COVID-Net USPro: An Explainable Few-Shot Deep Prototypical Network for COVID-19 Screening Using Point-of-Care Ultrasound.

As the Coronavirus Disease 2019 (COVID-19) continues to impact many aspects of life and the global healthcare systems, the adoption of rapid and effective screening methods to prevent the further spread of the virus and lessen the burden on healthcare providers is a necessity. As a cheap and widely accessible medical image modality, point-of-care ultrasound (POCUS) imaging allows radiologists to identify symptoms and assess severity through visual inspection of the chest ultrasound images. Combined with the recent advancements in computer science, applications of deep learning techniques in medical image analysis have shown promising results, demonstrating that artificial intelligence-based solutions can accelerate the diagnosis of COVID-19 and lower the burden on healthcare professionals. However, the lack of large, well annotated datasets poses a challenge in developing effective deep neural networks, especially in the case of rare diseases and new pandemics. To address this issue, we present COVID-Net USPro, an explainable few-shot deep prototypical network that is designed to detect COVID-19 cases from very few ultrasound images. Through intensive quantitative and qualitative assessments, the network not only demonstrates high performance in identifying COVID-19 positive cases, using an explainability component, but it is also shown that the network makes decisions based on the actual representative patterns of the disease. Specifically, COVID-Net USPro achieves 99.55% overall accuracy, 99.93% recall, and 99.83% precision for COVID-19-positive cases when trained with only five shots. In addition to the quantitative performance assessment, our contributing clinician with extensive experience in POCUS interpretation verified the analytic pipeline and results, ensuring that the network's decisions are based on clinically relevant image patterns integral to COVID-19 diagnosis. We believe that network explainability and clinical validation are integral components for the successful adoption of deep learning in the medical field. As part of the COVID-Net initiative, and to promote reproducibility and foster further innovation, the network is open-sourced and available to the public.

Towards Building a Trustworthy Deep Learning Framework for Medical Image Analysis.

Computer vision and deep learning have the potential to improve medical artificial intelligence (AI) by assisting in diagnosis, prediction, and prognosis. However, the application of deep learning to medical image analysis is challenging due to limited data availability and imbalanced data. While model performance is undoubtedly essential for medical image analysis, model trust is equally important. To address these challenges, we propose TRUDLMIA, a trustworthy deep learning framework for medical image analysis, which leverages image features learned through self-supervised learning and utilizes a novel surrogate loss function to build trustworthy models with optimal performance. The framework is validated on three benchmark data sets for detecting pneumonia, COVID-19, and melanoma, and the created models prove to be highly competitive, even outperforming those designed specifically for the tasks. Furthermore, we conduct ablation studies, cross-validation, and result visualization and demonstrate the contribution of proposed modules to both model performance (up to 21%) and model trust (up to 5%). We expect that the proposed framework will support researchers and clinicians in advancing the use of deep learning for dealing with public health crises, improving patient outcomes, increasing diagnostic accuracy, and enhancing the overall quality of healthcare delivery.

Postpandemic Evaluation of the Eco-Efficiency of Personal Protective Equipment Against COVID-19 in Emergency Departments: Proposal for a Mixed Methods Study.

BACKGROUND: The COVID-19 pandemic has had a profound impact on emergency department (ED) care in Canada and around the world. To prevent transmission of COVID-19, personal protective equipment (PPE) was required for all ED care providers in contact with suspected cases. With mass vaccination and improvements in several infection prevention components, our hypothesis is that the risks of transmission of COVID-19 will be significantly reduced and that current PPE use will have economic and ecological consequences that exceed its anticipated benefits. Evidence is needed to evaluate PPE use so that recommendations can ensure the clinical, economic, and environmental efficiency (ie, eco-efficiency) of its use. OBJECTIVE: To support the development of recommendations for the eco-efficient use of PPE, our research objectives are to (1) estimate the clinical effectiveness (reduced transmission, hospitalizations, mortality, and work absenteeism) of PPE against COVID-19 for health care workers; (2) estimate the financial cost of using PPE in the ED for the management of suspected or confirmed COVID-19 patients; and (3) estimate the ecological footprint of PPE use against COVID-19 in the ED. METHODS: We will conduct a mixed method study to evaluate the eco-efficiency of PPE use in the 5 EDs of the CHU de Québec-Université Laval (Québec, Canada). To achieve our goals, the project will include four phases: systematic review of the literature to assess the clinical effectiveness of PPE (objective 1; phase 1); cost estimation of PPE use in the ED using a time-driven activity-based costing method (objective 2; phase 2); ecological footprint estimation of PPE use using a life cycle assessment approach (objective 3; phase 3); and cost-consequence analysis and focus groups (integration of objectives 1 to 3; phase 4). RESULTS: The first 3 phases have started. The results of these phases will be available in 2023. Phase 4 will begin in 2023 and results will be available in 2024. CONCLUSIONS: While the benefits of PPE use are likely to diminish as health care workers' immunity increases, it is important to assess its economic and ecological impacts to develop recommendations to guide its eco-efficient use. TRIAL REGISTRATION: PROSPERO CRD42022302598; https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=302598. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): DERR1-10.2196/50682.

