Prior Training Research Articles

Error Management Training and Adaptive Expertise in Learning Computed Tomography Interpretation

Adaptive expertise helps physicians apply their skills to novel clinical cases and reduce preventable errors. Error management training (EMT) has been shown to improve adaptive expertise with procedural skills; however, its application to cognitive skills in medical education is unclear. To evaluate whether EMT improves adaptive expertise when learning the cognitive skill of head computed tomography (CT) interpretation. This 3-arm randomized clinical trial was conducted from July 8, 2022, to March 30, 2023, in 7 geographically diverse emergency medicine residency programs. Participants were postgraduate year 1 through 4 emergency medicine residents masked to the hypothesis. Participants were randomized 1:1:1 to a difficult EMT, easy EMT, or error avoidance training (EAT) control learning strategy for completing an online head CT curriculum. Both EMT cohorts received no didactic instruction before scrolling through head CT cases, whereas the EAT group did. The difficult EMT cohort answered difficult questions about the teaching cases, leading to errors, whereas the easy EMT cohort answered easy questions, leading to fewer errors. All 3 cohorts used the same cases. The primary outcome was a difference in adaptive expertise among the 3 cohorts, as measured using a head CT posttest. Secondary outcomes were (1) differences in routine expertise, (2) whether the quantity of errors during training mediated differences in adaptive expertise, and (3) the interaction between prior residency training and the learning strategies. Among 212 randomized participants (mean [SD] age, 28.8 [2.0] years; 107 men [50.5%]), 70 were allocated to the difficult EMT, 71 to the easy EMT, and 71 to the EAT control cohorts; 150 participants (70.8%) completed the posttest. The difficult EMT cohort outperformed both the easy EMT and EAT cohorts on adaptive expertise cases (60.6% [95% CI, 56.1%-65.1%] vs 45.2% [95% CI, 39.9%-50.6%], vs 40.9% [95% CI, 36.0%-45.7%], respectively; P < .001), with a large effect size (η2 = 0.19). There was no significant difference in routine expertise. The difficult EMT cohort made more errors during training than the easy EMT cohort. Mediation analysis showed that the number of errors during training explained 87.2% of the difficult EMT learning strategy's effect on improving adaptive expertise (P = .01). The difficult EMT learning strategy was more effective in improving adaptive expertise for residents earlier in training, with a large effect size (η2 = 0.25; P = .002). In this randomized clinical trial, the findings show that EMT is an effective method to develop physicians' adaptive expertise with cognitive skills. ClinicalTrials.gov Identifier: NCT05284838.

Movement-Based Prosthesis Control with Angular Trajectory Is Getting Closer to Natural Arm Coordination.

Traditional myoelectric controls of trans-humeral prostheses fail to provide intuitive coordination of the necessary degrees of freedom. We previously showed that by using artificial neural network predictions to reconstruct distal joints, based on the shoulder posture and movement goals (i.e., position and orientation of the targeted object), participants were able to position and orient an avatar hand to grasp objects with natural arm performances. However, this control involved rapid and unintended prosthesis movements at each modification of the movement goal, impractical for real-life scenarios. Here, we eliminate this abrupt change using novel methods based on an angular trajectory, determined from the speed of stump movement and the gap between the current and the 'goal' distal configurations. These new controls are tested offline and online (i.e., involving participants-in-the-loop) and compared to performances obtained with a natural control. Despite a slight increase in movement time, the new controls allowed twelve valid participants and six participants with trans-humeral limb loss to reach objects at various positions and orientations without prior training. Furthermore, no usability or workload degradation was perceived by participants with upper limb disabilities. The good performances achieved highlight the potential acceptability and effectiveness of those controls for our target population.

Implementation of a Virtual Point-of-Care Ultrasound Curriculum at Black Lion Hospital, Ethiopia.

