A pseudonymized corpus of occupational health narratives for clinical entity recognition in Spanish.

Despite the high creation cost, annotated corpora are indispensable for robust natural language processing systems. In the clinical field, in addition to annotating medical entities, corpus creators must also remove personally identifiable information (PII). This has become increasingly important in the era of large language models where unwanted memorization can occur. This paper presents a corpus annotated to anonymize personally identifiable information in 1,787 anamneses of work-related accidents and diseases in Spanish. Additionally, we applied a previously released model for Named Entity Recognition (NER) trained on referrals from primary care physicians to identify diseases, body parts, and medications in this work-related text. We analyzed the differences between the models and the gold standard curated by a physician in detail. Moreover, we compared the performance of the NER model on the original narratives, in narratives where personal information has been masked, and in texts where the personal data is replaced by another similar surrogate value (pseudonymization). Within this publication, we share the annotation guidelines and the annotated corpus.

Supporting the classification of patients in public hospitals in Chile by designing, deploying and validating a system based on natural language processing.

BACKGROUND: In Chile, a patient needing a specialty consultation or surgery has to first be referred by a general practitioner, then placed on a waiting list. The Explicit Health Guarantees (GES in Spanish) ensures, by law, the maximum time to solve 85 health problems. Usually, a health professional manually verifies if each referral, written in natural language, corresponds or not to a GES-covered disease. An error in this classification is catastrophic for patients, as it puts them on a non-prioritized waiting list, characterized by prolonged waiting times. METHODS: To support the manual process, we developed and deployed a system that automatically classifies referrals as GES-covered or not using historical data. Our system is based on word embeddings specially trained for clinical text produced in Chile. We used a vector representation of the reason for referral and patient's age as features for training machine learning models using human-labeled historical data. We constructed a ground truth dataset combining classifications made by three healthcare experts, which was used to validate our results. RESULTS: The best performing model over ground truth reached an AUC score of 0.94, with a weighted F1-score of 0.85 (0.87 in precision and 0.86 in recall). During seven months of continuous and voluntary use, the system has amended 87 patient misclassifications. CONCLUSION: This system is a result of a collaboration between technical and clinical experts, and the design of the classifier was custom-tailored for a hospital's clinical workflow, which encouraged the voluntary use of the platform. Our solution can be easily expanded across other hospitals since the registry is uniform in Chile.

[Construction of text resources for automatic identification of clinical information in unstructured narratives]. / Construcción de recursos de texto para la identificación automática de información clínica en narrativas no estructuradas.

BACKGROUND: A significant proportion of the clinical record is in free text format, making it difficult to extract key information and make secondary use of patient data. Automatic detection of information within narratives initially requires humans, following specific protocols and rules, to identify medical entities of interest. AIM: To build a linguistic resource of annotated medical entities on texts produced in Chilean hospitals. MATERIAL AND METHODS: A clinical corpus was constructed using 150 referrals in public hospitals. Three annotators identified six medical entities: clinical findings, diagnoses, body parts, medications, abbreviations, and family members. An annotation scheme was designed, and an iterative approach to train the annotators was applied. The F1-Score metric was used to assess the progress of the annotator's agreement during their training. RESULTS: An average F1-Score of 0.73 was observed at the beginning of the project. After the training period, it increased to 0.87. Annotation of clinical findings and body parts showed significant discrepancy, while abbreviations, medications, and family members showed high agreement. CONCLUSIONS: A linguistic resource with annotated medical entities on texts produced in Chilean hospitals was built and made available, working with annotators related to medicine. The iterative annotation approach allowed us to improve performance metrics. The corpus and annotation protocols will be released to the research community.

Prolonged wait time is associated with increased mortality for Chilean waiting list patients with non-prioritized conditions.

