Search

Objectives Lung ultrasound (LUS) is a portable, low-cost respiratory imaging tool but is challenged by user dependence and lack of diagnostic specificity. It is unknown whether the advantages of LUS implementation could be paired with deep learning (DL) techniques to match or exceed human-level, diagnostic specificity among similar appearing, pathological LUS images.Design A convolutional neural network (CNN) was trained on LUS images with B lines of different aetiologies. CNN diagnostic performance, as validated using a 10% data holdback set, was compared with surveyed LUS-competent physicians.Setting Two tertiary Canadian hospitals.Participants 612 LUS videos (121 381 frames) of B lines from 243 distinct patients with either (1) COVID-19 (COVID), non-COVID acute respiratory distress syndrome (NCOVID) or (3) hydrostatic pulmonary edema (HPE).Results The trained CNN performance on the independent dataset showed an ability to discriminate between COVID (area under the receiver operating characteristic curve (AUC) 1.0), NCOVID (AUC 0.934) and HPE (AUC 1.0) pathologies. This was significantly better than physician ability (AUCs of 0.697, 0.704, 0.967 for the COVID, NCOVID and HPE classes, respectively), p

Abstract Background Student-run free clinics are critical for medical education and medical care of underserved populations. This paper explores the feasibility of establishing ethical preventative health services at student-run free clinics such as the Rochester Education and Advocacy for Community Health (REACH) Clinic, affiliated with the Mayo Clinic Alix School of Medicine. Our analysis discusses the challenges and ethical considerations of providing preventative healthcare in student-run clinics. This paper also provides recommendations for implementing preventative care at REACH and other student-run clinics. Methods A literature search was conducted by an experienced librarian across four databases using keywords “student-run clinics” and “preventative healthcare.” Additionally, contributing authors identified and gathered data about community service organizations that provide preventative healthcare services to residents of Rochester, MN and surrounding areas. Results The literature search yielded 50 articles, 20 of which were included in the narrative review. Vaccinations and cancer screenings were the two preventative care initiatives most commonly undertaken by student-run clinics. The benefits of student-run clinic initiatives for both patients and students were an emerging theme. Clinics often achieved high vaccination and screening rates, and integration of preventative care topics into medical curricula was valued by students. However, the limitations of student-run clinics also emerged as a theme, particularly limitations to resources and comprehensive follow-up. 15 community health organizations in/around Rochester, MN were identified; of these, five provide preventative healthcare services. Conclusion Student-run clinics like REACH play a key role in delivering preventative care to underserved populations, bridging critical gaps in our healthcare system, and benefiting patients and students alike. However, the ethical principles of beneficence and non-maleficence should be considered when student-run clinics decide what services to offer. Preventative care has been successfully implemented at other student-run clinics but requires careful arrangement of adequate follow-up care, especially when it comes to cancer screening. Vaccinations are a low-cost, high-impact initiative that can be feasibly implemented at most student-run clinics.

Movement disorders such as cervical dystonia, blepharospasm, and hemifacial spasm negatively impact the quality of life of people living with these conditions. Botulinum toxin (BoNT) injections are commonly used to treat these disorders. We sought to describe patient characteristics, BoNT utilization, and potential adverse events (AEs) among patients with cervical dystonia, blepharospasm, and hemifacial spasm using Optum’s de-identified Clinformatics® Data Mart Database. Patients were required to have a diagnosis of the specific condition plus evidence of treatment with BoNT between 8/1/2010 and 5/31/2022. Cervical dystonia patients were commonly females (76%) and aged 45 and older (78%); both blepharospasm and hemifacial spasm patients were commonly females (both 69%) and aged 65 and older (61% and 56%, respectively). Anticholinergics were commonly used (65–82% across cohorts), as were peripheral muscle relaxants for cervical dystonia patients specifically (31%). The median number of injections per year was 2 with the median weeks between injections being between 13 and 15. Of the AEs evaluated, dyspnea was identified frequently across all the cohorts (14–20%). The findings were similar for different BoNT formulations. More research is needed to thoroughly describe BoNT utilization, such as the doses injected, and to optimize treatment for patients with these conditions.

Chronic sialorrhea is a condition characterized by excessive drooling, often associated with neurological and neuromuscular disorders such as Parkinson’s disease, cerebral palsy, and stroke. Despite its prevalence, it remains underdiagnosed and poorly understood, leading to a lack of comprehensive data on patient demographics, clinical characteristics, and treatment patterns. This study aimed to help fill these existing gaps by analyzing real-world data using Optum’s de-identified Clinformatics® Data Mart Database. Patients were required to have a diagnosis indicative of sialorrhea plus evidence of sialorrhea treatment between 1/1/2007 and 5/31/2022. Two cohorts were analyzed: patients with evidence of newly diagnosed sialorrhea and associated treatment, and sialorrhea patients initiating incobotulinumtoxinA. Clinical characteristics, comorbidities, symptoms, and treatment utilization were described before and after diagnosis and incobotulinumtoxinA initiation. No formal statistical comparisons were performed. Patients were predominantly aged 65 or older, male, and non-Hispanic white. Parkinson’s disease and cerebral palsy were the most common comorbidities among adults and children, respectively. Treatment patterns suggest that anticholinergics are more commonly used than botulinum toxin therapy. The findings offer valuable information for improving diagnosis and treatment approaches and suggest a need for further research into treatment effectiveness, safety, and disease burden.

