Search

Objective The study aimed to represent the content of nursing diagnosis and interventions in the openEHR standard. Methods This is a developmental study with the models developed according to ISO 18104: 2014. The Ocean Archetype Editor tool from the openEHR Foundation was used. Results Two archetypes were created; one to represent the nursing diagnosis concept and the other the nursing intervention concept. Existing archetypes available in the Clinical Knowledge Manager were reused in modeling. Conclusion The representation of nursing diagnosis and interventions based on the openEHR standard contributes to representing nursing care phenomena and needs in health information systems.

Background: Leprosy reactions (LR) are severe episodes of intense activation of the host inflammatory response, of uncertain etiology, today the leading cause of permanent nerve damage in leprosy patients. Several genetic and non-genetic risk factors for LR have been described; however, there are limited attempts to combine this information in order to estimate the risk of a leprosy patient to develop LR. Here we present an artificial intelligence (AI)-based system able to estimate risk of LR using clinical, demographic and genetic data.Methods: The study includes four datasets from different regions of Brazil, totalizing 1,450 leprosy patients followed prospectively for at least two years to assess the occurrence of LR. Data mining using WEKA software was performed following a two-step protocol to select the variables included in the AI system, based on Bayesian Networks and developed using the NETICA software.Results: Analysis of the complete database resulted in a system able to estimate LR-risk with 82.7% accuracy, 79.3% sensitivity, and 86.2% specificity. When using only databases for which host genetic information associated with LR was included, the performance increased to up to 87.7% accuracy, 85.7% sensitivity, and 89.4% specificity.Conclusion: We produced an easy-to-use, online, free-access system that allows the identification of leprosy patients at high risk of developing LR. Risk assessment of LR for individual patients may detect candidates close monitoring, with potential positive impact upon the prevention of permanent disabilities, the quality of life of the patients, as well as upon leprosy control programs.

Electronic health records (EHRs) contain patient-related information formed by structured and unstructured data, a valuable data source for Natural Language Processing (NLP) in the healthcare domain. The contextual word embeddings and Transformer-based models have proved their potential, reaching state-of-the-art for various NLP tasks. Although the performance for downstream NLP tasks with free-texts written in English has recently improved, less resource is available considering clinical texts and low-resource languages such as Portuguese. Our objective is to develop a Generative Pre-trained Transformer 2 (GPT-2) language model for Portuguese to support clinical and biomedical NLP tasks. We fine-tuned a generic Portuguese GPT-2 model to corpora of biomedical texts written in Portuguese, using transfer learning. We experimented on a public dataset, manually annotated for detecting patient fall, i.e., a classification task. Our in-domain GPT-2 model outperformed the generic Portuguese GPT-2 model by 3.43 in F1-score (weighted). Our preliminary results show that transfer learning with domain literature can benefit Portuguese biomedical NLP tasks, aligned with other languages' results.

Question answering (QA) systems aim to answer human questions made in natural language. This type of functionality can be very useful in the most diverse application domains, such as the biomedical and clinical. Considering the clinical context, where we have a growing volume of information stored in electronic health records, answering questions about the patient status can improve the decision-making and optimize the patient care. In this work, we carried out the first experiments to develop a QA model for clinical texts in Portuguese. To overcome the lack of corpora for the required language and context, we used a transfer learning approach supported by pre-trained attention-based models from the Transformers library. We fine-tuned the BioBERTpt model with a translated version of the SQuAD dataset. The evaluation showed promising results when evaluated in different clinical scenarios, even without the application of a clinical QA corpus to support a training process. The developed model is publicly available to the scientific community.

Automatic detection of negated content is often a prerequisite in information extraction systems in various domains. In the biomedical domain especially, this task is important because negation plays an important role. In this work, two main contributions are proposed. First, we work with languages which have been poorly addressed up to now: Brazilian Portuguese and French. Thus, we developed new corpora for these two languages which have been manually annotated for marking up the negation cues and their scope. Second, we propose automatic methods based on supervised machine learning approaches for the automatic detection of negation marks and of their scopes. The methods show to be robust in both languages (Brazilian Portuguese and French) and in cross-domain (general and biomedical languages) contexts. The approach is also validated on English data from the state of the art: it yields very good results and outperforms other existing approaches. Besides, the application is accessible and usable online. We assume that, through these issues (new annotated corpora, application accessible online, and cross-domain robustness), the reproducibility of the results and the robustness of the NLP applications will be augmented.

