Automated medical skill assessment facilitates medical education by merging varying clinical experiences across instructors for standardizing medical training. However, medical datasets for training such automated assessment rarely have satisfactory sizes due to the cost of data collection, safety concerns and privacy restrictions. Current medical training relies on evaluation rubrics that usually include multiple auxiliary labels to support the overall evaluation from varying aspects of the procedure. In this paper, we explore machine learning algorithms to design a generalizable auxiliary task-based framework for medical skill assessment to address training automated systems with limited data. Our framework exhaustively mines valid auxiliary information in the evaluation rubric to pre-train the feature extractor before training the skill assessment classifier. Notably, a new regression-based multitask weighting method is the key to pre-train a meaningful feature representation comprehensively, ensuring the evaluation rubric is well imitated in the final model. The overall evaluation task can be fine-tuned based on the pre-trained rubric-based feature representation. Our experimental results on two medical skill datasets show that our work can significantly improve performance, achieving 85.9% and 97.4% accuracy in the intubation dataset and surgical skill dataset, respectively.

Download full-text PDF

Source
http://dx.doi.org/10.1109/EMBC46164.2021.9630498DOI Listing

Publication Analysis

Top Keywords

medical skill
16
skill assessment
16
medical
8
automated medical
8
limited data
8
medical training
8
training automated
8
evaluation rubric
8
feature representation
8
skill
6

Similar Publications

Background: Involving parents in decisions about the care of their infant is common practice in most neonatal intensive care units. However, involvement is less common in neonatal research and a gap appears to exist in understanding the process of patient and public involvement. The aim of this study was to explore parents and researchers' experiences of patient and public involvement in a neonatal research project.

View Article and Find Full Text PDF

Background: Compassion Competence and the ability to strive to understand the suffering of patients in psychiatric ward is essential for nurses to establish effective therapeutic communication in the process of their recovery. Patient Safety Competency is of great importance for nurses to prevent adverse events and minimize errors. This study aimed to investigate the relationship between Compassion Competence and Patient Safety Competency in nurses working in psychiatric wards of Shiraz University of Medical Sciences affiliated hospitals in 2024.

View Article and Find Full Text PDF

Introduction: Ultrasound-guided regional anaesthesia enhances pain control, patient outcomes and lowers healthcare costs. However, teaching this skill effectively presents challenges with current training methods. Simulation-based medical education offers advantages over traditional methods.

View Article and Find Full Text PDF

Digital-based emergency prevention and control system: enhancing infection control in psychiatric hospitals.

BMC Med Inform Decis Mak

January 2025

Department of Nutritional and Metabolic Psychiatry, The Affiliated Brain Hospital, Guangzhou Medical University, No. 36 Fangcun Mingxin Road, Liwan District, Guangzhou, 510370, China.

Background: The practical application of infectious disease emergency plans in mental health institutions during the ongoing pandemic has revealed significant shortcomings. These manifest as chaotic management of mental health care, a lack of hospital infection prevention and control (IPC) knowledge among medical staff, and unskilled practical operation. These factors result in suboptimal decision-making and emergency response execution.

View Article and Find Full Text PDF

Evaluating medical learners' experiences with health literacy at a southeastern medical school.

BMC Med Educ

January 2025

Department of Health Services, Policy, and Management, University of South Carolina Arnold School of Public Health, 915 Greene Street, Discovery Bldg. Suite 349, Columbia, 378, South Carolina, USA.

Background: Health literacy (HL) is crucial for making informed health decisions. Over one-third of US adults have limited HL, leading to adverse health outcomes. Despite its importance, HL education lacks standardization in medical training.

View Article and Find Full Text PDF

Want AI Summaries of new PubMed Abstracts delivered to your In-box?

Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!