Introduction In low- and middle-income countries (LMICs), the availability of formal point-of-care ultrasound (POCUS) training remains limited, and there is limited data on how to train providers in these countries to use POCUS. This study aimed to describe a virtual training workshop for physicians in Ethiopia, with the intention of serving as a model that could guide similar initiatives. Methods The authors developed and implemented a three-day virtual workshop in 2022 for physicians in the Department of Medicine at Addis Ababa University in Ethiopia. Participants in the workshop completed pre-workshop and post-workshop surveys, as well as knowledge assessments. The authors examined POCUS use prior to the workshop, the impact of the workshop on ratings of comfort level in performing POCUS, and the change in scores on a knowledge assessment before and after the workshop. Results We found that very few of the participants had prior formal POCUS training. Participants reported a significantly higher level of comfort in using POCUS for the assessment of patients (p < 0.001) and for procedural guidance (p < 0.001) after attending the workshop, compared to before. Scores on the post-test knowledge assessment were significantly higher than scores on the pre-test knowledge assessment (p < 0.001). Conclusion Our POCUS workshop was successfully implemented and delivered virtually to a group of physicians in Ethiopia, and it increased comfort levels in performing POCUS and POCUS knowledge. We hope that similar workshops can be implemented in other LMICs.

Generative learning activities for online multimedia learning: when summarizing is effective but drawing is not.

The goal of this study is to determine whether two commonly used generative learning activities for text-based lessons-writing a summary or creating a drawing-help students learn from a multimedia lesson involving animations with short text captions without prior training in the generative activities. Students viewed a series of four annotated animations on greenhouse gases. During pauses between the animations, students were asked to generate a written summary, to create a drawing, or to do both, whereas a control group viewed the lesson without any generative learning activities. Students were tested immediately (Experiment 1) or after a one-week delay (Experiment 2). In both experiments, students who produced written summaries scored significantly higher on the posttest than those who engaged in no generative learning activities (d = 0.48 in Experiment 1, d = 0.54 in Experiment 2), but there was no significant difference on the posttest for students who generated drawings compared to those who engaged in no generative learning activities. In addition, those who engaged in drawing and summarizing did not have significantly different posttest performance than those engaged in summarizing alone. We conclude that writing summaries during a highly visual animated lesson is effective for learning, possibly because it encourages students to engage in generative processing during learning more than drawing and we discuss potential reasons for this in the discussion. This work helps extend generative learning theory by pinpointing potential boundary conditions for learning by drawing and learning by summarizing.

Abstract P107: Training for home blood pressure monitoring among women with a diagnosis or risk factors for hypertensive disorders of pregnancy

Introduction: Hypertensive disorders of pregnancy (HDP) are leading causes of morbidity and mortality among pregnant women. Despite the potential benefits of home blood pressure monitoring (HBPM) for early diagnosis and monitoring of patients with HDP, its utilization is limited by inconsistent training by healthcare providers. This study sought to examine the use of and training for HBPM in those at risk for HDP. Methods: Pregnant women, aged 20-50 (N = 103), diagnosed with either hypertension, HDP, or risk factors for HDP, were recruited from a large academic medical center to complete a survey on HBPM practices from June 2020 to January 2024. Thisincluded questions about HBPM, frequency of HBPM, and prior HBPM training provision by providers. Statistical analysis was performed using STATA/BE 18.0. Results: Of the 103 participants, 41 (39.8%) did HBPM and the most common frequency was daily (70.7%). Of those who did HBPM, 36.6% had received training from their healthcare providers and 31.7% had their BP monitors calibrated in the clinic. There was a trend towards Hispanic women being less likely to receive training compared to white women ( p =0.07). There were no significant differences in training across categories of education and location of residence. Conclusion: In this sample where nearly 40% of pregnant women used HBPM, we observed that only 1 in 3 received training for HBPM with little difference observed across demographic categories. Given the time constraints in clinic visits hindering appropriate training for HBPM, exploring alternative methods of HBPM training such as leveraging mobile health applications may be beneficial to ensure that pregnant women receive necessary training to effectively conduct HBPM.

Curriculum Innovations: Enhancing Skills in Serious Illness Communication in Neurology Residents Using Simulation: A Pilot Study.

