Article
Education & Educational Research
Xiaoming Zhai, Peng He, Joseph Krajcik
Summary: Involving students in scientific modeling practice is an effective approach to achieve science education learning goals. However, scoring student-developed models is time-consuming and challenging. This study utilized machine learning to automatically score student-drawn models and their descriptions, achieving good agreement with human scores.
JOURNAL OF RESEARCH IN SCIENCE TEACHING
(2022)
Article
Gastroenterology & Hepatology
Theresa Nguyen Wenker, Yamini Natarajan, Kadon Caskey, Francisco Novoa, Nabil Mansour, Huy Anh Pham, Jason K. Hou, Hashem B. El-Serag, Aaron P. Thrift
Summary: A natural language processing (NLP) algorithm was developed and validated to accurately identify dysplasia in Barrett's esophagus (BE) patients from pathology reports with varying formats. The algorithm demonstrated high sensitivity and accuracy, making it a valuable tool for research and clinical care.
CLINICAL GASTROENTEROLOGY AND HEPATOLOGY
(2023)
Article
Computer Science, Information Systems
Akshat Kumar, Heath Goodrum, Ashley Kim, Carly Stender, Kirk Roberts, Elmer Bernstam
Summary: This study successfully demonstrates the automatic identification of clinically significant abnormalities requiring follow-up in scanned documents using a ClinicalBERT-based model, achieving high recall and practically useful precision.
JOURNAL OF THE AMERICAN MEDICAL INFORMATICS ASSOCIATION
(2022)
Article
Computer Science, Software Engineering
Arianna Blasi, Alessandra Gorla, Michael D. Ernst, Mauro Pezze, Antonio Carzaniga
Summary: Software testing relies on effective oracles, with formal specification-based oracles revealing application-specific failures but being costly to obtain and maintain. MeMo is a technique and tool that automatically derives metamorphic equivalence relations from natural language documentation, effectively detecting defects when used as oracles in test cases.
JOURNAL OF SYSTEMS AND SOFTWARE
(2021)
Article
Education, Scientific Disciplines
Rebecca S. Gates, Kayla Marcotte, Rebecca Moreci, Brian C. George, Grace J. Kim, Kate H. Kraft, Tandis Soltani, Erkin Otles, Andrew E. Krumm
Summary: This study explores the quality of narrative feedback among trainee faculty gender dyads in an operative workplace-based assessment, revealing gender differences in the probability of receiving high-quality feedback. However, no significant differences were found based on faculty-resident gender dyad in providing high-quality narrative feedback.
JOURNAL OF SURGICAL EDUCATION
(2023)
Article
Computer Science, Artificial Intelligence
Abbas Akkasi, Jan Snajder
Summary: The proposed Word Sense Induction model utilizes automatically generated lexical substitutes to construct a graph and data preparation for Leader-Follower graph clustering. The method outperforms state-of-the-art solutions in terms of harmonic and geometric v-measure and f-score on the SemEval2010 dataset with a lower average number of sense groups.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Clinical Neurology
Bashar Zaidat, Justin Tang, Varun Arvind, Eric A. A. Geng, Brian Cho, Akiro H. Duey, Calista Dominy, Kiehyun D. Riew, Samuel K. K. Cho, Jun S. Kim
Summary: This study aims to automate the generation of CPT codes from operative notes using the XLNet machine learning algorithm. Results showed that the XLNet model performed close to human accuracy and successfully generated CPT billing codes from orthopedic surgeon's operative notes.
GLOBAL SPINE JOURNAL
(2023)
Article
Education & Educational Research
Rick Somers, Samuel Cunningham-Nelson, Wageeh Boles
Summary: This study utilized NLP techniques in an educational setting to automate assessment of students' conceptual understanding from short answer responses, providing insights and feedback. High performance NLP models were developed to assess the validity of students' justifications and their confidence levels in responses, with accuracies ranging from 91.46% to 99.46%.
