Conference Agenda

Pre-analytical laboratory processes are highly susceptible to operational errors due to manual handling and visual identification of blood collection tube boxes, negatively affecting efficiency, reliability, and workflow consistency. Although commercial automation solutions exist, their high acquisition and integration costs limit adoption in small and medium-sized laboratories, particularly in resource-constrained environments. This paper presents the operational validation of an integrated system combining machine vision and programmable logic control (PLC) for automated pre-analytical tube-box classification, using industrial simulation as the primary validation environment.

The proposed approach follows the VDI 2206 methodology for mechatronic system development and implements a three-level architecture inspired by IEC 62264, including a field level with virtual sensors and actuators, a control level based on a Siemens S7-1200 PLC programmed in Ladder logic, and a supervision level using a human–machine interface (HMI). A vision module based on HSV color-space processing classifies tube boxes according to color-coded attributes. High-fidelity simulation enables controlled experimentation without physical deployment.

Operational validation includes performance evaluation, robustness analysis, and extreme condition testing. The system achieves a classification accuracy of 97.3%, a sustained throughput of 219.5 boxes/h, and an average cycle time of 1.375 s. Integration feasibility is confirmed through a Vision–PLC latency compliance of 99.3% (≤200 ms). Reliability analysis via Monte Carlo simulation estimates an MTBF of 2,347 h, and stress tests confirm stable behavior under adverse scenarios.

Cardiac hemodynamics describes the behavior of blood flow and pressure within the cardiovascular system, constituting a central pillar in professional training for understanding cardiac mechanics and for the diagnosis of cardiac conditions in the clinical field. However, its understanding is limited by the complexity of the phenomenon and the limited accessibility to practical simulation and training tools. This work presents the development of a preliminary prototype of a physical–interactive cardiac flow simulator, designed as a low-cost educational tool for teaching cardiac physiology and pathophysiology by recreating essential hemodynamic aspects for the study of these phenomena. The methodology was structured in three stages: identification of relevant hemodynamic variables; design and integration of the physical, electronic, and control systems; and preliminary validation through expert evaluation, in addition to a pilot study with medical students. The results demonstrated high usability and adequate functional fidelity of the prototype, highlighting its usefulness in the educational field and the possibility of expanding its scope due to its design and structure. It is concluded that the simulator is technically viable and has the potential to strengthen the learning of complex hemodynamic concepts.

Acute diarrhea disease remains a major cause of morbidity and mortality in children under five years of age in Honduras and is closely related to water, sanitation and hygiene (WASH) conditions. To estimate the counterfactual causal effect of improvements in WASH on the prevalence of childhood diarrhea, a representative sample of 7,500 children adjusted to the distributions reported by ENDESA/MICS 2019 was simulated. Doubly robust estimators (IIPW) were used to estimate average causal effects, and five counterfactual intervention scenarios were analyzed, as well as heterogeneity by wealth quintile, area of residence, and age group. Simulated coverage of improved water was 88.1%, improved sanitation 76.3%, and basic hygiene 81.2%, with a prevalence of diarrhea of 4.1%. The estimated ATE showed decreases of -1.79 percentage points (pp) for improved water (95% CI: -3.63, 0.05), -1.93 pp for sanitation (95% CI: -3.72, -0.14) and -1.66 pp for hygiene (95% CI: -3.02, -0.30). The comprehensive WASH package was shown to have the greatest potential for reduction (0.74 pp), translating to 7.4 cases prevented per 1,000 children, with greater benefits in the poorest quintiles and in the 12–23 month group. Taken together, the findings indicate that integrated and targeted WASH interventions have the potential to significantly reduce the burden of childhood diarrhea in Honduras.