Patients with acute ischemic stroke are faced with prognostic uncertainty, progressive decline, and early mortality. Many neurologists report a lack of education and experience in providing palliative care. We developed a simulation-based curriculum to improve residents' confidence and comfort with conducting late-stage goals of care (GOC) conversations. To assess and improve neurology residents' self-reported confidence and comfort around GOC discussions, prognostication, and hospice; encourage neurology residents to conduct GOC conversations early in the illness; introduce neurology residents to a structured framework for conducting GOC conversations; facilitate the residents to build rapport and convey a mindful presence during GOC conversations; provide direct, real-time feedback and an opportunity for redo and practice; and identify gaps for education. The 3-hour experience included a didactic session followed by an interactive simulation and debriefing. The residents' objectives were to deliver difficult news, discuss prognosis, explore goals, navigate treatment options, and discuss end-of-life care including hospice. The faculty observed each interaction and called time-outs to allow the residents to self-assess and obtain feedback. Residents and faculty debriefed to identify take-home points and to reflect on their emotions, self-care, and sense of purpose in medicine. Twenty-six neurology residents filled out an anonymous presurvey to self-assess their confidence and comfort surrounding GOC conversations. More than 50% of residents reported being confident in conducting GOC discussions, whereas only 42% reported adequate prior training. Postsession, more than 90% of residents reported that training was relevant, helpful, organized, and clear. Faculty identified that residents had difficulty addressing prognosis, assessing goals, planning treatment, and using silence, responding to emotion, and displaying empathy. Fifteen residents filled out a postsurvey that revealed improved comfort with delivering prognosis, discussing hospice, and initiating early GOC discussions. Our project uniquely focuses on late-stage GOC conversations and builds on existing literature that supports a structured program with both didactic and simulation components to improve residents' abilities to effectively navigate GOC conversations with patients and families. Future work will focus on reinforcement and reassessment of communication skills.

Neuromuscular adaptations to perturbation-based balance training using treadmill belt accelerations do not transfer to an obstacle trip in older people: A cross-over randomised controlled trial

BackgroundThis study examined (i) adaptations in muscle activity following perturbation-based balance training (PBT) using treadmill belt-accelerations or PBT using walkway trips and (ii) whether adaptations during treadmill PBT transfer to a walkway trip. MethodsThirty-eight older people (65+ years) undertook two PBT sessions, including 11 treadmill belt-accelerations and 11 walkway trips. Surface electromyography (EMG) was measured bilaterally on the rectus femoris (RF), tibialis anterior (TA), semitendinosus (ST) and gastrocnemius medial head (GM) during the first (T1) and eleventh (T11) perturbations. Adaptations (within-subjects - 1st vs 11th perturbations for treadmill and walkway PBT) and their transfer (between-subjects – 1st walkway trip after treadmill PBT vs 1st walkway trip with no prior training) effects were examined for the EMG parameters. ResultsTreadmill PBT reduced post-perturbation peak muscle activation magnitude (left RF, TA, ST, right RF, ST, GM), onset latency (right TA), time to peak (right RF) and co-contraction index (knee muscles) (P < 0.05). Walkway PBT reduced post-trip onset latencies (right TA, ST), peak magnitude (left ST, right GM), time to peak (right RF, ST) and pre-perturbation muscle activity (right TA) (P < 0.05). Those who undertook treadmill PBT were not different to those without prior training during the first walkway trip (P > 0.05). ConclusionsBoth treadmill and walkway PBT induced earlier initiation and peak activation of right limb muscles responsible for the first recovery step. Treadmill PBT also reduced co-contraction of the knee muscles. Adaptations in muscle activity following treadmill PBT did not transfer to a walkway trip.

Predictors of hand hygiene attitudes among Saudi healthcare workers of the intensive care unit in Saudi Arabia

Healthcare-associated infections present a challenge to healthcare systems, particularly critical care units. Hand hygiene emerges as a crucial element in infection control, acting as a vital link between healthcare workers, patients, and pathogens. Positive attitudes, motivated by a genuine concern for patient safety, are recognized as major predictors of hand hygiene compliance among healthcare workers. This study aims to assess the attitudes of ICU staff toward hand hygiene and identify factors that influence these attitudes. A cross-sectional survey of intensive care unit staff in seven large Saudi hospitals was conducted using an anonymous, self-reporting questionnaire to examine the attitudes of ICU personnel about hand hygiene and determine the factors that impact these attitudes. A regression analysis was used to determine the determinants of hand hygiene attitudes. Of the 600 respondents, 93% rated their hand hygiene knowledge as good, and 71% received hand hygiene training from their hospital. Most respondents (78%) had previously experienced healthcare-associated infections. The majority reported a favorable overall attitude toward hand hygiene (M = 4.15, SD = 0.85). Attitudes toward hand hygiene were significantly associated with perceived knowledge (β = 0.32, p < 0.001), prior hand hygiene training (β = 0.13, p < 0.05), and years of experience (β = − 0.10, p < 0.05). Healthcare workers in the ICU have a generally positive attitude toward hand hygiene. Such attitudes were correlated with hand hygiene knowledge, prior training, and job experience. The study’s findings can help to inform health promotion initiatives and campaigns aimed at achieving long-term improvements in hand hygiene behaviors.

