Plos One
-
Observational Study
Combining patient visual timelines with deep learning to predict mortality.
Deep learning algorithms have achieved human-equivalent performance in image recognition. However, the majority of clinical data within electronic health records is inherently in a non-image format. Therefore, creating visual representations of clinical data could facilitate using cutting-edge deep learning models for predicting outcomes such as in-hospital mortality, while enabling clinician interpretability. The objective of this study was to develop a framework that first transforms longitudinal patient data into visual timelines and then utilizes deep learning to predict in-hospital mortality. ⋯ We converted longitudinal patient data into visual timelines and applied a deep neural network for predicting in-hospital mortality more accurately than current standard clinical models, while allowing for interpretation. Our framework holds promise for predicting several important outcomes in clinical medicine.
-
Among women, breast cancer is a leading cause of death. Breast cancer risk predictions can inform screening and preventative actions. Previous works found that adding inputs to the widely-used Gail model improved its ability to predict breast cancer risk. ⋯ However, the logistic regression, linear discriminant analysis, and neural network models with the broader set of inputs were all significantly stronger than the BCRAT. These results suggest that relative to the BCRAT, additional easy-to-obtain personal health inputs can improve five-year breast cancer risk prediction. Our models could be used as non-invasive and cost-effective risk stratification tools to increase early breast cancer detection and prevention, motivating both immediate actions like screening and long-term preventative measures such as hormone replacement therapy and chemoprevention.
-
Clinical Trial
How the service delivery works in the Iranian specialised burns hospitals? A qualitative approach.
As burn injuries are a major cause of death and infirmity, successful service delivery is vital in health systems. In Iran, a few specialised burns hospitals (SBHs) located in big provinces provide burn services in which burn patients with more severe conditions are referred to. However, SBHs are faced with several challenges for delivering due treatment for burn patients. ⋯ Themes and (subthemes) including burn care continuum (preventive care, pre-hospital care, hospital care, follow-up, and home care), regionalisation of burning services (access to other specialties and medical services, access to specialized care in provinces without a SBH, standardised regionalisation system for burn related services (BRSs), costs of providing BRSs (expensive services and supplies and long hospitalisation), and non-compliance with standardised care (guidelines to provide burn care and physical space to provide BRSs). Results suggest that improving BRSs delivery in Iran may be reached by strengthening burn care continuum, regionalising burn care, allocating sufficient budgets to burn services and formulating burn care guidelines. These policy actions can be better addressed via intra-sectoral collaborations.
-
Methicillin-resistant Staphylococcus aureus (MRSA) infection is an important clinical concern in patients, and is often associated with significant disease burden and metastatic infections. There is an increasing evidence of heterogeneous vancomycin-intermediate S. aureus (hVISA) associated treatment failure. In this study, we aim to understand the molecular mechanism of teicoplanin resistant MRSA (TR-MRSA) and hVISA. ⋯ However, none of the markers were reliable in differentiating hVISA from TR-MRSA. Significant pbp2 upregulation was noted in three TR-MRSA strains, which had teicoplanin MICs of 16 or 32 μg/ml, whilst significant pbp4 downregulation was not noted in these strains. In our study, multiple mutations were identified in the candidate genes, suggesting a complex evolutionary pathway involved in the development of TR-MRSA and hVISA strains.
-
In two studies we provide a novel investigation into the effects of monetary switching costs on choice-inertia (i.e., selection of the same option on consecutive choices). Study 1 employed a static decisions-from-feedback task and found that the introduction of, as well as larger, monetary switching costs led to increases in choice-inertia. ⋯ The effect of switching costs increasing choice-inertia for both the EV maximizing and the inferior option was replicated with little impact of the change in options values being detected. In sum, decision makers appear to be sensitive to switching costs, and this sensitivity can bias them towards inferior or superior options, revealing the good and the bad of choice-inertia.