One-shot skill assessment in high-stakes domains with limited data via meta learning.

Deep Learning (DL) has achieved robust competency assessment in various high-stakes fields. However, the applicability of DL models is often hampered by their substantial data requirements and confinement to specific training domains. This prevents them from transitioning to new tasks where data is scarce. Therefore, domain adaptation emerges as a critical element for the practical implementation of DL in real-world scenarios. Herein, we introduce A-VBANet, a novel meta-learning model capable of delivering domain-agnostic skill assessment via one-shot learning. Our methodology has been tested by assessing surgical skills on five laparoscopic and robotic simulators and real-life laparoscopic cholecystectomy. Our model successfully adapted with accuracies up to 99.5 % in one-shot and 99.9 % in few-shot settings for simulated tasks and 89.7 % for laparoscopic cholecystectomy. This study marks the first instance of a domain-agnostic methodology for skill assessment in critical fields setting a precedent for the broad application of DL across diverse real-life domains with limited data.

Assessment of Surgical Tasks Using Neuroimaging Dataset (ASTaUND).

Functional near-infrared spectroscopy (fNIRS) is a neuroimaging tool for studying brain activity in mobile subjects. Open-access fNIRS datasets are limited to simple and/or motion-restricted tasks. Here, we report a fNIRS dataset acquired on mobile subjects performing Fundamentals of Laparoscopic Surgery (FLS) tasks in a laboratory environment. Demonstrating competency in the FLS tasks is a prerequisite for board certification in general surgery in the United States. The ASTaUND data set was acquired over four different studies. We provide the relevant information about the hardware, FLS task execution protocols, and subject demographics to facilitate the use of this open-access data set. We also provide the concurrent FLS scores, a quantitative metric for surgical skill assessment developed by the FLS committee. This data set is expected to support the growing field of assessing surgical skills via neuroimaging data and provide an example of data processing pipeline for use in realistic, non-restrictive environments.

Automated Video Debriefing Using Computer Vision Techniques.

INTRODUCTION: Within any training event, debriefing is a vital component that highlights areas of proficiency and deficiency, enables reflection, and ultimately provides opportunity for remediation. Video-based debriefing is desirable to capture performance and replay events, but the reality is rife with challenges, principally lengthy video and occlusions that block line of sight from camera equipment to participants. METHODS: To address this issue, researchers automated the editing of a video debrief, using a system of person-worn cameras and computer vision techniques. The cameras record a simulation event, and the video is processed using computer vision. Researchers investigated a variety of computer vision techniques, ultimately focusing on the scale invariant feature transform detection method and a convolutional neural network. The system was trained to detect and tag medically relevant segments of video and assess a single exemplar medical intervention, in this case the application of a tourniquet. RESULTS: The system tagged medically relevant video segments with 92% recall and 66% precision, resulting in an F1 (harmonic mean of precision and recall) of 72% (N = 23). The exemplar medical intervention was successfully assessed in 39.5% of videos (N = 39). CONCLUSION: The system showed suitable accuracy tagging medically relevant video segments, but requires additional research to improve medical intervention assessment accuracy. Computer vision has the potential to automate video debrief creation to augment existing debriefing strategies.

Simulation-based education improves military trainees' skill performance and self-confidence in tourniquet placement: A randomized controlled trial.

BACKGROUND: Tactical Combat Casualty Care (TCCC) is the standard of care for stabilization and treatment of military trauma patients. The Department of Defense has mandated that all service members receive role-based TCCC training and certification. Simulation education can increase procedural skills by providing opportunities for deliberate practice in safe, controlled environments. We developed and evaluated the effectiveness of a simulation-based TCCC training intervention to improve participants' skill performance and self-confidence in tourniquet placement. METHODS: This study was a single-blinded, randomized trial with waitlist controls. Army Reserve Officers Training Corp cadets from a single training battalion comprised the study population. After randomization and baseline assessment of all participants, group A alone received focused, simulation-based TCCC tourniquet application training. Three months later, all participants underwent repeat testing, and after crossover, the waitlist group B received the same intervention. Two months later, all cadets underwent a third/final assessment. The primary outcome was tourniquet placement proficiency assessed by total score achieved on a standardized eight-item skill checklist. A secondary outcome was self-confidence in tourniquet application skill as judged by participants' Likert scale ratings. RESULTS: Forty-three Army Reserve Officers Training Corp cadets completed the study protocol. Participants in both group A (n = 25) and group B (n = 18) demonstrated significantly higher performance from baseline to final assessment at 5 months and 2 months, respectively, following the intervention. Mean total checklist score of the entire study cohort increased significantly from 5.53 (SD = 2.00) at baseline to 7.56 (SD = 1.08) at time 3, a gain of 36.7% ( p < 0.001). Both groups rated their self-confidence in tourniquet placement significantly higher following the training. CONCLUSION: A simulation-based TCCC curriculum resulted in significant, consistent, and sustained improvement in participants' skill proficiency and self-confidence in tourniquet placement. Participants maintained these gains 2 months to 5 months after initial training. LEVEL OF EVIDENCE: Therapeutic/care management; Level II.

