UNASSIGNED: Large language models (LLMs) have potential to increase the efficiency of information extraction from unstructured clinical notes in electronic medical records.
UNASSIGNED: To assess the utility and reliability of an LLM, ChatGPT-4 (OpenAI), to analyze clinical narratives and identify helmet use status of patients injured in micromobility-related accidents.
UNASSIGNED: This cross-sectional study used publicly available, deidentified 2019 to 2022 data from the US Consumer Product Safety Commission\'s National Electronic Injury Surveillance System, a nationally representative stratified probability sample of 96 hospitals in the US. Unweighted estimates of e-bike, bicycle, hoverboard, and powered scooter-related injuries that resulted in an emergency department visit were used. Statistical analysis was performed from November 2023 to April 2024.
UNASSIGNED: Patient helmet status (wearing vs not wearing vs unknown) was extracted from clinical narratives using (1) a text string search using researcher-generated text strings and (2) the LLM by prompting the system with low-, intermediate-, and high-detail prompts. The level of agreement between the 2 approaches across all 3 prompts was analyzed using Cohen κ test statistics. Fleiss κ was calculated to measure the test-retest reliability of the high-detail prompt across 5 new chat sessions and days. Performance statistics were calculated by comparing results from the high-detail prompt to classifications of helmet status generated by researchers reading the clinical notes (ie, a criterion standard review).
UNASSIGNED: Among 54 569 clinical notes, moderate (Cohen κ = 0.74 [95% CI, 0.73-0.75) and weak (Cohen κ = 0.53 [95% CI, 0.52-0.54]) agreement were found between the text string-search approach and the LLM for the low- and intermediate-detail prompts, respectively. The high-detail prompt had almost perfect agreement (κ = 1.00 [95% CI, 1.00-1.00]) but required the greatest amount of time to complete. The LLM did not perfectly replicate its analyses across new sessions and days (Fleiss κ = 0.91 across 5 trials; P < .001). The LLM often hallucinated and was consistent in replicating its hallucinations. It also showed high validity compared with the criterion standard (n = 400; κ = 0.98 [95% CI, 0.96-1.00]).
UNASSIGNED: This study\'s findings suggest that although there are efficiency gains for using the LLM to extract information from clinical notes, the inadequate reliability compared with a text string-search approach, hallucinations, and inconsistent performance significantly hinder the potential of the currently available LLM.

In the field of industrial safety, wearing helmets plays a vital role in ensuring workers\' health. Aiming at addressing the complex background in the industrial environment, caused by differences in distance, the helmet small target wearing detection methods for misdetection and omission detection problems are needed. An improved YOLOv8 safety helmet wearing detection network is proposed to enhance the capture of details, improve multiscale feature processing and improve the accuracy of small target detection by introducing Dilation-wise residual attention module, atrous spatial pyramid pooling and normalized Wasserstein distance loss function. Experiments were conducted on the SHWD dataset, and the results showed that the mAP of the improved network improved to 92.0%, which exceeded that of the traditional target detection network in terms of accuracy, recall, and other key metrics. These findings further improved the detection of helmet wearing in complex environments and greatly enhanced the accuracy of detection.