BACKGROUND: Most data on mortality and prognostic factors of universal healthcare waiting lists come from North America, Australasia, and Europe, with little information from South America. We aimed to determine the relationship between medical center-specific waiting time and waiting list mortality in Chile. METHOD: Using data from all new patients listed in medical specialist waitlists for non-prioritized health problems from 2008 to 2015 in three geographically distant regions of Chile, we constructed hierarchical multivariate survival models to predict mortality risk at two years after registration for each medical center. Kendall rank correlation analysis was used to measure the association between medical center-specific mortality hazard ratio and waiting times. RESULT: There were 987,497 patients waiting for care at 77 medical centers, including 33,546 (3.40%) who died within two years after registration. Male gender (hazard ratio [HR] = 1.17, 95% confidence interval [CI] 1.1-1.24), older age (HR = 2.88, 95% CI 2.72-3.05), urban residence (HR = 1.19, 95% CI 1.09-1.31), tertiary care (HR = 2.2, 95% CI 2.14-2.26), oncology (HR = 3.57, 95% CI 3.4-3.76), and hematology (HR = 1.6, 95% CI 1.49-1.73) were associated with higher risk of mortality at each medical center with large region-to-region variations. There was a statistically significant association between waiting time variability and death (Z = 2.16, P = 0.0308). CONCLUSION: Patient wait time for non-prioritized health conditions was associated with increased mortality in Chilean hospitals.

[Automatic keyword retrieval from clinical texts: an application of natural language processing to massive data of Chilean suspected diagnosis]. / Obtención automática de palabras clave en textos clínicos: una aplicación de procesamiento del lenguaje natural a datos masivos de sospecha diagnóstica en Chile.

BACKGROUND: Free-text imposes a challenge in health data analysis since the lack of structure makes the extraction and integration of information difficult, particularly in the case of massive data. An appropriate machine-interpretation of electronic health records in Chile can unleash knowledge contained in large volumes of clinical texts, expanding clinical management and national research capabilities. AIM: To illustrate the use of a weighted frequency algorithm to find keywords. This finding was carried out in the diagnostic suspicion field of the Chilean specialty consultation waiting list, for diseases not covered by the Chilean Explicit Health Guarantees plan. MATERIAL AND METHODS: The waiting lists for a first specialty consultation for the period 2008-2018 were obtained from 17 out of 29 Chilean health services, and total of 2,592,925 diagnostic suspicions were identified. A natural language processing technique called Term Frequency-Inverse Document Frequency was used for the retrieval of diagnostic suspicion keywords. RESULTS: For each specialty, four key words with the highest weighted frequency were determined. Word clouds showing words weighted by their importance were created to obtain a visual representation. These are available at cimt.uchile.cl/lechile/. CONCLUSIONS: The algorithm allowed to summarize unstructured clinical free-text data, improving its usefulness and accessibility.

Evaluating the 2014 sugar-sweetened beverage tax in Chile: An observational study in urban areas.

BACKGROUND: In October 2014, Chile implemented a tax modification on sugar-sweetened beverages (SSBs) called the Impuesto Adicional a las Bebidas Analcohólicas (IABA). The design of the tax was unique, increasing the tax on soft drinks above 6.25 grams of added sugar per 100 mL and decreasing the tax for those below this threshold. METHODS AND FINDINGS: This study evaluates Chile's SSB tax, which was announced in March 2014 and implemented in October 2014. We used household-level grocery-purchasing data from 2011 to 2015 for 2,836 households living in cities representative of the urban population of Chile. We employed a fixed-effects econometric approach and estimated the before-after change in purchasing of SSBs controlling for seasonality, general time trend, temperature, and economic fluctuations as well as time-invariant household characteristics. Results showed significant changes in purchasing for the statistically preferred model: while there was a barely significant decrease in the volume of all soft drinks, there was a highly significant decrease in the monthly purchased volume of the higher-taxed, sugary soft drinks by 21.6%. The direction of this reduction was robust to different empirical modelling approaches, but the statistical significance and the magnitude of the changes varied considerably. The reduction in soft drink purchasing was most evident amongst higher socioeconomic groups and higher pretax purchasers of sugary soft drinks. There was no systematic, robust pattern in the estimates by household obesity status. After tax implementation, the purchase prices of soft drinks decreased for the items for which the tax rate was reduced, but they remained unchanged for sugary items, for which the tax was increased. However, the purchase prices increased for sugary soft drinks at the time of the policy announcement. The main limitations include a lack of a randomised design, limiting the extent of causal inference possible, and the focus on purchasing data rather than consumption or health outcomes. CONCLUSIONS: The results of subgroup analyses suggest that the policy may have been partially effective, though not necessarily in ways that are likely to reduce socioeconomic inequalities in diet-related health. It remains unclear whether the policy has had a major, overall population-level impact. Additionally, because the present study examined purchasing of soft drinks for only 1 year, a longer-term evaluation-ideally including an assessment of consumption and health impacts-should be conducted in future research. TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT02926001.