COVIDx-US: An Open-Access Benchmark Dataset of Ultrasound Imaging Data for AI-Driven COVID-19 Analytics.

BACKGROUND: The Coronavirus Disease 2019 (COVID-19) pandemic continues to have a devastating effect on the health and well-being of the global population. Apart from the global health crises, the pandemic has also caused significant economic and financial difficulties and socio-physiological implications. Effective screening, triage, treatment planning, and prognostication of outcome play a key role in controlling the pandemic. Recent studies have highlighted the role of point-of-care ultrasound imaging for COVID-19 screening and prognosis, particularly given that it is non-invasive, globally available, and easy-to-sanitize. COVIDx-US Dataset: Motivated by these attributes and the promise of artificial intelligence tools to aid clinicians, we introduce COVIDx-US, an open-access benchmark dataset of COVID-19 related ultrasound imaging data. The COVIDx-US dataset was curated from multiple data sources and its current version, i.e., v1.5., consists of 173 ultrasound videos and 21,570 processed images across 147 patients with COVID-19 infection, non-COVID-19 infection, other lung diseases/conditions, as well as normal control cases. CONCLUSIONS: The COVIDx-US dataset was released as part of a large open-source initiative, the COVID-Net initiative, and will be continuously growing, as more data sources become available. To the best of the authors' knowledge, COVIDx-US is the first and largest open-access fully-curated benchmark lung ultrasound imaging dataset that contains a standardized and unified lung ultrasound score per video file, providing better interpretation while enabling other research avenues such as severity assessment. In addition, the dataset is reproducible, easy-to-use, and easy-to-scale thanks to the well-documented modular design.

Understanding the temporal evolution of COVID-19 research through machine learning and natural language processing.

The outbreak of the novel coronavirus disease 2019 (COVID-19), caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has been continuously affecting human lives and communities around the world in many ways, from cities under lockdown to new social experiences. Although in most cases COVID-19 results in mild illness, it has drawn global attention due to the extremely contagious nature of SARS-CoV-2. Governments and healthcare professionals, along with people and society as a whole, have taken any measures to break the chain of transition and flatten the epidemic curve. In this study, we used multiple data sources, i.e., PubMed and ArXiv, and built several machine learning models to characterize the landscape of current COVID-19 research by identifying the latent topics and analyzing the temporal evolution of the extracted research themes, publications similarity, and sentiments, within the time-frame of January-May 2020. Our findings confirm the types of research available in PubMed and ArXiv differ significantly, with the former exhibiting greater diversity in terms of COVID-19 related issues and the latter focusing more on intelligent systems/tools to predict/diagnose COVID-19. The special attention of the research community to the high-risk groups and people with complications was also confirmed.

Direct-access physiotherapy to help manage patients with musculoskeletal disorders in an emergency department: Results of a randomized controlled trial.

OBJECTIVES: The objective was to evaluate the effects of direct-access physiotherapy on patients presenting with a musculoskeletal disorder (MSKD) to the emergency department (ED) on clinical outcomes and use of health care resources. METHODS: We conducted a randomized controlled trial in an academic ED in Québec City, Canada. We included patients aged 18 to 80 years with minor MSKD. The intervention group had direct access to a physiotherapist (PT) in the ED immediately after triage and prior to physician assessment, and the control group received usual care by the emergency physician without PT intervention. The key variables included clinical outcomes (pain, interference of pain on function) and resources use (ED return visit, medications, diagnostic tests, additional consultations). They were analyzed using descriptive statistics and compared between groups using two-way analyses of variance, log-linear analysis, and chi-square tests. RESULTS: Seventy-eight patients suffering from MSKDs were included (40.2 ± 17.6 years old; 44% women). For the primary clinical outcome, participants in the PT group (n = 40) had statistically lower levels of pain and pain interference at 1 and 3 months. In terms of resource use, participants in the PT group returned significantly less often to the ED. At baseline and 1 month, less prescription medication was used, including opioids, but there were no differences at 3 months. Although over-the-counter medication was recommended more at baseline in the PT group, there were no differences in use at 1 month, and the PT group had used them less at 3 months. There were no differences between groups at follow-up for imaging tests, other professionals consulted, and hospitalization rates. CONCLUSION: Patients presenting with a MSKD to the ED with direct access to a PT had better clinical outcomes and used less services and resources than those in the usual care group after ED discharge and up to 3 months after discharge.