UNASSIGNED: Headgear designed to protect girls\' lacrosse athletes is widely available and permitted for voluntary use; however, it remains unknown how policies mandating headgear use may change the sport and, particularly regarding impacts during game-play. Therefore, this study compares the impact rates and game play characteristics of girls\' high school lacrosse in Florida which mandates headgear use (HM), with states having no headgear mandate (NHM).
UNASSIGNED: Video from 189 randomly-selected games (HM: 64, NHM: 125) were analyzed. Descriptive statistics, Impact Rates (IR), Impact Rate Ratios (IRR), Impact Proportion Ratios (IPR), and 95% Confidence Intervals (CI) were calculated. IRRs and IPRs with corresponding CIs that excluded 1.00 were deemed statistically significant.
UNASSIGNED: 16,340 impacts (HM:5,821 NHM: 10,519; 86.6 impacts/game, CI: 88.6-93.3) were identified using the Lacrosse Incident Analysis Instrument (LIAI). Most impacts directly struck the body (n = 16,010, 98%). A minority of impacts directly struck a player\'s head (n = 330, 2%). The rate of head impacts was significantly higher in the HM cohort than NHM cohort (IRR = 2.1; 95% CI = 1.7-2.6). Most head impacts (n = 271, 82%) were caused by stick contact in both groups. There was no difference in the proportion of penalties administered for head impacts caused by stick contact between the HM and NHM cohorts (IPR IRRHM/NHM = 0.98; CI = 0.79-1.16). However, there was a significantly greater proportion of head impacts caused by player contact that resulted in a penalty administered in the HM cohort (IPR = 1.44 CI = 1.17-1.54).
UNASSIGNED: These findings demonstrate that mandating headgear use was associated with a two-fold greater likelihood of sustaining a head impact during game play compared to NHM states. A majority of head impacts in both HM and NHM states were caused by illegal stick contact that did not result in penalty.
High school girls’ lacrosse athletes participating in a state with a headgear mandate was twice as likely to sustain a head impact than those participating in states without headgear mandates.Stick contact remains the most common mechanism of head impacts in girls’ lacrosse, regardless of mandating headgear.Regardless of whether headgear was or was not mandated, most head impacts caused by stick contact did not result in a penalty.

BACKGROUND: Wearing a helmet reduces the risk of head injuries substantially in the event of a motorcycle crash. Countries around the world are committed to promoting helmet use, but the progress has been slow and uneven. There is an urgent need for large-scale data collection for situation assessment and intervention evaluation.
METHODS: This study proposes a scalable, low-cost algorithm to estimate helmet-wearing rates. Applying the state-of-the-art deep learning technique for object detection to images acquired from Google Street View, the algorithm has the potential to provide accurate estimates at the global level.
RESULTS: Trained on a sample of 3995 images, the algorithm achieved high accuracy. The out-of-sample prediction results for all three object classes (helmets, drivers, and passengers) reveal a precision of 0.927, a recall value of 0.922, and a mean average precision at 50 (mAP50) of 0.956.
CONCLUSIONS: The remarkable model performance suggests the algorithm\'s capacity to generate accurate estimates of helmet-wearing rates from an image source with global coverage. The significant enhancement in the availability of helmet usage data resulting from this approach could bolster progress tracking and facilitate evidence-based policymaking for helmet wearing globally.

Wearing helmets is essential in two-wheeler traffic to reduce the incidence of injuries caused by accidents. We present FB-YOLOv7, an improved detection network based on the YOLOv7-tiny model. The objective of this network is to tackle the problems of both missed detection and false detection that result from the difficulties in identifying small targets and the constraints in equipment performance during helmet detection. By applying an enhanced Bi-Level Routing Attention, the network can improve its capacity to extract global characteristics and reduce information distortion. Furthermore, we deploy the AFPN framework and effectively resolve information conflict using asymptotic adaptive feature fusion technology. Incorporating the EfficiCIoU loss significantly improves the prediction box\'s accuracy. Experimental trials done on specific datasets reveal that FB-YOLOv7 attains an accuracy of 87.2% and 94.6% on the mean average precision (mAP@.5). Additionally, it maintains a high level of efficiency with frame rates of 129 and 126 frames per second (FPS). FB-YOLOv7 surpasses the other six widely-used detection networks in terms of detection accuracy, network implementation requirements, sensitivity in detecting small targets, and potential for practical applications.