Using simulated patients to teach de-escalation during Registered Nurses' onboarding

BackgroundWorkplace violence and aggression toward healthcare professionals are concerning trends in the industry, with increasing incidents over the last decade. The COVID-19 pandemic put additional stress on health systems, healthcare workers, patients, and family members. These additional stressors resulted in increased aggressive behaviors. Education and training on de-escalation techniques is paramount to increased safety of health professionals and patients. Nursing education programs and hospital orientation programs often lack de-escalation training. MethodsFrom January to October 2023, 102 nurses participated in a de-escalation simulation encounter. Quantitative and qualitative data was captured via Redcap survey. Facilitators tracked behaviors of the simulated patients and reactions of the nurses, in real-time, using a survey designed by the education team. Participants completed a post-survey which elicited information about prior de-escalation training and application of skills learned during the de-escalation simulation encounter to future practice. ResultsSixty-five percent of new graduate nurses revealed that they had some form of de-escalation training most of which was provided through a lecture (36%) or online module (19%). Following participation in the simulation encounter, 98.5% of participants felt better prepared to manage an escalating patient/family member situation. Structured debriefings identified educational gaps regarding hospital resources, behavioral strategies for averting escalating incidents, methods for maintaining safety, and discreetly seeking assistance. ConclusionThe simulated patient training offered participants an opportunity to apply de-escalation techniques in a safe environment. Targeted de-escalation training using simulated patients increased the authenticity and realism of the immersive experience and resulted in enhanced de-escalation skill development in new hire nurses.

DTL-5G: Deep transfer learning-based DDoS attack detection in 5G and beyond networks

Network slicing is considered as a key enabler for 5G and beyond mobile networks for supporting a variety of new services, including enhanced mobile broadband, ultra-reliable and low-latency communication, and massive connectivity, on the same physical infrastructure. However, this technology increases the susceptibility of networks to cyber threats, particularly Distributed Denial-of-Service (DDoS) attacks. These attacks have the potential to cause service quality degradation by overloading network function(s) that are central to network slices to operate seamlessly. This calls for an Intrusion Detection System (IDS) as a shield against a wide array of DDoS attacks. In this regard, one promising solution would be the use of Deep Learning (DL) models for detecting possible DDoS attacks, an approach that has already made its way into the field given its manifest effectiveness. However, one particular challenge with DL models is that they require large volumes of labeled data for efficient training, which are not readily available in operational networks. A possible workaround is to resort to Transfer Learning (TL) approaches that can utilize the knowledge learned from prior training to a target domain with limited labeled data. This paper investigates how Deep Transfer Learning (DTL) based approaches can improve the detection of DDoS attacks in 5G networks by leveraging DL models, such as Bidirectional Long Short-Term Memory (BiLSTM), Convolutional Neural Network (CNN), Residual Network (ResNet), and Inception as base models. A comprehensive dataset generated in our 5G network slicing testbed serves as the source dataset for DTL, which includes both benign and different types of DDoS attack traffic. After learning features, patterns, and representations from the source dataset using initial training, we fine-tune base models using a variety of TL processes on a target DDoS attack dataset. The 5G-NIDD dataset, which has a sparse amount of annotated traffic pertaining to several DDoS attack generated in a real 5G network, is chosen as the target dataset. The results show that the proposed DTL models have performance improvements in detecting different types of DDoS attacks in 5G-NIDD dataset compared to the case when no TL is applied. According to the results, the BiLSTM and Inception models being identified as the top-performing models. BiLSTM indicates an improvement of 13.90%, 21.48%, and 12.22% in terms of accuracy, recall, and F1-score, respectively, whereas, Inception demonstrates an enhancement of 10.09% in terms of precision, compared to the models that do not adopt TL.

