas J. Reese, Chelsey R. Schlechter, Heidi Kramer, Polina Kukhareva, Charlene R. Weir, Guilherme Del Fiol, Tanner Caverly, Rachel Hess, Michael C. Flynn, Teresa Taft, Kensaku Kawamoto
Summary: Lung cancer screening using low-dose computed tomography (CT) can prevent thousands of deaths annually. To effectively implement this screening, systematic and theory-based strategies are needed. This study explored the implementation of lung cancer screening in primary care by integrating a decision aid into electronic health records. Implementation strategies were designed to target key behaviors and determinants, and these strategies can be replicated and tested in other institutions. Further research is needed to evaluate the effectiveness of these strategies and determine their applicability in different contexts.
TRANSLATIONAL BEHAVIORAL MEDICINE
(2022)
Article
Hematology
Greig Blamey, Becky Van Tassel, Elizabeth Sagermann, Ann Marie Stain, Linda Waterhouse, Alfonso Iorio
Summary: The HERO Study identified sexual health as an important issue for people with hemophilia worldwide, but this issue is often inadequately addressed at treatment centers due to barriers in communication. The pilot program in Canada successfully demonstrated improvements in HCPs' knowledge, skills, and comfort levels in discussing sexual health, highlighting the effectiveness of educational programs in facilitating these conversations.
Review
Health Care Sciences & Services
Federico Germini, Noella Noronha, Victoria Borg Debono, Binu Abraham Philip, Drashti Pete, Tamara Navarro, Arun Keepanasseril, Sameer Parpia, Kerstin de Wit, Alfonso Iorio
Summary: This study conducted a systematic review of the accuracy and acceptability of wrist-wearable activity trackers. The results showed that Fitbit Charge and Fitbit Charge HR had good accuracy in step counts, while Apple Watch was accurate in measuring heart rate. However, none of the tested devices proved to be accurate in measuring energy expenditure. Efforts should be made to reduce heterogeneity among studies.
JOURNAL OF MEDICAL INTERNET RESEARCH
(2022)
Review
Hematology
Amy D. Shapiro, Brandon M. Hardesty, Flora Peyvandi, Alfonso Iorio
Summary: Life expectancy for persons with hemophilia has increased due to advances in treatment, but they are now more likely to be affected by conditions associated with aging. This study analyzed the prevalence of bleeding and thrombotic events in hemophilia populations compared to the general population. The results showed consistently higher rates of bleeding events in hemophilia populations, while the prevalence of arterial thrombosis varied.
RESEARCH AND PRACTICE IN THROMBOSIS AND HAEMOSTASIS
(2023)
Article
Hematology
Jamie M. M. O'Sullivan, Ellia Tootoonchian, Baiba Ziemele, Michael Makris, Augusto B. B. Federici, Claudia Khayat Djambas, Magdy El Ekiaby, Dawn Rotellini, Robert F. F. Sidonio, Alfonso Iorio, Donna Coffin, Glenn F. F. Pierce, Jeffrey Stonebraker, Paula D. D. James, Michelle Lavin
Summary: Recent guidelines for von Willebrand Disease (VWD) highlighted the challenges in diagnosis and management. Identifying the number of persons with VWD (PwVWD) internationally will help target support to aid diagnosis of PwVWD. Registration rates of PwVWD vary internationally and are influenced by national income status.
Review
Hematology
Jeffrey S. Stonebraker, Alfonso Iorio, Michelle Lavin, Suely M. Rezende, Alok Srivastava, Glenn F. Pierce, Donna Coffin, Ellia Tootoonchian, Michael Makris
Summary: This study analyzed the reported prevalence of von Willebrand disease (VWD) worldwide in relation to income classification. The results showed that the prevalence of VWD was significantly higher in high-income countries compared to other income classifications, and the prevalence was higher in females than males. The detection and diagnosis of type 1 VWD presented challenges in forming a consistent prevalence value across countries and income classifications.
Review
Hematology
Pier Mannuccio Mannucci, Craig M. Kessler, Federico Germini, Francis Nissen, Richard Ofori-Asenso, Cristian Brocchieri, Sara Bendinelli, Alfonso Iorio
Summary: This systematic literature review assessed bleeding outcomes in people with congenital haemophilia A (PwcHA) using FVIII-containing products as prophylactic treatment. The results showed that PwcHA without inhibitors still experience bleeding despite FVIII prophylaxis. Therefore, improved standardization on capturing and reporting bleeding outcomes is needed for effective treatment comparisons.