AUSTRALASIAN JOURNAL OF EDUCATIONAL TECHNOLOGY
(2021)
Article
Clinical Neurology
Nicolas Vandenbussche, Cynthia Van Hee, Veronique Hoste, Koen Paemeleire
Summary: This study applies natural language processing (NLP) and machine learning (ML) algorithms to analyze and classify self-reported narratives of headache patients. The results show the potential of NLP in differentiating between patients with different headache disorders and the good performance of ML algorithms in classifying headache attack descriptions. This research highlights the importance of NLP in clinical information extraction.
JOURNAL OF HEADACHE AND PAIN
(2022)
Article
Computer Science, Artificial Intelligence
Hafiz Rizwan Iqbal, Rashad Maqsood, Agha Ali Raza, Saeed Ul Hassan
Summary: Automatic paraphrase detection is challenging for South Asian languages like Urdu due to the lack of standard evaluation resources. This study addresses this problem by proposing a semi-automatic pipeline for generating a paraphrased corpus for Urdu and presenting two novel DNN-based approaches for paraphrase detection. The evaluation results show that the proposed approaches outperform state-of-the-art methods for paraphrase detection and text reuse/plagiarism detection in Urdu.
NATURAL LANGUAGE ENGINEERING
(2023)
Article
Oncology
Sunil Bhatt, P. Connor Johnson, Netana H. Markovitz, Tamryn Gray, Ryan D. Nipp, Nneka Ufere, Julia Rice, Matthew J. Reynolds, Mitchell W. Lavoie, Madison A. Clay, Charlotta Lindvall, Areej El-Jawahri
Summary: This study used natural language processing to assess the extent of social support in hospitalized patients with advanced cancer and explored its association with survival and healthcare utilization. The results showed no significant associations between social support and survival or healthcare utilization, but cancer type may moderate the relationship between social support and survival.
Article
Psychology, Mathematical
Michael J. Tanana, Christina S. Soma, Patty B. Kuo, Nicolas M. Bertagnolli, Aaron Dembe, Brian T. Pace, Vivek Srikumar, David C. Atkins, Zac E. Imel
Summary: Emotional distress is a common reason for seeking psychotherapy, and sharing emotional material is central to the process of psychotherapy. Recent advances in technology have shown that using natural language processing can better identify emotions in therapist-client interactions.
BEHAVIOR RESEARCH METHODS
(2021)
Article
Radiology, Nuclear Medicine & Medical Imaging
Kaitlin M. Zaki-Metias, Jeffrey J. MacLean, Alexander M. Satei, Serguei Medvedev, Huijuan Wang, Christopher C. Zarour, Paul J. Arpasi
Summary: Incidental findings are often unrelated to the original purpose of imaging and require follow-up. The FIND Program was established to improve the follow-up process. This study analyzed the frequency of incidental findings and adherence to follow-up recommendations before and after the implementation of a tracking system.
JOURNAL OF DIGITAL IMAGING
(2023)
Article
Public, Environmental & Occupational Health
Alex Gillespie, Tom W. Reader
Summary: Safety reporting systems are widely used in healthcare, but their effectiveness can be undermined if staff fail to notice or report incidents. Online patient feedback, which allows for anonymous reporting, can be valuable in identifying safety incidents that have been overlooked by staff. This study developed an automated language analysis method to measure the likelihood of patient-reported safety incidents in online feedback. The identified incidents were often reported as unnoticed or unresolved, suggesting that patients use online platforms to raise awareness of safety concerns they believe have been ignored.
Article
Computer Science, Information Systems
Dongjin Yu, Lin Wang, Xin Chen, Jie Chen
Summary: In this paper, a novel approach based on BiLSTM networks with the attention mechanism is proposed to automatically detect self-admitted technical debt by leveraging source code comments. Experimental results show that the approach outperforms the state-of-the-art text mining-based method in terms of precision, recall, and F1-score.
