Explore the vast range of titles available.

We need to consider the ethical challenges inherent in implementing machine learning in health care if its benefits are to be realized. Some of these challenges are straightforward, whereas others have less obvious risks but raise broader ethical concerns.

Proton pump inhibitors (PPIs) have been associated with adverse clinical outcomes amongst clopidogrel users after an acute coronary syndrome. Recent pre-clinical results suggest that this risk might extend to subjects without any prior history of cardiovascular disease. We explore this potential risk in the general population via data-mining approaches. Using a novel approach for mining clinical data for pharmacovigilance, we queried over 16 million clinical documents on 2.9 million individuals to examine whether PPI usage was associated with cardiovascular risk in the general population. In multiple data sources, we found gastroesophageal reflux disease (GERD) patients exposed to PPIs to have a 1.16 fold increased association (95% CI 1.09-1.24) with myocardial infarction (MI). Survival analysis in a prospective cohort found a two-fold (HR = 2.00; 95% CI 1.07-3.78; P = 0.031) increase in association with cardiovascular mortality. We found that this association exists regardless of clopidogrel use. We also found that H2 blockers, an alternate treatment for GERD, were not associated with increased cardiovascular risk; had they been in place, such pharmacovigilance algorithms could have flagged this risk as early as the year 2000. Consistent with our pre-clinical findings that PPIs may adversely impact vascular function, our data-mining study supports the association of PPI exposure with risk for MI in the general population. These data provide an example of how a combination of experimental studies and data-mining approaches can be applied to prioritize drug safety signals for further investigation.

Background Access to palliative care is a key quality metric which most healthcare organizations strive to improve. The primary challenges to increasing palliative care access are a combination of physicians over-estimating patient prognoses, and a shortage of palliative staff in general. This, in combination with treatment inertia can result in a mismatch between patient wishes, and their actual care towards the end of life. Methods In this work, we address this problem, with Institutional Review Board approval, using machine learning and Electronic Health Record (EHR) data of patients. We train a Deep Neural Network model on the EHR data of patients from previous years, to predict mortality of patients within the next 3-12 month period. This prediction is used as a proxy decision for identifying patients who could benefit from palliative care. Results The EHR data of all admitted patients are evaluated every night by this algorithm, and the palliative care team is automatically notified of the list of patients with a positive prediction. In addition, we present a novel technique for decision interpretation, using which we provide explanations for the model’s predictions. Conclusion The automatic screening and notification saves the palliative care team the burden of time consuming chart reviews of all patients, and allows them to take a proactive approach in reaching out to such patients rather then relying on referrals from the treating physicians.

In the global South, rural and resource-based livelihoods increasingly face water-related risks. The conceptualization and application of the water security concept in relation to rural livelihoods has not been reviewed in this context. To fill this gap, a systematic scoping review of refereed journal articles (2000–2019) was conducted to examine how water security is defined, driven, and addressed for rural livelihoods in the global South. Publications (n = 99) featured diverse methodologies and geographical contexts, and recognized simultaneous drivers of water insecurity and solution strategies for water security. Several shortcomings were evident. First, only 30.3% of publications defined the concept, mostly using frames of ‘adequate’, ‘sufficient’, and ‘acceptable’ water-related risks. Few definitions recognized the role of water security interventions in increasing capabilities and prosperity. Second, technical and managerial responses to proximate drivers of water-related risk – namely climate-related dynamics, water re-allocation, extraction, and mismanagement – outnumbered efforts to identify and transform the underlying social, economic, and political inequities that create and sustain water insecurity. Last, studies focused heavily on agriculture, while labour, transhumance pastoralism, and aquaculture were underrepresented. A research agenda that increases the synergies between the wider water security and rural livelihoods scholarship is advanced to address these shortcomings.

Data of abortion incidence and trends are needed to monitor progress toward improvement of maternal health and access to family planning. To date, estimates of safe and unsafe abortion worldwide have only been made for 1995 and 2003. We used the standard WHO definition of unsafe abortions. Safe abortion estimates were based largely on official statistics and nationally representative surveys. Unsafe abortion estimates were based primarily on information from published studies, hospital records, and surveys of women. We used additional sources and systematic approaches to make corrections and projections as needed where data were misreported, incomplete, or from earlier years. We assessed trends in abortion incidence using rates developed for 1995, 2003, and 2008 with the same methodology. We used linear regression models to explore the association of the legal status of abortion with the abortion rate across subregions of the world in 2008. The global abortion rate was stable between 2003 and 2008, with rates of 29 and 28 abortions per 1000 women aged 15–44 years, respectively, following a period of decline from 35 abortions per 1000 women in 1995. The average annual percent change in the rate was nearly 2·4% between 1995 and 2003 and 0·3% between 2003 and 2008. Worldwide, 49% of abortions were unsafe in 2008, compared to 44% in 1995. About one in five pregnancies ended in abortion in 2008. The abortion rate was lower in subregions where more women live under liberal abortion laws (p<0·05). The substantial decline in the abortion rate observed earlier has stalled, and the proportion of all abortions that are unsafe has increased. Restrictive abortion laws are not associated with lower abortion rates. Measures to reduce the incidence of unintended pregnancy and unsafe abortion, including investments in family planning services and safe abortion care, are crucial steps toward achieving the Millennium Development Goals. UK Department for International Development, Dutch Ministry of Foreign Affairs, and John D and Catherine T MacArthur Foundation.