Editorial Material
Medicine, General & Internal
Kensaku Kawamoto, Joseph Finkelstein, Guilherme Del Fiol
MAYO CLINIC PROCEEDINGS
(2023)
Article
Oncology
Lorena Gonzalez-Castro, Marcela Chavez, Patrick Duflot, Valerie Bleret, Alistair G. Martin, Marc Zobel, Jama Nateqi, Simon Lin, Jose J. Pazos-Arias, Guilherme Del Fiol, Martin Lopez-Nores
Summary: Breast cancer is a heterogeneous disease with varying risks of relapse, making it difficult to predict progression and choose appropriate follow-up strategies. Machine Learning algorithms can analyze large amounts of data and provide insights for predicting breast cancer recurrence. Structured data yields the best prediction results, while natural language processing offers comparable results with less mapping effort.
Article
Hematology
Alfonso Iorio, Vance Macdonald, Alexandre Caillaud, Maria D. Luckevich, Pia Christoffersen, Davide Matino, Arun Keepanasseril, Emma Iserman, Federico Germini, Anthony Bentley, Man-Chiu Poon
Summary: This study evaluates the impact of switching to N9-GP on treatment costs for patients with hemophilia. The results show that the switch to N9-GP leads to improvements in bleed rates and factor consumption, resulting in a decrease in treatment costs by 9.4% to 10.5%.
RESEARCH AND PRACTICE IN THROMBOSIS AND HAEMOSTASIS
(2023)
Article
Computer Science, Information Systems
Andrew D. Boyd, Rosa Gonzalez-Guarda, Katharine Lawrence, Crystal L. Patil, Miriam O. Ezenwa, Emily C. O'Brien, Hyung Paek, Jordan M. Braciszewski, Oluwaseun Adeyemi, Allison M. Cuthel, Juanita E. Darby, Christina K. Zigler, P. Michael Ho, Keturah R. Faurot, Karen L. Staman, Jonathan W. Leigh, Dana L. Dailey, Andrea Cheville, Guilherme Del Fiol, Mitchell R. Knisely, Corita R. Grudzen, Keith Marsolo, Rachel L. Richesson, Judith M. Schlaeger
Summary: Embedded pragmatic clinical trials (ePCTs) play a vital role in addressing population health problems by utilizing electronic health record (EHR) systems to increase the speed and volume of relevant research. However, the growing number of ePCTs using EHR-derived data increases the risk of biases due to differences in data capture and access to care, perpetuating health inequalities. This article identifies 3 challenges - incomplete data on social determinants of health, lack of representation of vulnerable populations, and data loss due to variable technology use - that exacerbate bias when working with EHR data and provides recommendations and examples to actively mitigate bias.
JOURNAL OF THE AMERICAN MEDICAL INFORMATICS ASSOCIATION
(2023)
Review
Computer Science, Information Systems
Yik-Ki Jacob Wan, Melanie C. Wright, Mary M. McFarland, Deniz Dishman, Mary A. Nies, Adriana Rush, Karl Madaras-Kelly, Amanda Jeppesen, Guilherme Del Fiol
Summary: This scoping review aimed to investigate the integration of surveillance algorithms that predict patient decompensation with clinical workflows and their impact on process and patient outcomes. The study found that most information displays continue to rely on well-accepted score-based algorithms.
JOURNAL OF THE AMERICAN MEDICAL INFORMATICS ASSOCIATION
(2023)
Article
Health Care Sciences & Services
Federico Germini, Victoria Borg Debono, David Page, Victoria Zuk, Alexandra Kucher, Chris Cotoi, Nicholas Hobson, Michael Sevestre, Mark W. Skinner, Alfonso Iorio
Summary: This study aimed to assess the needs of stakeholders involved in the use of the PROBE questionnaire, develop the necessary software infrastructure, and test the usability of the final product. Through interviews and user evaluations, an updated online survey and a mobile app were successfully developed with a high user satisfaction score. This will facilitate data collection for research and advocacy purposes in clinical practice.
JMIR HUMAN FACTORS
(2022)
Article
Hematology
Davide Matino, Alfonso Iorio, Arun Keepanasseril, Federico Germini, Alexandre Caillaud, Manuel Carcao, Julia Hews-Girard, Emma Iserman, Paula James, Adrienne Lee, Chai W. Phua, Haowei (Linda) Sun, Jerome Teitel, Man-Chiu Poon
Summary: This study assessed the treatment outcomes of using Nonacog beta pegol (N9-GP) in a real-world setting for patients with hemophilia B. The results showed that switching to N9-GP resulted in optimal bleeding control and lower factor consumption, regardless of the previous product used.
RESEARCH AND PRACTICE IN THROMBOSIS AND HAEMOSTASIS
(2022)