FRONTIERS OF COMPUTER SCIENCE
(2021)
Article
Urology & Nephrology
Erkin Otles, Brian T. Denton, Bo Qu, Adharsh Murali, Selin Merdan, Gregory B. Auffenberg, Spencer C. Hiller, Brian R. Lane, Arvin K. George, Karandeep Singh
Summary: New models developed using the MUSIC registry outperformed existing models and should be considered as potential replacements for the prediction of pathological outcomes in prostate cancer.
JOURNAL OF UROLOGY
(2022)
Letter
Education, Scientific Disciplines
Sanaya Irani, Serena S. Bidwell, Quintin P. Solano
Article
Medicine, General & Internal
Fahad Kamran, Shengpu Tang, Erkin Otles, Dustin S. McEvoy, Sameh N. Saleh, Jen Gong, Benjamin Y. Li, Sayon Dutta, Xinran Liu, Richard J. Medford, Thomas S. Valley, Lauren R. West, Karandeep Singh, Seth Blumberg, John P. Donnelly, Erica S. Shenoy, John Z. Ayanian, Brahmajee K. Nallamothu, Michael W. Sjoding, Jenna Wiens
Summary: A machine learning model was created and validated to accurately predict clinical deterioration in covid-19 patients across different institutions. The model performed well in multiple medical centers and patient subgroups, showing potential for optimizing healthcare resources.
BMJ-BRITISH MEDICAL JOURNAL
(2022)
Article
Surgery
Quintin P. Solano, Jyothi R. Thumma, Cody Mullens, Ryan Howard, Anne Ehlers, Lia Delaney, Brian Fry, Mary Shen, Michael Englesbe, Justin Dimick, Dana Telem
Summary: This study evaluated the hospital-level variation of ventral or incisional hernia repair in the kidney transplant population and found significant differences in hernia repair rates among different hospitals. Patient and hospital characteristics also varied across tertiles, particularly in diabetes and obesity.
SURGICAL ENDOSCOPY AND OTHER INTERVENTIONAL TECHNIQUES
(2023)
Article
Surgery
Ryan Howard, Anne Ehlers, Lia Delaney, Quintin Solano, Mary Shen, Michael Englesbe, Justin Dimick, Dana Telem
Summary: Despite evidence supporting the use of mesh in ventral and incisional hernia repair, there is significant variation in practice patterns between hospitals that is not explained by patient characteristics or operative approach. This suggests opportunities to standardize surgical practice for better alignment with the evidence supporting mesh use.
SURGICAL ENDOSCOPY AND OTHER INTERVENTIONAL TECHNIQUES
(2023)
Article
Surgery
Ryan Howard, Anne Ehlers, Lia Delaney, Quintin Solano, Brian Fry, Michael Englesbe, Justin Dimick, Dana Telem
Summary: This study assessed decision regret among patients who underwent surgical management of ventral and inguinal hernias. The results showed that roughly 1 in 10 patients reported regret with their decision to undergo surgery, and regret was associated with complications and readmission.
SURGICAL ENDOSCOPY AND OTHER INTERVENTIONAL TECHNIQUES
(2022)
Article
Surgery
Anne P. Ehlers, Ryan Howard, Lia D. Delaney, Quintin Solano, Dana A. Telem
Summary: The use of mesh for small hernias is controversial. This study found that patients who had mesh placed during surgery may have a higher risk for complications, suggesting that the decision to use mesh may be driven by patient-related factors rather than evidence indicating its superiority in this population.
SURGICAL ENDOSCOPY AND OTHER INTERVENTIONAL TECHNIQUES
(2022)
Article
Surgery
Lia D. Delaney, Jyothi Thumma, Ryan Howard, Quintin Solano, Brian Fry, Justin B. Dimick, Dana A. Telem, Anne P. Ehlers
Summary: This study explores the motivating factors associated with surgeons' decisions to utilize a robotic approach for abdominal hernia repair. The qualitative analysis revealed three dominant themes: access and resources, surgeon comfort, and market factors. The study found significant variability in robotic utilization rates among different surgeons, with hernia location being the only factor associated with the use of robotic repair technique.
