Search

Background: Dementia is a significant cause of death in the older population and is becoming an important public health issue as the population ages and the prevalence of dementia increases. The Braden score is one of the most commonly used clinical tools to assess the risk of skin pressure injury in patients, and some studies have reported that it may reflect the state of frailty of patients. The present study attempted to explore the association between Braden score and 90‐day mortality, pressure injury, and aspiration pneumonia in older patients with dementia in the intensive care unit (ICU). Methods: The study involved extracting crucial data from the Medical Information Market for Intensive Care IV (MIMIC‐IV) database using Structured Query Language, with a license certificate obtained after completing the necessary training and examination available on the MIMIC‐IV website. A retrospective analysis was performed on older patients with dementia, aged 65 or older, who were first admitted to the ICU. Ninth and tenth revision International Classification of Diseases codes were used to identify patients with dementia. The primary outcome was 90‐day mortality. Cox proportional hazards models were used to determine the association between Braden score and death, and hazard ratios (HR) and 95% confidence intervals (CI) were calculated. Propensity score matching and E‐value assessments were employed for sensitivity analysis. Results: A total of 2892 patients with a median age of approximately 85 years (interquartile range 78.74–89.59) were included, of whom 1625 were female (56.2%). Patients had a median Braden score of 14 (interquartile range 12–15) at ICU admission. Braden score at ICU admission was inversely associated with 90‐day mortality risk after adjustment for demographics, severity of illness, treatment and medications, delirium, and sepsis (adjusted HR: 0.92, 95% CI: 0.87–0.98, p = 0.006). Patients were divided into two groups with a cut‐off value of 15: high‐risk group and low‐risk group. Compared to the low‐risk group (Braden score >15), the risk of 90‐day mortality was significantly increased in the high‐risk group (Braden score ≤15) (adjusted HR: 1.52, 95% CI: 1.10–2.09, p = 0.011, E‐value: 2.01), the risk of pressure injury (adjusted OR: 2.62, 95% CI: 2.02–3.43, E‐value: 2.62) and aspiration pneumonia (adjusted OR: 2.55, 95% CI: 1.84–3.61, E‐value: 2.57) was also significantly higher. Conclusions: The Braden score may be a quick and simple screening tool to identify the risk of adverse outcomes in critically ill older adults with dementia. Key points: The Braden score, which is commonly used to assess the risk of pressure injury, may reflect frailty in patients.The Braden score at ICU admission was inversely associated with 90‐day mortality risk, suggesting its potential as a mortality predictor.The findings suggest that the Braden score could be a valuable tool for quickly identifying critically ill older adults with dementia who are at risk of adverse outcomes. [ABSTRACT FROM AUTHOR]

This systematic review examines prediction models for adverse drug events (ADEs) related to antineoplastic drugs, which are used in cancer therapy. The review analyzes 55 studies and evaluates the quality and reporting of these prediction models. The majority of models focused on ADEs related to conventional chemotherapy drugs, with fewer models addressing novel anticancer drugs and additional toxicities. The study finds that machine learning algorithms generally had better methodological quality but weaker reporting quality compared to traditional statistical models. The authors recommend expanding research focus, optimizing model presentation, and improving the design, conduct, validation, and reporting of prediction models to enhance their application in clinical practice. [Extracted from the article]

One challenge in power-system control designs is the gap between numerical model-based analysis and complex real-world power systems. With increased data and measurements being collected from power systems, data-driven analysis (e.g., machine learning) may provide an alternative approach to reveal hidden information through learning from the real system data, and provide insights for better control scheme design during the utility planning process. In this study, data-driven feature analysis is performed to evaluate the relationships between series compensation, power generation, and path flows in a real transmission system, as well as temporal patterns. The main data-driven analysis methods, including statistical cross correlation, multinomial logistical regression, and classification and regression trees, are integrated for feature selection and developing predictive models of series compensation. Analysis results demonstrates the effectiveness of the proposed methodology in feature analysis and the potential to help improve power-system control scheme design. [ABSTRACT FROM AUTHOR]