Background Cross-mapping establishes equivalence between terms from different terminology systems, which is useful for interoperability, updated terminological versions, and reuse of terms. Due to the number of terms to be mapped, this work can be extensive, tedious, and thorough, and it is susceptible to errors; this can be minimized by automated processes, which use computational tools. Objective The aim of this study was to compare the results of manual and automated term mapping processes. Methods In this descriptive, quantitative study, we used the results of two mapping processes as an empirical basis: manual, which used 2638 terms of nurses’ records from a university hospital in southern Brazil and the International Classification for Nursing Practice (ICNP); and automated, which used the same university hospital terms and the primitive terms of the ICNP through MappICNP, an algorithm based on rules of natural language processing. The two processes were compared via equality and exclusivity assessments of new terms of the automated process and of candidate terms. Results The automated process mapped 569/2638 (21.56%) of the source bank’s terms as identical, and the manual process mapped 650/2638 (24.63%) as identical. Regarding new terms, the automated process mapped 1031/2638 (39.08%) of the source bank’s terms as new, while the manual process mapped 1251 (47.42%). In particular, manual mapping identified 101/2638 (3.82%) terms as identical and 429 (16.26%) as new, whereas the automated process identified 20 (0.75%) terms as identical and 209 (7.92%) as new. Of the 209 terms mapped as new by the automated process, it was possible to establish an equivalence with ICNP terms in 48 (23.0%) cases. An analysis of the candidate terms offered by the automated process to the 429 new terms mapped exclusively by the manual process resulted in 100 (23.3%) candidates that had a semantic relationship with the source term. Conclusions The automated and manual processes map identical and new terms in similar ways and can be considered complementary. Direct identification of identical terms and the offering of candidate terms through the automated process facilitate and enhance the results of the mapping; confirmation of the precision of the automated mapping requires further analysis by researchers.

Objective To describe a method of analysis for understanding the health care process, enriched with information on the clinical and profile characteristics of the patients. To apply the proposed technique to analyze an ischemic stroke dataset. Materials and methods We analyzed 4,830 electronic health records (EHRs) from patients with ischemic stroke (2010–2017), containing information about events realized during treatment and clinical and profile information of the patients. The proposed method combined process mining techniques with data analysis, grouping the data by primary care units (PCU — units responsible for the primary care of patients residing in a geographical area). Results A novel method, named process, data, and management (PDM) analysis method was used for ischemic stroke data and it provided the following outcomes: health care process for patients with ischemic stroke with time statistics; analysis of potential factors for slow hospital admission indicating an increase in the time to hospital admission of 3.4 h (mean value) for patients with an origin at the urgent care center (UCC) – 30% of patients; analysis of PCUs with distinct secondary stroke rates indicating that the social class of patients is the main difference between them; and the visualization of risk factors (before the stroke) by the PCU to inform the health manager about the potential of prevention. Discussion PDM analysis describes a step-by-step method for combining process analysis with data analysis considering a management focus. The results obtained on the stroke context can support the definition of more refined action plans by the health manager, improving the stroke health care process and preventing new events. Conclusion When a patient is diagnosed with ischemic stroke, immediate treatment is needed. Moreover, it is possible to prevent new events to some degree by monitoring and treating risk factors. PDM analysis provides an overview of the health care process with time, combining elements that affect the treatment flow and factors, which can indicate a potential for preventing new events. We also can apply PDM analysis in different scenarios, when there is information about activities from treatment flow and other characteristics related to the treatment or the prevention of the analyzed disease. The management focus of the results aids in the formulation of service policies, action plans, and resource allocation.

Objectives: Clinical Named Entity Recognition is a critical Natural Language Processing task, as it could support biomedical research and healthcare systems. While most extracted clinical entities are based on single-label concepts, it is very common in the clinical domain entities with more than one semantic category simultaneously. This work proposes BERT-based models to support multilabel clinical named entity recognition in the Portuguese language. Methods: For the experiment, we used the Label Powerset method applied to the multilabel corpus SemClinBr. Results: We compare our results with a Conditional Random Fields baseline, reaching +2.1 in precision, +11.2 in recall, and +7.4 in F1 with a clinical-biomedical BERT model (BioBERTpt). Conclusion: We achieved higher results for both exact and partial metrics, contributing to the multilabel semantic processing of clinical narratives in Portuguese.

In this paper, we trained a set of Portuguese clinical word embedding models of different granularities from multi-specialty and multi-institutional clinical narrative datasets. Then, we assessed their impact on a downstream biomedical NLP task of Urinary Tract Infection disease identification. Additionally, we intrinsically evaluated our main model using an adapted version of Bio-SimLex for the Portuguese language. Our empirical results showed that the larger, coarse-grained model achieved a slightly better outcome when compared with the small, fine-grained model in the proposed task. Moreover, we obtained satisfactory results with Bio-SimLex intrinsic evaluation.

Diabetic Retinopathy (DR) is one of the most common microvascular complications presenting by patients diagnosticated with diabetic diseases. Uncontrolled hyperglycemia may manifest as visual impairment and blindness. The early detection of DR is essential to minimize the risk and consequence of visual diminishing. The standard gold diagnoses tool relies on different imaging modalities and requires a judgment of expert photographers, which are not available in most of the primary care centers or remote location. In that scenario, an automate or semiautomated DR screening systems can contribute to improving the accuracy of the diagnostic. Thus, we performed a Systematic Review and Meta-Analysis to evaluate the Decision Support Systems (DSS) in diagnosing DR. The overall Diagnostic Odds Ratio was 73.15 (95%CI: 37.54-142.50), sensitivity was 97.70 (95%CI: 97.50-97.90) and specificity was 90.30 (95%CI: 90.00-90.60). Our results corroborate with the concept of usefulness of DSSs in early diagnosis, screening and preliminary evaluation of suspicious images of DR.