BackgroundThough there has been considerable effort to implement machine learning (ML) methods for health care, clinical implementation has lagged. Incorporating explainable machine learning (XML) methods through the development of a decision support tool using a design thinking approach is expected to lead to greater uptake of such tools. ObjectiveThis work aimed to explore how constant engagement of clinician end users can address the lack of adoption of ML tools in clinical contexts due to their lack of transparency and address challenges related to presenting explainability in a decision support interface. MethodsWe used a design thinking approach augmented with additional theoretical frameworks to provide more robust approaches to different phases of design. In particular, in the problem definition phase, we incorporated the nonadoption, abandonment, scale-up, spread, and sustainability of technology in health care (NASSS) framework to assess these aspects in a health care network. This process helped focus on the development of a prognostic tool that predicted the likelihood of admission to an intensive care ward based on disease severity in chest x-ray images. In the ideate, prototype, and test phases, we incorporated a metric framework to assess physician trust in artificial intelligence (AI) tools. This allowed us to compare physicians’ assessments of the domain representation, action ability, and consistency of the tool. ResultsPhysicians found the design of the prototype elegant, and domain appropriate representation of data was displayed in the tool. They appreciated the simplified explainability overlay, which only displayed the most predictive patches that cumulatively explained 90% of the final admission risk score. Finally, in terms of consistency, physicians unanimously appreciated the capacity to compare multiple x-ray images in the same view. They also appreciated the ability to toggle the explainability overlay so that both options made it easier for them to assess how consistently the tool was identifying elements of the x-ray image they felt would contribute to overall disease severity. ConclusionsThe adopted approach is situated in an evolving space concerned with incorporating XML or AI technologies into health care software. We addressed the alignment of AI as it relates to clinician trust, describing an approach to wire framing and prototyping, which incorporates the use of a theoretical framework for trust in the design process itself. Moreover, we proposed that alignment of AI is dependent upon integration of end users throughout the larger design process. Our work shows the importance and value of engaging end users prior to tool development. We believe that the described approach is a unique and valuable contribution that outlines a direction for ML experts, user experience designers, and clinician end users on how to collaborate in the creation of trustworthy and usable XML-based clinical decision support tools.

Abstract During HIV infection, intron-containing viral mRNAs are exported from the cell nucleus to the cytoplasm to complete the replication cycle. Cellular restrictions on the export of incompletely spliced transcripts are overcome by a viral protein, Rev, and an RNA structure found in all unspliced and incompletely spliced viral mRNAs, the Rev Response Element (RRE). Primary HIV isolates display substantial variation in the sequence and functional activity of Rev proteins. We analyzed Rev from two primary isolates with disparate activity that resulted in differences in in vitro fitness of replication-competent viral constructs. The results showed that amino acid differences within the oligomerization domain, but not the arginine-rich motif or the nuclear export signal, determined the level of Rev activity. Two specific amino acid substitutions were sufficient to alter the low-activity Rev to a high-activity phenotype. Other mutations in Rev sequences had unpredictable effects on activity that differed between the two Rev backbones. The sensitivity of Rev function level to small sequence changes likely permits modulation of Rev-RRE activity during HIV infection, which may play a role in pathogenesis. The functional consequences of Rev mutations differed between primary isolates, highlighting the challenge of generalizing studies of Rev conducted using laboratory HIV strains.

PurposeBevacizumab-related imaging abnormality (BRIA), appearing as areas of restricted diffusion on magnetic resonance imaging (MRI) and representing atypical coagulative necrosis pathologically, has been observed in patients with brain tumors receiving radiotherapy and bevacizumab. We investigated the role of cumulative radiation dose in BRIA development in a voxel-wise analysis.MethodsPatients (n = 18) with BRIA were identified. All had high-grade gliomas or brain metastases treated with radiotherapy and bevacizumab. Areas of BRIA were segmented semi-automatically on diffusion-weighted MRI with apparent diffusion coefficient (ADC) images. To avoid confounding by possible tumor, hypoperfusion was confirmed with perfusion imaging. ADC images and radiation dose maps were co-registered to a high-resolution T1-weighted MRI and registration accuracy was verified. Voxel-wise normal tissue complication probability analyses were performed using a logistic model analyzing the relationship between cumulative voxel equivalent total dose in 2 Gy fractions (EQD2) and BRIA development at each voxel. Confidence intervals for regression model predictions were estimated with bootstrapping.ResultsAmong 18 patients, 39 brain tumors were treated. Patients received a median of 4.5 cycles of bevacizumab and 1-4 radiation courses prior to BRIA appearance. Most (64%) treated tumors overlapped with areas of BRIA. The median proportion of each BRIA region of interest volume overlapping with tumor was 98%. We found a dose-dependent association between cumulative voxel EQD2 and the relative probability of BRIA (β0 = -5.1, β1 = 0.03 Gy-1, γ = 1.3).ConclusionsBRIA is likely a radiation dose-dependent phenomenon in patients with brain tumors receiving bevacizumab and radiotherapy. The combination of radiation effects and tumor microenvironmental factors in potentiating BRIA in this population should be further investigated.

Introduction: Mounting evidence suggests that high-fidelity mannequin-based (HFMBS) and computer-based simulation are useful adjunctive educational tools for advanced cardiac life support (ACLS) instruction. We sought to determine whether access to a supplemental, online computer-based ACLS simulator would improve students’ performance on a standardized Mega Code using high-fidelity mannequin based simulation (HFMBS). Methods: Sixty-five third-year medical students were randomized. Intervention group subjects (n = 29) each received a two-week access code to the online ACLS simulator, whereas the control group subjects (n = 36) did not. Primary outcome measures included students’ time to initiate chest compressions, defibrillate ventricular fibrillation, and pace symptomatic bradycardia. Secondary outcome measures included students’ subjective self-assessment of ACLS knowledge and confidence. Results: Students with access to the online simulator on average defibrillated ventricular fibrillation in 112 seconds, whereas those without defibrillated in 149.9 seconds, an average of 38 seconds faster [p