BurnCare tablet trainer to enhance burn injury care and treatment.

BACKGROUND: Applied Research Associates (ARA) and the United States Army Institute of Surgical Research (USAISR) have been developing a tablet-based simulation environment for burn wound assessment and burn shock resuscitation. This application aims to supplement the current gold standard in burn care education, the Advanced Burn Life Support (ABLS) curriculum. RESULTS: Subject matter experts validate total body surface area (TBSA) identification and analysis and show that the visual fidelity of the tablet virtual patients is consistent with real life thermal injuries. We show this by noting that the error between their burn mapping and the actual patient burns was sufficiently less than that of a random sample population. Statistical analysis is used to confirm this hypothesis. In addition a full body physiology model developed for this project is detailed. Physiological results, and responses to standard care treatment, are detailed and validated. Future updates will include training modules that leverage this model. CONCLUSION: We have created an accurate, whole-body model of burn TBSA training experience in Unreal 4 on a mobile platform, provided for free to the medical community. We hope to provide learners with more a realistic experience and with rapid feedback as they practice patient assessment, intervention, and reassessment.

The effect of autostereoscopic holograms on anatomical knowledge: a randomised trial.

CONTEXT: Three-dimensional (3-D) visualisation in anatomical education has been shown to be broadly beneficial for students. However, there is limited research on the relative efficacy of 3-D modalities. This study compares knowledge performance, mental effort and instructional efficiency between autostereoscopic 3-D visualisation (holograms), monoscopic 3-D visualisation (3-DPDFs) and a control (2-D printed images). METHODS: A cardiac anatomy model was used to generate holograms, 3-DPDFs and 2-D printed images. Nursing student participants (n = 179) were randomised into three groups: holograms (n = 60), 3-DPDFs (n = 60) and printed images (n = 59). Participants completed a pre-test followed by a self-study period using the anatomical visualisation. Afterwards, participants completed the NASA-Task Load Index (NASA-TLX) cognitive load instrument and a knowledge post-test. RESULTS: Post-test results showed participants studying with holograms (median = 80.0, interquartile range [IQR] = 66.7-86.7) performed significantly better regarding cardiac anatomy knowledge than participants using 3-DPDF (median = 66.7, IQR = 53.3-80.0, p = 0.008) or printed images (median = 66.7, IQR = 53.3-80.0, p = 0.007). Mental effort scores, on a scale from 1 to 20, showed hologram (mean = 4.9, standard deviation [SD] = 3.56) and 3-DPDF participants (mean = 4.9, SD = 3.79) reported significantly lower cognitive load than printed images (mean = 7.5, SD = 4.9, p < 0.005). Instructional efficiency (E) of holograms (E = 0.35) was significantly higher than printed images (E = -0.36, p < 0.001), although not significantly higher than 3-DPDF (E = 0.03, p = 0.097). CONCLUSIONS: Participants using holograms demonstrated significant knowledge improvement over printed images and monoscopic 3-DPDF models, suggesting additional depth cues from holographic visualisation provide benefit in understanding spatial anatomy. Mental effort scores and instructional efficiency of holograms indicate holograms are a cognitively efficient instructional medium. These findings highlight the need for further study of novel 3-D technologies and learning performance.

Real-time passive tracking for multi-touch medical modeling and simulation.

The military medical community expects to minimize use of live tissue and cadavers for training purposes. This research demonstrates an innovative use of synthetic tissue and passive tracking computer vision to create a real-time, interactive environment for training medical staff.