Search

The optimization of locomotive drivers' scheduling in rail freight transportation comes as a necessity for minimizing economic expenses and training investments. The Ferodata AI engine, an artificial intelligence (AI)/machine learning (ML) software module, developed by our team, has integrated a supervised random forest model that automatically assigns conductors to freight transportation orders based on the data about locomotive driver's tiredness score, distance of the driver to the departure point of a transportation order, driver availability, and circulation history. The model proposed by us obtained very good performance metrics on the train set (accuracy: 95%, AUC: 0.9905) and reasonably good and encouraging performance on the test set (accuracy: 84%, AUC: 0.8357). After rigorous testing and validation on external and larger datasets, the automated optimization of locomotive driver assignments could bring operational efficiency, cost savings, regulatory compliance, and improved safety to scheduled rail freight transports. [ABSTRACT FROM AUTHOR]

Our paper proposes the first machine learning model to predict long-term mortality in patients with diabetic foot ulcers (DFUs). The study includes 635 patients with DFUs admitted from January 2007 to December 2017, with a follow-up period extending until December 2020. Two multilayer perceptron (MLP) classifiers were developed. The first MLP model was developed to predict whether the patient will die in the next 5 years after the current hospitalization. The second MLP classifier was built to estimate whether the patient will die in the following 10 years. The 5-year and 10-year mortality models were based on the following predictors: age; the University of Texas Staging System for Diabetic Foot Ulcers score; the Wagner–Meggitt classification; the Saint Elian Wound Score System; glomerular filtration rate; topographic aspects and the depth of the lesion; and the presence of foot ischemia, cardiovascular disease, diabetic nephropathy, and hypertension. The accuracy for the 5-year and 10-year models was 0.7717 and 0.7598, respectively (for the training set) and 0.7244 and 0.7087, respectively (for the test set). Our findings indicate that it is possible to predict with good accuracy the risk of death in patients with DFUs using non-invasive and low-cost predictors. [ABSTRACT FROM AUTHOR]

Left ventricular (LV) aneurysm following acute myocardial infarction (MI) represents a less common complication, but with worse clinical outcomes. Ventricular surgical reconstruction is not always the intervention of choice due to high surgical risk. There were proposed less invasive LV aneurysm exclusion techniques such as the less invasive ventricular enhancement (LIVE) procedure. Our paper represents the first systematic approach to investigate the efficacy and safety of LIVE procedure using Revivent TCTM anchor system for LV aneurysm exclusion. Studies were considered if they reported original data regarding LIVE procedure's efficacy and/or safety using the Revivent TCTM system in patients with LV aneurysms. Five studies met the inclusion criteria. The procedure is associated with a reduction in LV volumes and an improvement in LV ejection fraction (LVEF). The means of preoperative LVEF varied between 22.8% and 35.6%, while postoperative LVEF means ranged between 34% and 45.9% (P < 0.005) across studies. All included papers reported a significant difference between preoperative and postoperative LV end-systolic volume index (P ≤ 0.001) and LV end-diastolic volume index (P ≤ 0.001). Three out of four studies achieved statistical significance (P ≤ 0.001) when comparing preoperative (means range: 2.6-3.4) and postoperative (means range: 1.4-1.9) New York Heart Association (NYHA) class. One study reported a survival rate of 90.6 (95% CI, 84.6-97.0) at 12 months following the procedure. LIVE appears to be a promising and appropriate treatment strategy for a complex condition, which could extend the indication of LV aneurysm exclusion in the future. [ABSTRACT FROM AUTHOR]

Acute kidney injury following cardiac surgery (CS-AKI) represents a severe postoperative complication, negatively impacting short-term and long-term mortality. Due to the lack of a specific treatment, effective prevention remains the most powerful tool to overcome the CS-AKI burden. Improving the preventive strategies is possible by establishing appropriate preoperative risk profiles. Various clinical models were proposed as a means to assist physicians in stratifying the risk of CS-AKI. However, these models are used for predicting severe forms of CS-AKI, while their predictive power for mild forms is insufficient. Our paper represents the first systematic approach to review all proposed preoperative risk factors and their predictive power. Our strategy is the starting point for selecting and comparing the predictive elements to be integrated into future risk models. Heart failure, chronic hyperglycemia, anemia, obesity, preoperative exposure to nephrotoxic drugs or contrast media, inflammation, proteinuria, and pre-existing kidney disease were systematically reviewed and were found to be associated with an increased risk of postoperative CS-AKI. As no externally validated and universally accepted risk models currently exist, the clinical judgment and a good knowledge of the preoperative risk factors in the light of new evidence may help personalize preoperative risk profiles as the cornerstone of prevention measures. [ABSTRACT FROM AUTHOR]

The issue of the COVID-19 pandemic occupies the agenda of the whole world. The pivot of this pandemic is a crucial element that has become almost as important as the virus itself, namely the lockdown. Although, the rationale for lockdown is well-sustained by strong epidemiological arguments, exploring the 'other' unwanted consequences of the contemporary COVID-19 pandemic is mandatory for coagulating a robust agreed position against the numerous problems generated by the SARS-CoV-2 virus. Starting from the rationale of the lockdown, in this paper we explored and exposed the other consequences of the COVID-19 pandemic measures such as the use or abuse of human rights and freedom restrictions, economic issues, marginalized groups and eclipse of all other diseases. Our scientific attempt is to coagulate a stable position and integrate current opposing views by advancing the idea that rather than applying the uniform lockdown policy, one could recommend instead an improved model targeting more strict and more prolonged lockdowns to vulnerable risk/age groups while enabling less stringent measures for the lower-risk groups, minimizing both economic losses and deaths. Rigorous (and also governed by freedom) debating may be able to synchronize the opposed perspectives between those advocating an extreme lockdown (e.g., most of the epidemiologists and health experts), and those criticizing all restrictive measures (e.g., economists and human rights experts). Confronting the multiple facets of the public health mitigation measures is the only way to avoid contributing to history with yet another failure, as seen in other past epidemics. [ABSTRACT FROM AUTHOR]