Barriers to diffusion and implementation of pediatric minimally invasive surgery in Brazil

BackgroundThe main barriers to the broad implementation of pediatric minimally invasive surgery (MIS) are technological, technical, and epistemological barriers, as well as the rarity of certain pathologies. These issues are presumed to be more significant in low- and middle-income countries. This study aimed to identify and analyze the factors limiting the diffusion and implementation of pediatric MIS in Brazil.MethodsA nationwide cross-sectional survey was conducted via an online questionnaire in Brazil from January 2022 to July 2022. The sample was taken by convenience from the population of pediatric surgeons in Brazil. A total of 187 surgeons were included. The collected data were divided into three sections to evaluate technological, technical, and epistemological limitations to pediatric MIS implementation.ResultsAlthough 85% of the participants had previous training, a lack of adequate training was identified as a significant limiting factor, particularly among those who had taken only short courses (42.3% vs. 64.3%, p = 0.033). Only 14% of the participants reported performing MIS for major pediatric procedures. With respect to intracorporeal suturing, 38.1% of the surgeons with extensive training considered it a limiting factor compared with 60.7% (p = 0.029) of those without prior training. Among those without previous training, 61% cited a lack of financial support or encouragement from their department as the reason. Additionally, 65% of the surgeons considered the lack of basic instruments a limiting factor. Although 95% of the participants agreed that simulation training is indispensable, pediatric surgery fellowship programs in Brazil do not include a standardized curriculum or mandatory training in MIS, and only 47% reported providing training space for their current fellows.ConclusionA combination of technological, technical, and epistemological barriers hinders the implementation of pediatric MIS. Despite its limitations, this study serves as a foundational guide for future analysis and overcoming the identified barriers.

Integration of Animal Behavioral Assessment and Convolutional Neural Network to Study Wasabi-Alcohol Taste-Smell Interaction.

The commercial wasabi pastes commonly used for food preparation contain a homologous compound of chemosensory isothiocyanates (ITCs) that elicit an irritating sensation upon consumption. The impact of sniffing dietary alcoholic beverages on the sensation of wasabi spiciness has never been studied. While most sensory evaluation studies focus on individual food and beverages separately, there is a lack of research on the olfactory study of sniffing liquor while consuming wasabi. Here, a methodology is developed that combines the use of an animal behavioral study and a convolutional neural network to analyze the facial expressions of mice when they simultaneously sniff liquor and consume wasabi. The results demonstrate that the trained and validated deep learning model recognizes 29% of the images depicting co-treatment of wasabi and alcohol belonging to the class of the wasabi-negative liquor-positive group without the need for prior training materials filtering. Statistical analysis of mouse grimace scale scores obtained from the selected video frame images reveals a significant difference (P < 0.01) between the presence and absence of liquor. This finding suggests that dietary alcoholic beverages might have a diminishing effect on the wasabi-elicited reactions in mice. This combinatory methodology holds potential for individual ITC compound screening and sensory analyses of spirit components in the future. However, further study is required to investigate the underlying mechanism of alcohol-induced suppression of wasabi pungency.

Comparing Male and Female Resident Physicians in Central Venous Catheter Insertion Self-confidence and Competency: A Retrospective Cohort Study.

Female physicians often report lower self-confidence in their procedural and clinical competency compared to male physicians. There is limited data regarding self-reported confidence of female versus male trainees and any relation to objective competency in central venous catheter insertion. To analyze differences between male and female trainees in self-confidence and skill-based outcomes in placing central venous catheters. Using data from a central venous catheter simulation training program at a large tertiary medical center, we performed linear regressions to analyze confidence difference pre- and post-training, number of restarts, and number of cannulation attempts while controlling for baseline demographic characteristics of the sample. PGY-1 physician residents in all residency specialties who insert central venous catheters in the clinical setting at a tertiary academic center with a sample size of 281 residents. Confidence difference pre- and post-training measured on a Likert scale 1-5, number of restarts (novel global assessment variable), and number of cannulation attempts during the competency evaluation. Female trainees had both lower pre-program confidence (1.35 versus 1.74 out of 5, p < 0.001) and lower post-program confidence (3.77 versus 4.12 out of 5, p = 0.0021) as compared to male trainees. There was no statistically significant difference in number of restarts (95% CI - 0.073 to 0.368, p = 0.185) or cannulation attempts (95% CI - 0.039 to 0.342, p = 0.117) between sexes in linear regressions controlled for age, specialty designation, prior central venous catheter training, prior ultrasound guided vessel cannulation training, and pre-training confidence level. Female trainees rated their confidence significantly lower than their male counterparts both before and after the training program, despite no significant difference in skill-based outcomes. We discuss potential implications for trainees acquiring procedural skills during residency and for physician educators as they design training programs and delegate procedural opportunities.

