We present an innovative extension of the Set Covering Problem, transitioning from a traditional radial covering to an angular covering structure. The decisions are based on locating the facilities and identifying the directional servers installed in each, covering a set of points in a geographic area. The strategic placement of surveillance cameras inspired this novel approach. We aim to minimize the cost of covering demand points by strategically installing directional servers on facilities. We propose an integer linear model and an initial approach based on column generation decomposition. We conducted extensive computational experiments on different test instances, including a practical case study involving positioning security cameras for surveillance in Valparaíso, Chile. The results highlight the effectiveness of column generation in achieving optimal solutions or significantly improving solution quality compared with solving the model directly with standard optimization solvers, especially for larger instances. In particular, the column generation algorithm achieves up to a 67% improvement in solution quality compared to the optimization model for real-world size instances, demonstrating its practical applicability and potential for enhancing surveillance infrastructure design.

Background: Pressure injury (PI) prevention relies heavily on the application of the Braden Scale, yet its predictive performance is limited mainly due to subjective measurements and fixed cut-off points, yielding potential errors in terms of false-positive alarm rates and workforce misuse. Machine learning (ML) has emerged as a promising approach to improve discrimination, but fair comparisons with traditional tools remain scarce in clinical settings. Methods: Using data from 446 hospitalized patients in a tertiary hospital in Chile, we compared five classic ML models for classification against the Braden Scale. ML models were trained exclusively on routinely collected clinical and nursing variables, excluding Braden inputs. A matched operating-point framework was applied, aligning ML decision thresholds with Braden cutoffs based on equivalent recall or specificity. Results: The XGB model showed the highest overall discrimination performance (AUC = 0.835). When matched to Braden recall, XGB achieved about 17% gains in specificity, substantially reducing false positives. When matched on specificity, recall improvements ranged from 13% to 25%. These gains were consistent across clinically relevant thresholds considered during the comparison. Conclusions: ML models, particularly XGB, outperform the Braden Scale under equivalent clinical operating conditions. Rather than replacing Braden, ML has emerged as a promising approach that preserves patient safety while improving precision and resource allocation.

Background: Early unplanned readmission to critical care units is associated with adverse outcomes and consumes scarce capacity twice, yet the decision to transfer a patient out of the unit is still made without objective criteria. Machine learning may support this decision, but metaheuristic hyperparameter search remains largely unexplored here. Methods: We analyzed 371 patients discharged from the critical care unit of a Chilean public hospital, of whom 41 were readmitted within seven days of transfer to a general ward. Five models (logistic regression, support vector classification, decision tree, random forest, and XGBoost) were trained on clinical and nursing variables available at transfer, using nested stratified five-fold cross-validation with preprocessing fitted within the training folds. Bayesian optimization and a genetic algorithm were compared under identical folds. Results: Under nested cross-validation, discrimination was moderate (AUC 53.3–75.2%), and no single model performed best across all metrics. Logistic regression achieved the highest AUC (75.2%) with the lowest variability across folds, followed by XGBoost (75.0%), whose results varied markedly across folds. The genetic algorithm yielded higher AUC than Bayesian optimization for most tree-based models. Days in the critical care unit, age, and unit of origin ranked highest in impurity-based importance. Conclusions: This proof-of-concept study suggests that routinely collected clinical and nursing information available at transfer may inform early readmission prediction. As the sample was case-enriched and derived from a single center, recalibration to the natural readmission prevalence and prospective external validation are required before clinical implementation.

Background: Pressure injuries (PIs) remain a major global healthcare challenge, increasing morbidity and costs. While Machine Learning (ML) models effectively predict PI risk, their lack of uncertainty quantification limits clinical trust. Conformal Prediction (CP) addresses this issue by providing statistically valid confidence estimates, enhancing model transparency and reliability in clinical practice. Objective: This study aims to improve the reliability and clinical applicability of PI prediction models by integrating CP with traditional ML algorithms, enabling uncertainty-aware predictions for safer and more transparent clinical decisions. Methods: A methodological study was conducted to evaluate ML classifiers for PI risk classification using routinely collected clinical data and alternative data-processing strategies. Model performance was assessed through paired statistical comparisons across resampling folds. Models showing superior performance were subsequently calibrated using a CP framework to generate uncertainty-aware prediction sets, which were evaluated using coverage and efficiency metrics across different confidence levels. Results: The overall incidence of PIs was 27%, with significant predictors including hospital ward, risk level, mattress use, containment, dermatitis, and age (p < 0.05). Among all models, XGBoost (XGB) and Random Forest (RF) showed the best predictive performance. After calibration with CP, both produced well-aligned uncertainty estimates. At α = 0.10, XGB achieved coverage = 0.949 and efficiency = 1.34, outperforming RF with tighter, more informative prediction sets, with a loss of 5% of cases outside the prediction set and an average size of the prediction set closer to 1.0 compared to the RF model. This framework enhanced model interpretability in clinical settings without compromising accuracy. Conclusions: Integrating CP into ML models may improve the interpretability and reliability of risk predictions by quantifying uncertainty. Although the findings are promising, they should be interpreted with caution given the modest sample size, event rate, single-center design, and potential variability in clinical documentation. This framework provides a foundation for uncertainty-aware decision support in PI prevention.