Body mass index in urban Canada: neighborhood and metropolitan area effects.

OBJECTIVES: We investigated the influence of neighborhood and metropolitan area characteristics on body mass index (BMI) in urban Canada in 2001. METHODS: We conducted a multilevel analysis with data collected from a cross-sectional survey of men and women nested in neighborhoods and metropolitan areas in urban Canada during 2001. RESULTS: After we controlled for individual sociodemographic characteristics and behaviors, the average BMIs of residents of neighborhoods in which a large proportion of individuals had less than a high school education were higher than those BMIs of residents in neighborhoods with small proportions of such individuals (P< .01). Living in a neighborhood with a high proportion of recent immigrants was associated with lower BMI for men (P<.01), but not for women. Neighborhood dwelling density was not associated with BMI for either gender. Metropolitan sprawl was associated with higher BMI for men (P=.02), but the effect was not significant for women (P= .09). CONCLUSIONS: BMI is strongly patterned by an individual's social position in urban Canada. A neighborhood's social condition has an incremental influence on the average BMI of its residents. However, BMI is not influenced by dwelling density. Metropolitan sprawl is associated with higher BMI for Canadian men, which supports recent evidence of this same association among American men. Individuals and their environments collectively influence BMI in urban Canada.

Impact of the Ottawa Decision Support Framework on the agreement and the difference between patients' and physicians' decisional conflict.

BACKGROUND: The Ottawa Decision Support Framework (ODSF) provides a process that facilitates shared decision making. OBJECTIVE: To assess the impact of implementing the ODSF on the agreement and the difference between patients' and physicians' decisional conflict scores. DESIGN: In total, 120 physicians and 903 patients enrolled in this before-and-after study. Implementation of the ODSF was composed of an interactive workshop, feedback, and a reminder at the point of care. The Decisional Conflict Scale (DCS) was completed by physicians and patients after a clinical encounter. RESULTS: The intraclass correlation coefficient was-0.205 +/- 0.096 (95% confidence interval [CI]= - 0.224 to -0.186) before implementing the ODSF and- 0.013 +/- 0.114 (95% CI = - 0.036 to 0.009) after. At the patient level, the following factors were significantly associated with the difference between the patients' and physicians' DCS: unemployed (P = 0.023), implementing the ODSF (P = 0.008), high school degree (P = 0.04), male (P = 0.03), and unilateral role in decision making (P = 0.03). At the physician level, provincial committee (P = 0.001), national committee (P = 0.045), clinical site (P = 0.016), reluctance to share uncertainty (P = 0.023), and anxiety due to uncertainty (P = 0.017) were significantly associated with this outcome. CONCLUSION: After implementing the ODSF, there was less dissimilarity between patients' and physicians' DSC than expected by chance than before. Implementing the ODSF was also found to be associated with the difference between patients' and physicians' DSC. The physician level explained a significant amount of the variance in this outcome, thus emphasizing the importance of an intervention at this level.

Effects of physical activity on pediatric reference data for obesity.

PURPOSE: To examine the influence of physical activity on pediatric obesity reference data for Canada. METHODS: The sample included 3527 boys and 3554 girls, 6 to 18 years of age, from the 2004 Canadian Community Health Survey: Nutrition component. The heights and weights of the participants were directly measured, and the body mass index was calculated (BMI: kg/m(2)). Physical activity levels were reported using an interviewer-administered questionnaire. Participants were divided into low and high physical activity groups, based on age-specific physical activity levels (lower and upper quartiles). BMI percentiles (25th, 50th, 75th, 85th, 95th) were generated by sex using the LMS method, separately by physical activity groups. RESULTS: There were only minor differences in BMI at the 25th and 50th percentiles between physical activity groups in both boys and girls. However, in boys, the low active group had somewhat higher BMI values at the 85th and 95th percentiles than the high active group after the age of 10 years. In girls, the differences in BMI across groups was similar to that of boys at the 95th percentile, but inconsistent at the other percentiles. CONCLUSION: The results suggest that screening for physical activity may be important for the development of national reference data for obesity.