Gut-brain Axis Disorders (GBAD) afflict 15-20% of Hondurans, making them a major public health issue. At-risk people can be identified early to improve Primary Health Care (PHC) prevention. Objective: This work developed and validated machine learning models to predict GBAD presence using sociodemographic, behavioral, and environmental characteristics from standardized PHC surveys. Methods: We generated 1,200 synthetic patient records using epidemiological parameters from the literature and a validated survey instrument for communities on the Public Health IV rotation at the Universidad Nacional Autonoma de Honduras (UNAH). The following classification techniques were tested: Random Forest (RF), Support Vector Machine (SVM), Gradient Boosting (GB), and Multilayer Perceptron Neural Network. All algorithms were analyzed for feature relevance and model performance using 10-fold stratified cross-validation. Results: The simulated dataset had 27.5% GBAD prevalence, matching regional estimates. Random Forest performed best in cross-validation (AUC = 0.594 ± 0.067), followed by Gradient Boosting (AUC = 0.591 ± 0.068). Cross-model feature importance analysis found age, perceived stress, fiber consumption, and education as the most influential predictors. Conclusions: Machine learning has moderate potential for PHC GBAD screening. The identified risk factors support epidemiological research and give prevention program targets. Integrating these models into PHC operations could aid early detection and intervention.

Chronic child malnutrition is a persistent public health problem in Honduras, with irreversible consequences on cognitive and physical development. The aim of this study was to develop and validate predictive models of chronic malnutrition (stunting) using explainable machine learning, to identify the most influential determinants and estimate their individual contribution to risk. Data from 8,713 children aged 0 to 59 months from ENDESA/MICS 2019 were analyzed. A wealth index was constructed using principal component analysis on 21 household assets and three algorithms (logistic regression, Random Forest and Gradient Boosting) were trained with stratified cross-validation. Interpretability was evaluated with SHAP (SHapley Additive exPlanations) values. The prevalence of stunting was 18.9%. Random Forest presented the best performance (AUC-ROC=0.691; AUC-PR=0.362). SHAP analyses identified wealth index as the primary predictor (mean SHAP=0.489), followed by child age (0.323) and maternal education (0.251). A marked socioeconomic gradient was observed (Q1: 39.2% vs Q5: 7.1%), with amplification of the effect in rural areas. In conclusion, explainable machine learning allows the identification and quantification of key determinants of chronic child malnutrition, supporting interventions focused on economic transfers, improvements in water and sanitation, and women's education, with territorial prioritization in the western corridor of the country.

Child mortality remains a critical public health indicator in low- and middle-income countries, and risk stratification using machine learning represents a promising approach to target targeted interventions. This study aimed to develop and validate predictive models for neonatal (0–27 days) and postneonatal (28–364 days) mortality in Honduras, and to identify their main determinants using interpretability methods. A retrospective cohort study was conducted with data from ENDESA/MICS 2019, comparing three classification models (logistic regression, Random Forest and Gradient Boosting). Performance was assessed with AUC-ROC, AUC-PR, and Brier score, and interpretability was addressed using SHAP values. Among 9,579 live births, neonatal mortality was 14.2 per 1,000 and postneonatal mortality was 8.6 per 1,000. Gradient Boosting showed the best calibration (Brier=0.029 for neonatal; 0.019 for postneonatal). The most influential determinants of neonatal mortality were parity, wealth quintile and access to improved water; for post-neonatal mortality, the wealth quintile, parity and rural residence stood out. In conclusion, machine learning models interpretable using SHAP allow the identification of differential risk profiles for neonatal and postneonatal mortality, supporting the prioritization of maternal and child health interventions.

Facial paralysis affects motor function, communication, emotional expression, and patient self-esteem, making rehabilitation a complex clinical challenge. This study presents the development of an intelligent therapeutic assistant for facial rehabilitation based on surface electromyography (sEMG). The system processes muscle biopotentials through filtering, normalization, and feature extraction to identify activation patterns that define optimal time windows for therapeutic interventions such as electrical stimulation or massage therapy. A mixed-method design was applied with patients presenting facial motor dysfunction, classified using the House-Brackmann (HB) scale, who participated in rehabilitation sessions supported by the system. A protocol for electrode placement and size was established to ensure signal quality and minimize contamination. Evaluation included algorithm performance metrics, muscle activation records, and clinical observations regarding detection accuracy and practical usefulness. Sessions were conducted in a controlled environment, considering patient comfort and variability in facial movements. Statistical analyses compared muscle activation detection before and after intervention and assessed the system’s capacity to support therapeutic decision-making. Preliminary results indicate that the assistant enhances precision and personalization in therapy, reduces therapist uncertainty, and improves intervention effectiveness. This research highlights the potential of sEMG-based intelligent systems in facial rehabilitation and their integration into clinical
and telemedicine contexts. Future studies should expand sample size, extend intervention duration, incorporate additional physiological markers, and establish standardized protocols for broader clinical application.