Monocyte homing to the liver and adhesion to the liver sinusoidal endothelial cells (LSECs) are key elements in nonalcoholic steatohepatitis (NASH) pathogenesis. We reported previously that VCAM-1 mediates monocyte adhesion to LSECs. However, the pathogenic role of VCAM-1 in NASH is unclear. Herein, we report that VCAM-1 was a top upregulated adhesion molecule in the NASH mouse liver transcriptome. Open chromatin landscape profiling combined with genome-wide transcriptome analysis showed robust transcriptional upregulation of LSEC VCAM-1 in murine NASH. Moreover, LSEC VCAM-1 expression was significantly increased in human NASH. LSEC VCAM-1 expression was upregulated by palmitate treatment in vitro and reduced with inhibition of the mitogen-activated protein 3 kinase (MAP3K) mixed lineage kinase 3 (MLK3). Likewise, LSEC VCAM-1 expression was reduced in the Mlk3-/- mice with diet-induced NASH. Furthermore, VCAM-1 neutralizing Ab or pharmacological inhibition attenuated diet-induced NASH in mice, mainly via reducing the proinflammatory monocyte hepatic population as examined by mass cytometry by time of flight (CyTOF). Moreover, endothelium-specific Vcam1 knockout mice were also protected against NASH. In summary, lipotoxic stress enhances the expression of LSEC VCAM-1, in part, through MLK3 signaling. Inhibition of VCAM-1 was salutary in murine NASH and might serve as a potential therapeutic strategy for human NASH.

The success of foundation models such as ChatGPT and AlphaFold has spurred significant interest in building similar models for electronic medical records (EMRs) to improve patient care and hospital operations. However, recent hype has obscured critical gaps in our understanding of these models’ capabilities. In this narrative review, we examine 84 foundation models trained on non-imaging EMR data (i.e., clinical text and/or structured data) and create a taxonomy delineating their architectures, training data, and potential use cases. We find that most models are trained on small, narrowly-scoped clinical datasets (e.g., MIMIC-III) or broad, public biomedical corpora (e.g., PubMed) and are evaluated on tasks that do not provide meaningful insights on their usefulness to health systems. Considering these findings, we propose an improved evaluation framework for measuring the benefits of clinical foundation models that is more closely grounded to metrics that matter in healthcare.

As challenging as it is to use face recognition with a Single Sample Per Person, it becomes even more difficult when face recognition based on a single sample is performed in an unconstrained environment. The unconstrained environment is normally considered irregular in facial expressions, pose, occlusion, and illumination. This degree of difficulty increases as a result of the single sample and in the presence of occlusion. Extensive research has been done on face recognition under pose and expression changes. Comparatively, less research has been reported on the occlusion problem that occurs in facial images. Occlusion may alter the appearance of facial images and cause deterioration in recognition. A robust method is required to handle the occlusion in the face image to improve the recognition performance. This study aimed to implement an effective augmentation technique that improves the performance of the Single Sample Per Person face recognition system in unconstrained environments. Virtual samples were created to expand the sample size to address the problem of a single sample. A local region-based technique was proposed to deal with occlusion by creating virtual samples. A deep neural network-based model, FaceNet, was used to extract the features and a support vector machine was used for classification. The performance of the proposed approach was evaluated, demonstrating its superiority in handling occlusion compared to that of its state-of-the-art counterparts. The proposed method achieved significant accuracy improvements, specifically 94.83% for the occlusion with sunglasses and 98% for the occlusion with scarves in the AR dataset.

Poorly differentiated thyroid cancer (PDTC) and anaplastic thyroid cancer (ATC) are rare and frequently lethal tumors that so far have not been subjected to comprehensive genetic characterization. We performed next-generation sequencing of 341 cancer genes from 117 patient-derived PDTCs and ATCs and analyzed the transcriptome of a representative subset of 37 tumors. Results were analyzed in the context of The Cancer Genome Atlas study (TCGA study) of papillary thyroid cancers (PTC). Compared to PDTCs, ATCs had a greater mutation burden, including a higher frequency of mutations in TP53, TERT promoter, PI3K/AKT/mTOR pathway effectors, SWI/SNF subunits, and histone methyltransferases. BRAF and RAS were the predominant drivers and dictated distinct tropism for nodal versus distant metastases in PDTC. RAS and BRAF sharply distinguished between PDTCs defined by the Turin (PDTC-Turin) versus MSKCC (PDTC-MSK) criteria, respectively. Mutations of EIF1AX, a component of the translational preinitiation complex, were markedly enriched in PDTCs and ATCs and had a striking pattern of co-occurrence with RAS mutations. While TERT promoter mutations were rare and subclonal in PTCs, they were clonal and highly prevalent in advanced cancers. Application of the TCGA-derived BRAF-RAS score (a measure of MAPK transcriptional output) revealed a preserved relationship with BRAF/RAS mutation in PDTCs, whereas ATCs were BRAF-like irrespective of driver mutation. These data support a model of tumorigenesis whereby PDTCs and ATCs arise from well-differentiated tumors through the accumulation of key additional genetic abnormalities, many of which have prognostic and possible therapeutic relevance. The widespread genomic disruptions in ATC compared with PDTC underscore their greater virulence and higher mortality. This work was supported in part by NIH grants CA50706, CA72597, P50-CA72012, P30-CA008748, and 5T32-CA160001; the Lefkovsky Family Foundation; the Society of Memorial Sloan Kettering; the Byrne fund; and Cycle for Survival.