Physics-informed neural networks for parameter estimation in blood flow models.

BACKGROUND: Physics-informed neural networks (PINNs) have emerged as a powerful tool for solving inverse problems, especially in cases where no complete information about the system is known and scatter measurements are available. This is especially useful in hemodynamics since the boundary information is often difficult to model, and high-quality blood flow measurements are generally hard to obtain. METHODS: In this work, we use the PINNs methodology for estimating reduced-order model parameters and the full velocity field from scatter 2D noisy measurements in the aorta. Two different flow regimes, stationary and transient were studied. RESULTS: We show robust and relatively accurate parameter estimations when using the method with simulated data, while the velocity reconstruction accuracy shows dependence on the measurement quality and the flow pattern complexity. Comparison with a Kalman filter approach shows similar results when the number of parameters to be estimated is low to medium. For a higher number of parameters, only PINNs were capable of achieving good results. CONCLUSION: The method opens a door to deep-learning-driven methods in the simulations of complex coupled physical systems.

Automatic Detection of Distant Metastasis Mentions in Radiology Reports in Spanish.

PURPOSE: A critical task in oncology is extracting information related to cancer metastasis from electronic health records. Metastasis-related information is crucial for planning treatment, evaluating patient prognoses, and cancer research. However, the unstructured way in which findings of distant metastasis are often written in radiology reports makes it difficult to extract information automatically. The main aim of this study was to extract distant metastasis findings from free-text imaging and nuclear medicine reports to classify the patient status according to the presence or absence of distant metastasis. MATERIALS AND METHODS: We created a distant metastasis annotated corpus using positron emission tomography-computed tomography and computed tomography reports of patients with prostate, colorectal, and breast cancers. Entities were labeled M1 or M0 according to affirmative or negative metastasis descriptions. We used a named entity recognition model on the basis of a bidirectional long short-term memory model and conditional random fields to identify entities. Mentions were subsequently used to classify whole reports into M1 or M0. RESULTS: The model detected distant metastasis mentions with a weighted average F1 score performance of 0.84. Whole reports were classified with an F1 score of 0.92 for M0 documents and 0.90 for M1 documents. CONCLUSION: These results show the usefulness of the model in detecting distant metastasis findings in three different types of cancer and the consequent classification of reports. The relevance of this study is to generate structured distant metastasis information from free-text imaging reports in Spanish. In addition, the manually annotated corpus, annotation guidelines, and code are freely released to the research community.

Training and intrinsic evaluation of lightweight word embeddings for the clinical domain in Spanish.

Resources for Natural Language Processing (NLP) are less numerous for languages different from English. In the clinical domain, where these resources are vital for obtaining new knowledge about human health and diseases, creating new resources for the Spanish language is imperative. One of the most common approaches in NLP is word embeddings, which are dense vector representations of a word, considering the word's context. This vector representation is usually the first step in various NLP tasks, such as text classification or information extraction. Therefore, in order to enrich Spanish language NLP tools, we built a Spanish clinical corpus from waiting list diagnostic suspicions, a biomedical corpus from medical journals, and term sequences sampled from the Unified Medical Language System (UMLS). These three corpora can be used to compute word embeddings models from scratch using Word2vec and fastText algorithms. Furthermore, to validate the quality of the calculated embeddings, we adapted several evaluation datasets in English, including some tests that have not been used in Spanish to the best of our knowledge. These translations were validated by two bilingual clinicians following an ad hoc validation standard for the translation. Even though contextualized word embeddings nowadays receive enormous attention, their calculation and deployment require specialized hardware and giant training corpora. Our static embeddings can be used in clinical applications with limited computational resources. The validation of the intrinsic test we present here can help groups working on static and contextualized word embeddings. We are releasing the training corpus and the embeddings within this publication.

Enhanced diffusion due to active swimmers at a solid surface.

