Related Articles
Socially vulnerable communities face disproportionate exposure and susceptibility to U.S. wildfire and prescribed burn smoke
While air pollution from most U.S. sources has decreased, emissions from wildland fires have risen. Here, we use an integrated assessment model to estimate that wildfire and prescribed burn smoke caused $200 billion in health damages in 2017, associated with 20,000 premature deaths. Nearly half of this damage came from wildfires, predominantly in the West, with the remainder from prescribed burns, mostly in the Southeast. Our analysis reveals positive correlations between smoke exposure and various social vulnerability measures; however, when also considering smoke susceptibility, these disparities are systematically influenced by age. Senior citizens, who are disproportionately White, represented 16% of the population but incurred 75% of the damages. Nonetheless, within most age groups, Native American and Black communities experienced the greatest damages per capita. Our work highlights the extraordinary and disproportionate effects of the growing threat of fire smoke and calls for targeted, equitable policy solutions for a healthier future.
Addressing myocardial infarction in South-Asian populations: risk factors and machine learning approaches
Cardiovascular diseases, especially myocardial infarction (MI), are an important and up-trending public health challenge in the South Asian population. With urbanization and economic development, there has been a rise in obesity, dyslipidemia, diabetes mellitus, and hypertension in these regions, which, combined with genetic predisposition, create a unique cardiovascular risk profile among South Asians. Traditional risk assessment tools often underestimate the cardiovascular risk in South Asians due to a lack of phenotypic representation in their development. In this review, we explore the risk factors for MI in South Asians and highlight the potential role of machine learning (ML) and deep learning (DL) in enhancing diagnostic and predictive accuracy. These ML algorithms, including convolutional neural networks (CNNs) and transformer-based models, show potential in analyzing complex information from clinical characteristics, electrocardiograms (ECG), and cardiac biomarkers while integrating multimodal data. We also explore the challenges in accessing high-quality datasets and enabling applicability in clinical settings. We believe that future research should focus on developing comprehensive cardiovascular risk scores that incorporate South Asian-specific risk factors and leverage advanced ML models to enhance risk prediction, diagnosis, and management.
Expert perspectives on incorporating justice considerations into integrated assessment modelling
There is growing criticism aimed towards global integrated assessment models (IAMs) and an ongoing academic debate on how justice considerations can be incorporated in those models. By relying on 39 interviews with a multidisciplinary group of experts, we map three shapes of change containing multiple avenues for incorporating justice considerations into IAM tools and scenarios: to improve representation within IAMs (Shape 1), to couple to new models and expand points of access to disciplines and users (Shape 2), and to refine the role of IAMs within a wider array of practices (Shape 3). These shapes reflect multi-disciplinary agreements and divergences over the capacity of IAMs to incorporate justice considerations—regarding kinds of representation, greater involvement of new disciplines and users, and the objective of mitigation scenarios in climate policy. Our analysis is among the first to describe and integrate a variety of opinions from different communities, fostering a more holistic understanding of the opportunities and challenges of incorporating justice into IAMs.
Multi-expert ensemble ECG diagnostic algorithm using mutually exclusive–symbiotic correlation between 254 hierarchical multiple labels
Electrocardiograms (ECGs) are a cheap and convenient means of assessing heart health and provide an important basis for diagnosis and treatment by cardiologists. However, existing intelligent ECG diagnostic approaches can only detect up to several tens of ECG terms, which barely cover the most common arrhythmias. Thus, further diagnosis is required by cardiologists in clinical settings. This paper describes the development of a multi-expert ensemble learning model that can recognize 254 ECG terms. Based on data from 191,804 wearable 12-lead ECGs, mutually exclusive–symbiotic correlations between hierarchical multiple labels are applied at the loss level to improve the diagnostic performance of the model and make its predictions more reasonable while alleviating the difficulty of class imbalance. The model achieves an average area under the receiver operating characteristics curve of 0.973 and 0.956 on offline and online test sets, respectively. We select 130 terms from the 254 available for clinical settings by considering the classification performance and clinical significance, providing real-time and comprehensive ancillary support for the public.
Toward change in the uneven geographies of urban knowledge production
More than four-fifths of the global urban population live in the Global South and East. Most urban theories, however, originate in the Global North. Building on recent efforts to address this mismatch, this paper examines the geographies of urban knowledge production. It analyzes the institutional affiliations of contributions in 25 leading Anglophone journals (n = 14,582) and nine urban handbooks (n = 252). We show that 42% of the journal articles and 17% of the handbook chapters were authored outside the Global North. However, only 15% of the editor positions (handbooks: 10%) were held by scholars based outside the Global North. This indicates that Global Northern institutions still dominate knowledge gatekeeping, whereas authors are more diverse. Additionally, more empirical journals and those with fewer Northern board members tend to publish more non-Northern authors. Our findings underscore the need for greater epistemic diversity in gatekeeping positions and broader understandings of what counts as theory to better incorporate diverse urban knowledge.
Responses