Mitigation of Subsequent Ovariectomy Responses through Prior Exercise Training in Rats.

It is well known that cardiometabolic dysfunction gradually increases after menopause, and the sedentary lifestyle can aggravate this condition. Therefore, we compared the effects of aerobic exercise training during the premenopausal period and after ovariectomy (OVX) on metabolic, hemodynamic, and autonomic parameters in an experimental rat model of menopause. The female rats were divided into four groups: control (C), sedentary OVX (SO), trained OVX (TO), and previously trained OVX (PTO). The PTO group was trained for 4 weeks prior to+8 weeks after OVX, and the TO group trained only after OVX on a motor treadmill. Autonomic modulation was evaluated, white adipose tissue (WAT) was removed and weighed, and lipolysis was assessed. The citrate synthase activity in the soleus muscle was analyzed. The trained groups prevented the impairment of baroreceptor reflex sensitivity in relation to SO; however, only PTO reduced the low-frequency band of the pulse interval compared to SO. PTO reduced the weight of WAT compared to the other groups; lipolysis in PTO was similar to that in C. PTO preserved muscle metabolic injury in all types of fibers analyzed. In conclusion, this study suggests that exercise training should be recommended in a premenopausal model to prevent cardiometabolic and autonomic menopause-induced deleterious effects.

Exploring pain management practices among nurses in a provincial city of Vietnam: A cross-sectional investigation of influential factors

Background &amp; Aim: There is limited evidence regarding pain management practice and its associated factors among nurses in Vietnam. Therefore, the study aimed 1) To examine the current state of pain management practices among nurses at Hai Duong Provincial General Hospital, Vietnam; and 2) to identify its associated factors. Methods &amp; Materials: A cross-sectional study design was performed on 200 nurses and 200 medical records needed to be reviewed to match the number of nursing samples. The data collection was conducted from May to June 2022. The modified Cancer Pain Practice Index included 13 evidence-based pain management Indicators applied to measure nurses’ pain management practice by reviewing medical records, and the Knowledge and Attitudes Survey Regarding Pain tool was used to assess the pain management knowledge and attitude of nurses by self-completed questionnaire. Results: A significant majority, comprising 146 nurses (73%), demonstrated inadequate pain management practices. Several crucial indicators of pain management were suboptimally performed by the majority of nurses. These included failure to document the execution of pain assessments for new patients, inadequate monitoring for side effects of opioid analgesics, and insufficient assessment and care for bowel status when opioids were prescribed. Additionally, participation in prior pain management training courses, along with knowledge and attitude toward pain management, emerged as significant variables associated with pain management practices (p&lt;0.001). Conclusion: The study highlighted widespread poor pain management practices among nurses. Participation in training courses and good knowledge and attitude correlate with better practices.

A Graph Reinforcement Learning Framework for Neural Adaptive Large Neighbourhood Search

Adaptive Large Neighbourhood Search (ALNS) is a popular metaheuristic with renowned efficiency in solving combinatorial optimisation problems. However, despite 18 years of intensive research into ALNS, the design of an effective adaptive layer for selecting operators to improve the solution remains an open question. In this work, we isolate this problem by formulating it as a Markov Decision Process, in which an agent is rewarded proportionally to the improvement of the incumbent. We propose Graph Reinforcement Learning for Operator Selection (GRLOS), a method based on Deep Reinforcement Learning and Graph Neural Networks, as well as Learned Roulette Wheel (LRW), a lightweight approach inspired by the classic Roulette Wheel adaptive layer. The methods, which are broadly applicable to optimisation problems that can be represented as graphs, are comprehensively evaluated on 5 routing problems using a large portfolio of 28 destroy and 7 repair operators. Results show that both GRLOS and LRW outperform the classic selection mechanism in ALNS, owing to the operator choices being learned in a prior training phase. GRLOS is also shown to consistently achieve better performance than a recent Deep Reinforcement Learning method due to its substantially more flexible state representation. The evaluation further examines the impact of the operator budget and type of initial solution, and is applied to problem instances with up to 1000 customers. The findings arising from our extensive benchmarking bear relevance to the wider literature of hybrid methods combining metaheuristics and machine learning.