Waiting for care in Canada: findings from the health services access survey.

Waiting for care has been and continues to be a major issue for the healthcare sector in Canada. While considerable gains have been made regarding valid and reliable information on waiting times, gaps remain. Statistics Canada continues to provide information regarding patients' experiences in accessing care at the national and provincial levels, including how long individuals waited for specialized services, through the Health Services Access Survey. The survey offers several advantages, including waiting time information that is comparable across time and space, enhanced patient information and information regarding patients' experiences in waiting for care. The results for 2005 indicate that median waiting time for all specialized services was between 3 and 4 weeks and remained relatively stable between 2003 and 2005. Waiting times for specialist visits did not vary by income. In addition to being asked how long they waited, individuals were asked about their experiences in waiting for care. While the majority of individuals waiting for care indicated that their waiting time was acceptable, there continues to be a proportion of Canadians who feel they are waiting an unacceptably long time for care. Between 11% and 18% of individuals waiting for care indicated that their life was affected by waiting.

Factors related to adolescents' self-perceived health.

OBJECTIVES: This analysis examines self-perceived health among Canadian adolescents aged 12 to 17, and factors associated with ratings of very good/excellent health. DATA SOURCE: The data are from cycle 1.1 of the 2000/01 Canadian Community Health Survey (CCHS), conducted by Statistics Canada. The sample consisted of 12,715 adolescents aged 12 to 17. ANALYTICAL TECHNIQUES: Cross-tabulations were used to estimate the prevalence of various characteristics and health behaviours for the 12-to-14 and 15-to-17 age groups. Multiple logistic regression was used to model associations between very good/excellent self-reported health and selected characteristics. MAIN RESULTS: In 2000/01, nearly 30% of 12- to 17-year-olds rated their health as poor, fair or good. At ages 15 to 17, girls were less likely than boys to report very good/excellent health and were more likely to have a chronic condition and to have experienced depression in the past year. When other factors were taken into account, the odds of reporting very good/excellent health were significantly lower for teens who were daily smokers, episodic heavy drinkers, physically inactive during leisure time, infrequent consumers of fruit and vegetables, or obese, compared with teens who did not have these characteristics.

Changes in unmet health care needs.

OBJECTIVES: This article examines recent trends in self-reported unmet health care needs among the household population aged 12 or older, and explores various explanations for the increase observed. DATA SOURCES: The data are from the first half (September 2000 through February 2001) of data collection for cycle 1.1 of the Canadian Community Health Survey and from cross-sectional (1994/95 through 1998/99) household components of the National Population Health Survey. ANALYTICAL TECHNIQUES: Weighted frequencies and cross-tabulations were used to estimate the proportion of people aged 12 or older who reported that they did not receive health care when they thought they needed it. Estimates were also produced for the type of care sought, and specific reasons for unmet health care needs. MAIN RESULTS: The percentage of people reporting unmet health care needs rose gradually between 1994/95 and 1998/99, then doubled (from 6% to over 12%) between 1998/99 and 2000/01. Long waiting time was the reason most frequently reported for unmet needs.

Factors associated with the difference in score between women's and doctors' decisional conflict about hormone therapy: a multilevel regression analysis.

OBJECTIVE: To explore factors associated with the difference in score between women's and doctors' decisional conflict about hormone therapy (HT). DESIGN: Secondary analysis. SETTING AND PARTICIPANTS: family doctors were randomized to prepare women for counselling about HT using either a decision aid or a pamphlet. MAIN VARIABLES STUDIED: After each counselling session, decisional conflict was assessed in women and doctors using the Decisional Conflict Scale (DCS) and the Provider Decision Process Assessment Instrument (PDPAI), respectively. The difference in score between the DCS and PDPAI was computed and entered as the dependent variable in a multilevel regression analysis. MAIN OUTCOME RESULTS: A total of 40 doctors and 167 women were included in the analysis. The intra-doctor correlation coefficient was 0.25. Factors associated with women experiencing higher decisional conflict than their doctor were: age of doctor >45 years, women who were undecided about the best choice after the counselling session, women with a university degree and women who said that their doctor usually does not give them control over treatment decision. Factors associated with doctors experiencing more decisional conflict than women were: doctors who were undecided about the quality of the decision, length of visit <30 min and women who thought that the decision was shared with their doctor. CONCLUSION: In order to reduce the disparities between women's and doctors' decisional conflict about HT, interventions aimed at raising awareness of doctors about shared decision-making should be encouraged.