UNASSIGNED: A cross-sectional survey was conducted using a structured questionnaire involving 402 motorcyclists from four major southeastern towns, comprising 350 (86.07%) males and 52 (12.93%) females. The chi-square test was applied in bivariate analysis, and binary multivariable logistic regression was performed to determine the risk factors of road traffic crashes.
UNASSIGNED: This study\'s findings revealed that the overall reported prevalence of road traffic crashes involving motorcycle drivers over one year was 68.66%. Multivariable logistic regression analysis revealed several factors that significantly impacted road traffic crashes. These factors included driving without a valid driving license, the young age (<20) of motorcyclists, driving in rainy weather, exceeding the speed limit, per-week working hours, smoking status, motorcycle ownership, the brand of motorcycle, and not wearing a helmet while driving.
UNASSIGNED: The study findings highlight the need for improving motorcycle safety by implementing measures such as imposing per-week work hour limits for riders, enforcing traffic regulations, and promoting helmet use among motorcycle drivers. The results of this study draw attention to the Bangladesh Road Transport Authority (BRTA) and motorcycle drivers in the country to decrease motorcycle crashes and the severity of injuries by implementing efficient guidelines and strategies for driving motorcycles. The findings of this study can assist policymakers and concerned authorities in taking the essential steps to lessen road traffic crashes among motorcyclists in Bangladesh.

Concussions, a prevalent public health concern in the United States, often result from mild traumatic brain injuries (mTBI), notably in sports such as American football. There is limited exploration of smart-textile-based sensors for measuring the head impacts associated with concussions in sports and recreational activities. In this paper, we describe the development and construction of a smart textile impact sensor (STIS) and validate STIS functionality under high magnitude impacts. This STIS can be inserted into helmet cushioning to determine head impact force. The designed 2 × 2 STIS matrix is composed of a number of material layered structures, with a sensing surface made of semiconducting polymer composite (SPC). The SPC dimension was modified in the design iteration to increase sensor range, responsiveness, and linearity. This was to be applicable in high impact situations. A microcontroller board with a biasing circuit was used to interface the STIS and read the sensor\'s response. A pendulum test setup was constructed to evaluate various STISs with impact forces. A camera and Tracker software were used to monitor the pendulum swing. The impact forces were calculated by measuring the pendulum bob\'s velocity and acceleration. The performance of the various STISs was measured in terms of voltage due to impact force, with forces varying from 180 to 722 N. Through data analysis, the threshold impact forces in the linear range were determined. Through an analysis of linear regression, the sensors\' sensitivity was assessed. Also, a simplified model was developed to measure the force distribution in the 2 × 2 STIS areas from the measured voltages. The results showed that improving the SPC thickness could obtain improved sensor behavior. However, for impacts that exceeded the threshold, the suggested sensor did not respond by reflecting the actual impact forces, but it gave helpful information about the impact distribution on the sensor regardless of the accurate expected linear response. Results showed that the proposed STIS performs satisfactorily within a range and has the potential to be used in the development of an e-helmet with a large STIS matrix that could cover the whole head within the e-helmet. This work also encourages future research, especially on the structure of the sensor that could withstand impacts which in turn could improve the overall range and performance and would accurately measure the impact in concussion-causing impact ranges.