Diagnostic Accuracy of an Integrated AI Tool to Estimate Gestational Age From Blind Ultrasound Sweeps

Accurate assessment of gestational age (GA) is essential to good pregnancy care but often requires ultrasonography, which may not be available in low-resource settings. This study developed a deep learning artificial intelligence (AI) model to estimate GA from blind ultrasonography sweeps and incorporated it into the software of a low-cost, battery-powered device. To evaluate GA estimation accuracy of an AI-enabled ultrasonography tool when used by novice users with no prior training in sonography. This prospective diagnostic accuracy study enrolled 400 individuals with viable, single, nonanomalous, first-trimester pregnancies in Lusaka, Zambia, and Chapel Hill, North Carolina. Credentialed sonographers established the "ground truth" GA via transvaginal crown-rump length measurement. At random follow-up visits throughout gestation, including a primary evaluation window from 14 0/7 weeks' to 27 6/7 weeks' gestation, novice users obtained blind sweeps of the maternal abdomen using the AI-enabled device (index test) and credentialed sonographers performed fetal biometry with a high-specification machine (study standard). The primary outcome was the mean absolute error (MAE) of the index test and study standard, which was calculated by comparing each method's estimate to the previously established GA and considered equivalent if the difference fell within a prespecified margin of ±2 days. In the primary evaluation window, the AI-enabled device met criteria for equivalence to the study standard, with an MAE (SE) of 3.2 (0.1) days vs 3.0 (0.1) days (difference, 0.2 days [95% CI, -0.1 to 0.5]). Additionally, the percentage of assessments within 7 days of the ground truth GA was comparable (90.7% for the index test vs 92.5% for the study standard). Performance was consistent in prespecified subgroups, including the Zambia and North Carolina cohorts and those with high body mass index. Between 14 and 27 weeks' gestation, novice users with no prior training in ultrasonography estimated GA as accurately with the low-cost, point-of-care AI tool as credentialed sonographers performing standard biometry on high-specification machines. These findings have immediate implications for obstetrical care in low-resource settings, advancing the World Health Organization goal of ultrasonography estimation of GA for all pregnant people. ClinicalTrials.gov Identifier: NCT05433519.

Youth Mental Health First Aid Training: Impact on the Ability to Recognize and Support Youth Needs

Youth Mental Health First Aid (YMHFA) trains individuals who regularly interact with youth to identify youth experiencing mental health challenges. Several studies demonstrate positive training impacts, but few assess whether the training equally impacts participants of different demographic and professional backgrounds or those who participate in different training modalities. Using a pre-post follow-up design with a comparison group, this study examined changes in participants’ confidence in their ability to recognize and support youth mental health needs 1 to 2 months after training. Data were collected over two years (2021–2023) from training participants (n = 480) and comparable non-participants (n = 51). The authors examined whether changes in confidence varied by participant race/ethnicity, professional role in the education or mental health fields, and training modality (online versus hybrid). Training participants’ confidence in supporting youth mental health increased significantly compared to non-participants. Although the training was effective for all participants, those with less mental health experience benefited more, consistent with previous research. While both in-person and hybrid training were effective, in-person training participants reported slightly higher confidence scores than virtual at follow-up. Study findings suggest that educational and social service organizations should offer this training to their staff and community members who interact with youth, prioritizing participants with less prior mental health training and delivering training through an in-person training modality when possible. However, additional research is needed to explore how aspects of in-person training, such as trainer characteristics and group dynamics, impact outcomes.

