http://hdl.handle.net/11624/151 Tue, 14 Jul 2026 02:13:50 GMT 2026-07-14T02:13:50Z https://repositorio.unisc.br:443/jspui/retrieve/372/thumb_4c4592cd12911.jpg http://hdl.handle.net/11624/151 http://hdl.handle.net/11624/4309 Title: Sistema anonimizado de dados a partir de prontuários de saúde. Authors: Martini, Patrick Luiz Abstract: The increasing digitalization of healthcare services and the adoption of electronic health records by healthcare institutions have generated a significant volume of clinical data, representing a valuable resource for academic research and the advancement of evidence-based medicine. The enactment of the General Data Protection Law (GDPL) established a fundamental regulatory framework for protecting citizens' privacy, setting guidelines for the processing of sensitive personal data, including health-related information. In this scenario, it is essential to develop technological solutions that allow for the secondary use of clinical data for research purposes while ensuring full compliance with the data protection principles established by legislation. In this context, this study aims to develop a system for the anonymization of health records, enabling their safe and regulated use in academic research in compliance with the LGPD. The adopted methodology combined a systematic literature review, following the PRISMA method, with the implementation of a system based on Large Language Models (LLMs) and Natural Language Processing (NLP) techniques. The selection of the language model considered requirements for Portuguese language support, data privacy, and local processing capacity, resulting in the adoption of the Gemma 3 4B model. The system implements a two-stage anonymization pipeline, using Named Entity Recognition (NER) techniques to identify and suppress five categories of personal identifiers present in medical records: personal names, ages, service dates, geographic locations, and organizations. For structured data, deterministic encryption (BLAKE2b) is applied for the pseudonymization of identifiers and date transformation, preserving temporal relationships and data consistency. Additionally, an LGPD-based audit prompt acts as a verification layer to mitigate false negatives. The result is an anonymized database that protects patient privacy while preserving data utility for academic analysis, contributing to the advancement of healthcare research by facilitating the secure sharing of information between institutions and researchers. Thu, 01 Jan 2026 00:00:00 GMT http://hdl.handle.net/11624/4309 2026-01-01T00:00:00Z http://hdl.handle.net/11624/4307 Title: Proposta de um modelo de detecção de tuberculose por emissão de plasma em ar pulmonar. Authors: Gündel, Mateus Elias Abstract: Tuberculosis (TB) remains a major global public-health concern, with approximately 10.8 million incident cases and 1.25 million deaths reported in 2023 (World Health Organization, 2024), and disproportionately high incidence in vulnerable populations. Conventional diagnostic methods (smear microscopy, microbiological culture, and molecular tests such as GeneXpert MTB/RIF) rely on laboratory infrastructure that limits their access in low-resource settings. This dissertation investigates the feasibility of a machine-learning model for TB screening based on plasma emission images generated by the BDEE-Device (Breath Diagnostic Electronic Eye Device), a portable instrument that produces microplasma from exhaled pulmonary breath and captures its optical emission as a digital image. The work is organized in the Scandinavian (article-based) format and comprises two papers. Paper I, a systematic review conducted under the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) protocol covering studies published between 2019 and 2024, identified six relevant works whose performance varies widely (sensitivities between 0.52 and 1.00) for technologies such as comprehensive two-dimensional gas chromatography time-of-flight mass spectrometry (GC×GC-TOFMS) and electronic noses combined with algorithms such as Random Forest and support vector machines (SVM); none of the reviewed studies, however, combine plasma emission, image capture, and computer vision-based classification for TB detection, defining the gap that motivates the second paper. Paper II proposes and experimentally evaluates a classification approach on the PEVA dataset (2,154 images from 60 patients; 7 positive, 53 negative), comparing classical classifiers with traditional visual descriptors and three convolutional neural network (CNN) architectures with transfer learning, under stratified cross-validation with strict patient-wise split and a hold-out test set. MobileNetV2 achieved the best performance in cross-validation (area under the ROC curve, AUC = 0.911 ± 0.166) and on the hold-out test set (AUC = 0.873; 95% confidence interval [0.820; 0.915]), with limited sensitivity (0.150). Gradient-weighted Class Activation Mapping (Grad-CAM) interpretability analysis revealed evidence of shortcut learning, and the complementary blob cropping experiment doubled test sensitivity (to 0.300) without degrading AUC. Adomain shift analysis between versions 1.0 and 2.0 of the device revealed a severe drop in performance, suggesting that part of the learned patterns may be instrumental. The reported metrics must be interpreted with caution given the high statistical variance associated with the small number of positive patients (n=7), which reinforces the proof-of-concept nature of this work. The results demonstrate the technical feasibility of this modality as a proof of concept for point-of-care TB screening and point, as central directions, to dataset expansion, prospective multicenter validation, and the investigation of domain-adaptation techniques. Thu, 01 Jan 2026 00:00:00 GMT http://hdl.handle.net/11624/4307 2026-01-01T00:00:00Z http://hdl.handle.net/11624/4271 Title: Plataforma digital com agente conversacional e gamificação para educação em resíduos de saúde. Authors: Ferreira, Guilherme Abstract: Healthcare waste management presents critical challenges for public health and environmental sustainability, particularly in contexts where improper disposal practices, such as syringes and expired medications, have severe environmental and collective health impacts. Despite clear regulations, effective implementation of these guidelines faces barriers related to the lack of technical knowledge and engagement from both healthcare professionals and the general population. To mitigate this problem, the goal is to develop a digital platform that integrates a conversational agent, named Dóris®, and gamification elements, promoting practical and interactive education on healthcare waste management. The platform includes educational games, interactive articles, and quizzes, offering dynamic learning with content personalization through the virtual agent Dóris®, who guides users in real time. Validation conducted with 25 participants from various professional categories, such as nurses, radiology technicians, and cleaning professionals, demonstrated high efficacy in knowledge retention and adherence to proper disposal practices. The results show that integrating artificial intelligence and gamification contributes not only to learning but also promotes sustainable behavioral changes, making waste management an accessible and impactful practice on a large scale. This innovative solution stands out for its ability to combine advanced technology and interactive methods to transform contemporary challenges into opportunities for environmental education and sustainability. Wed, 01 Jan 2025 00:00:00 GMT http://hdl.handle.net/11624/4271 2025-01-01T00:00:00Z http://hdl.handle.net/11624/4270 Title: Sistema de gestão farmacêutica na área de atenção primária. Authors: Rutsatz, Rafael Fernando Wed, 01 Jan 2025 00:00:00 GMT http://hdl.handle.net/11624/4270 2025-01-01T00:00:00Z