JOURNAL OF SURGICAL RESEARCH
(2022)
Article
Computer Science, Information Systems
Erkin Otle, Jon Seymour, Haozhu Wang, Brian T. Denton
Summary: This study investigates the benefits of dynamically estimating return to work (RTW) in occupational injuries using longitudinal observation data. The proposed longitudinal approach outperforms the baseline model in predicting future work status, showing potential for updating RTW predictions dynamically in injured workers.
JOURNAL OF THE AMERICAN MEDICAL INFORMATICS ASSOCIATION
(2022)
Correction
Public, Environmental & Occupational Health
Meghana Kamineni, Erkin Otles, Jeeheh Oh, Krishna Rao, Vincent B. Young, Benjamin Y. Li, Lauren R. West, David C. Hooper, Erica S. Shenoy, John G. Guttag, Jenna Wiens, Maggie Makar
INFECTION CONTROL AND HOSPITAL EPIDEMIOLOGY
(2023)
Article
Surgery
Quintin P. Solano, Ryan Howard, Anne Ehlers, Lia D. Delaney, Brian Fry, Michael Englesbe, Justin Dimick, Dana Telem
Summary: This study compares the application and short-term outcomes of anterior component separation (aCS) and posterior component separation (pCS) techniques in ventral hernia repair (VHR). The results show that, compared to patients without CS, patients undergoing aCS have a higher rate of 30-day adverse events. There were no significant differences in adverse events or surgical site infection (SSI) between pCS and aCS techniques.
JOURNAL OF SURGICAL RESEARCH
(2023)
Article
Surgery
Quintin P. Solano, Ryan Howard, Cody L. Mullens, Anne P. Ehlers, Lia D. Delaney, Brian Fry, Mary Shen, Michael Englesbe, Justin Dimick, Dana Telem
Summary: This study examined the association of frailty with outcomes after ventral hernia repair (VHR) and found that frailty was associated with postoperative complications, highlighting the importance of preoperative frailty assessment for risk stratification and patient counseling.
SURGICAL ENDOSCOPY AND OTHER INTERVENTIONAL TECHNIQUES
(2023)
Editorial Material
Cell Biology
Erkin Otles, Cornelius A. James, Kimberly D. Lomis, James O. Woolliscroft
Summary: Artificial intelligence is transforming medical practice, but medical students are not adequately prepared to utilize and evaluate AI systems. We propose integrating AI into medical curricula to equip graduating medical students with the skills to solve challenges at the intersection of AI and medicine.
CELL REPORTS MEDICINE
(2022)
Article
Public, Environmental & Occupational Health
Erkin Otles, Emily A. Balczewski, Micah Keidan, Jeeheh Oh, Alieysa Patel, Vincent B. Young, Krishna Rao, Jenna Wiens
Summary: This study compared the effectiveness of swab surveillance and daily risk estimates from a machine learning model in identifying patients likely to develop C. difficile infection in the ICU. The results showed that the ML model identified the same number of infections as swab surveillance with limited resources, and it also identified at-risk patients before disease onset, providing opportunities for prevention.
INFECTION CONTROL AND HOSPITAL EPIDEMIOLOGY
(2023)
Article
Education, Scientific Disciplines
Rebecca S. Gates, Kayla Marcotte, Rebecca Moreci, Brian C. George, Grace J. Kim, Kate H. Kraft, Tandis Soltani, Erkin Otles, Andrew E. Krumm
Summary: This study explores the quality of narrative feedback among trainee faculty gender dyads in an operative workplace-based assessment, revealing gender differences in the probability of receiving high-quality feedback. However, no significant differences were found based on faculty-resident gender dyad in providing high-quality narrative feedback.
JOURNAL OF SURGICAL EDUCATION
(2023)