Shared Decision-Making in Colorectal Cancer Screening for Older Adults

Decisions about whether to stop colorectal cancer (CRC) screening tests in older adults can be difficult and may benefit from shared decision-making (SDM). To evaluate the effect of physician training in SDM and electronic previsit reminders (intervention) vs reminders only (comparator) on receipt of the patient-preferred approach to CRC screening and on overall CRC screening rates of older adults at 12 months. This was a secondary analysis of the Promoting Informed Decisions About Colorectal Cancer Screening in Older Adults (PRIMED) cluster randomized clinical trial. In the PRIMED trial, primary care physicians (PCPs) from 36 primary care practices in Massachusetts and Maine were enrolled between May 1 and August 30, 2019, and were randomized to the intervention group or the comparator group. Patients aged 76 to 85 years who were overdue for CRC screening and did not have a prior diagnosis of CRC enrolled between October 21, 2019, and April 8, 2021. Data analysis was performed between May 24, 2022, and May 10, 2023. Primary care physicians in the intervention group completed an SDM training course and received previsit reminders of patients eligible for CRC testing discussion, whereas PCPs in the comparator group received reminders only. The primary outcome was concordance, or the percentage of patients who received their preferred screening approach. Postvisit surveys were administered to assess patient preference for testing, and electronic health record review was used to assess CRC testing at 12 months. Heterogeneity of treatment effect analyses examined interaction between study groups and different factors on concordance rates. This study included 59 physicians and 466 older adults. Physicians had a mean (SD) age of 52.7 (9.4) years and a mean (SD) of 21.6 (10.2) years in practice; 30 (50.8%) were women and 16 (27.1%) reported prior training in SDM. Patients had a mean (SD) age of 80.3 (2.8) years; 249 (53.4%) were women and 238 (51.1%) reported excellent or very good overall health. Patients preferred stool-based tests (161 [34.5%]), followed by colonoscopy (116 [24.8%]) or no further screening (97 [20.8%]); 75 (16.1%) were not sure. The distribution of patient preferences was similar across groups (P = .36). At 12 months, test uptake was also similar for both the intervention group (29 [12.3%] for colonoscopy, 62 [26.3%] for stool-based tests, and 145 [61.4%] for no testing) and the comparator group (32 [13.9%] for colonoscopy, 35 [15.2%] for stool-based tests, and 163 [70.9%] for no testing; P = .08). Approximately half of patients in the intervention group received their preferred approach vs the comparator group (115 of 226 [50.9%] vs 103 of 223 [46.2%]; P = .47). Heterogeneity of treatment effect analyses found significantly higher rates with the intervention vs the comparator for patients with a strong intention to follow through with the preferred approach (adjusted odds ratio [AOR], 1.79 [95% CI, 1.11-2.89]; P = .02, P = .05 for interaction) and for patients who reported more than 5 minutes (AOR, 3.27 [95% CI, 1.25-8.59]; P = .02, P = .05 for interaction) of discussion with their PCP regarding screening. Higher rates were also observed among patients who reported 2 to 5 minutes of discussion with their PCP, although this finding was not significant (AOR, 1.89 [95% CI, 0.93-3.84]; P = .08, P = .05 for interaction). In this secondary analysis of a cluster randomized clinical trial, approximately half of older patients received their preferred approach to CRC screening. Physician training in SDM did not result in higher concordance rates overall but may have benefitted some subgroups. Future work to refine and evaluate clinical decision support (in the form of an electronic advisory or reminder) as well as focused SDM skills training for PCPs may promote high-quality, preference-concordant decisions about CRC testing for older adults. ClinicalTrials.gov Identifier: NCT03959696.

Fun with electricity: A novel ballistics gelatin model with LED tracking for ultrasound needle guidance.

Use of ultrasound (US) for procedural needle guidance can improve success rates, safety, and accuracy. Often, training is performed on task trainers, which can be prohibitively expensive. Determining undesired needle placement is difficult when the needle is poorly visualized with US. Currently available simulation phantoms cannot provide real-time feedback on the location of needle placement. The primary objective was to develop and determine feasibility of a low-cost simulation phantom with an internal circuit and LED light system to determine when a needle contacts internal structures. We also aimed to determine whether its use was associated with increased comfort level. Emergency medicine (EM) residents (PGY-1 to PGY-3) performed in-plane and out-of-plane US needle guidance using homemade phantoms. Comfort levels were assessed by pre- and post intervention survey. Outcomes were measured on Likert scale (minimum = 1, maximum = 5). The primary outcome was change in confidence markers before and after the simulation task. Secondary outcomes were survey results of comparisons of these models to prior training experiences on simulators and humans. All EM residents (30) in our program were invited to participate. Twenty participants enrolled and completed the study. In the primary outcome, median comfort with out-of-plane and in-plane guidance increased after using the model but was more pronounced for out-of-plane guidance. On a posttest survey, residents rated the models overall very similar to prior experience on simulators (median 5/5 [IQR 4.0-5.0]) and moderately similar to humans (median 3/5 [IQR 3.0-4.0]). We created a low-cost ballistic gelatin phantom with an internal electric needle guidance system. Use of the phantom for training was associated with increased learner comfort with the procedure. Learners rated the characteristics of the phantom as similar to higher-cost commercial equipment and humans.