Background and Objectives: In recent years, electronic scooters (e-scooters) have gained popularity, whether for private use or as a publicly available transportation method. With the introduction of these vehicles, reports of e-scooter-related accidents have surged, sparking public debate and concern. The aim of this study was to analyze the epidemiological data, characteristics, and severity of traumatic brain injury (TBI) related to e-scooter accidents. Materials and Methods: This retrospective case series evaluated patients who were admitted to the three largest neurosurgery clinics in Riga, Latvia, from the time period of April to October in two separate years-2022 and 2023-after e-scooter-related accidents. The data were collected on patient demographics, the time of the accident, alcohol consumption, helmet use, the type of TBI, other related injuries, and the treatment and assessment at discharge. Results: A total of 28 patients were admitted with TBI related to e-scooter use, with a median age of 30 years (Q1-Q3, 20.25-37.25), four individuals under the age of 18, and the majority (64%) being male. In 23 cases, the injury mechanism was falling, in 5 cases, collision. None were wearing a helmet at the time of the injury. Alcohol intoxication was evident in over half of the patients (51.5%), with severe intoxication (>1.2 g/L) in 75% of cases among them. Neurological symptoms upon admission were noted in 50% of cases. All patients had intracranial trauma: 50% had brain contusions, 43% traumatic subdural hematoma, and almost 30% epidural hematoma. Craniofacial fractures were evident in 71% of cases, and there were fractures in other parts of body in three patients. Six patients required emergency neurosurgical intervention. Neurological complications were noted in two patients; one patient died. Conclusions: e-scooter-related accidents result in a significant number of brain and other associated injuries, with notable frequency linked to alcohol influence and a lack of helmet use. Prevention campaigns to raise the awareness of potential risks and the implementation of more strict regulations should be conducted.

UNASSIGNED: Wearing protective helmets is an important prevention strategy to reduce work-related traumatic brain injuries. The existing standardized testing systems are used for quality control and do not provide a quantitative measure of the helmet performance.
UNASSIGNED: To analyze the failure characterizations of Type I industrial helmets and develop a generalized approach to quantify the shock absorption performance of Type I industrial helmets based on the existing standardized setups.
UNASSIGNED: A representative basic Type I construction helmet model was selected for the study. Top impact tests were performed on the helmets at different drop heights using two different impactor masses (3.6 and 5.0 kg).
UNASSIGNED: When the helmets were impacted with potential impact energies smaller than the critical potential impact energy values, there was a consistent relationship between the peak impact force and the potential impact energy. When the helmets were impacted under potential impact energies greater than the critical potential impact energy values, the peak impact forces increased steeply with increasing potential impact energy.
UNASSIGNED: A concept of safety margin for construction helmets based on potential impact energy was introduced to quantify the helmets\' shock absorption performance. The proposed method will help helmet manufacturers improve their product quality.

Introduction: In this study, we sought to develop a thermoplastic patient-specific helmet bolus that could deliver a uniform therapeutic dose to the target and minimize the dose to the normal brain during whole-scalp treatment with a humanoid head phantom. Methods: The bolus material was a commercial thermoplastic used for patient immobilization, and the holes in the netting were filled with melted paraffin. We compared volumetric-modulated arc therapy treatment plans with and without the bolus for quantitative dose distribution analysis. We analyzed the dose distribution in the region of interest to compare dose differences between target and normal organs. For quantitative analysis of treatment dose, OSLD chips were attached at the vertex (VX), posterior occipital (PO), right (RT), and left temporal (LT) locations. Results: The average dose in the clinical target volume was 6553.8 cGy (99.3%) with bolus and 5874 cGy (89%) without bolus, differing by more than 10% from the prescribed dose (6600 cGy) to the scalp target. For the normal brain, it was 3747.8 cGy (56.8%) with bolus and 5484.6 cGy (83.1%) without bolus. These results show that while the dose to the treatment target decreased, the average dose to the normal brain, which is mostly inside the treatment target, increased by more than 25%. With the bolus, the OSLD measured dose was 102.5 ± 1.2% for VX and 101.5 ± 1.9%, 95.9 ± 1.9%, and 81.8 ± 2.1% for PO, RT, and LT, respectively. In addition, the average dose in the treatment plan was 102%, 101%, 93.6%, and 80.7% for VX, PO, RT, and LT. When no bolus was administered, 59.6 ± 2.4%, 112.6 ± 1.8%, 47.1 ± 1.6%, and 53.1 ± 2.3% were assessed as OSLD doses for VX, PO, RT, and LT, respectively. Conclusion: This study proposed a method to fabricate patient-specific boluses that are highly reproducible, accessible, and easy to fabricate for radiotherapy to the entire scalp and can effectively spare normal tissue while delivering sufficient surface dose.