We consider two systems of active swimmers moving close to a solid surface, one being a living population of wild-type E. coli and the other being an assembly of self-propelled Au-Pt rods. In both situations, we have identified two different types of motion at the surface and evaluated the fraction of the population that displayed ballistic trajectories (active swimmers) with respect to those showing randomlike behavior. We studied the effect of this complex swimming activity on the diffusivity of passive tracers also present at the surface. We found that the tracer diffusivity is enhanced with respect to standard Brownian motion and increases linearly with the activity of the fluid, defined as the product of the fraction of active swimmers and their mean velocity. This result can be understood in terms of series of elementary encounters between the active swimmers and the tracers.

On the Construction of Multilingual Corpora for Clinical Text Mining.

The amount of digital data derived from healthcare processes have increased tremendously in the last years. This applies especially to unstructured data, which are often hard to analyze due to the lack of available tools to process and extract information. Natural language processing is often used in medicine, but the majority of tools used by researchers are developed primarily for the English language. For developing and testing natural language processing methods, it is important to have a suitable corpus, specific to the medical domain that covers the intended target language. To improve the potential of natural language processing research, we developed tools to derive language specific medical corpora from publicly available text sources. n order to extract medicine-specific unstructured text data, openly available pub-lications from biomedical journals were used in a four-step process: (1) medical journal databases were scraped to download the articles, (2) the articles were parsed and consolidated into a single repository, (3) the content of the repository was de-scribed, and (4) the text data and the codes were released. In total, 93 969 articles were retrieved, with a word count of 83 868 501 in three different languages (German, English, and Spanish) from two medical journal databases Our results show that unstructured text data extraction from openly available medical journal databases for the construction of unified corpora of medical text data can be achieved through web scraping techniques.

Predicting nationwide obesity from food sales using machine learning.

The obesity epidemic progresses everywhere across the globe, and implementing frequent nationwide surveys to measure the percentage of obese population is costly. Conversely, country-level food sales information can be accessed inexpensively through different suppliers on a regular basis. This study applies a methodology to predict obesity prevalence at the country-level based on national sales of a small subset of food and beverage categories. Three machine learning algorithms for nonlinear regression were implemented using purchase and obesity prevalence data from 79 countries: support vector machines, random forests and extreme gradient boosting. The proposed method was validated in terms of both the absolute prediction error and the proportion of countries for which the obesity prevalence was predicted satisfactorily. We found that the most-relevant food category to predict obesity is baked goods and flours, followed by cheese and carbonated drinks.

Effects of a sugar-sweetened beverage tax on prices and affordability of soft drinks in Chile: A time series analysis.

Chile is one of several countries that recently implemented a fiscal policy to reduce soft drink (SD) intake and obesity. In 2014 the government increased the existing ad-valorem tax on high-sugar SD by 5% and decreased by 3% the tax on low-sugar SD, based on a 6.25gr/100 ml sugar threshold. This study aims to evaluate the tax modification passed-on to consumers through prices, and to calculate changes in affordability of SDs. We analysed nationally representative consumer price index data of 41 soft drinks within 6 beverage categories between 2009 and 2016. Price change post-tax implementation was estimated for different categories (carbonates, juices, concentrates, waters and energy-sport drinks), using time-series analyses. In addition, changes in affordability were evaluated by estimating the changes in prices relative to wages. The price of carbonates increased by 5.60% (CI 95% 3.18-8.03%) immediately after the tax was implemented. A sustained increase in the prices of concentrates was observed after the implementation. Unexpectedly, a smaller increase was also seen for the price of bottled water - a category that saw no tax change. There were no effects for juices and energy-sports drinks. There was a reduction in affordability for carbonates, concentrates and waters. Overall, the fiscal policy was effective in increasing prices and there are some signs of reduced affordability. Results varied substantially among categories directly affected by the tax policy. While for carbonates the price increase exceeded the tax change ('over-shifting'), in other categories subject to a tax cut, a price reduction was expected but the opposite occurred. As the effect of the tax on prices differed between categories, the effects of the tax policy on consumption patterns are likely to be mixed. Our findings underline the need to better understand and anticipate price setting behaviour of firms in response to a tax.

Associations of LEP, CRH, ICAM-1, and LINE-1 methylation, measured in saliva, with waist circumference, body mass index, and percent body fat in mid-childhood.

BACKGROUND: Genetics explains a small proportion of variance in body mass index at the population level. Epigenetics, commonly measured by gene methylation, holds promise for understanding obesity risk factors and mechanisms. METHODS: Participants were 431 adolescents aged 10-15 years. BMI z-score, waist circumference z-score, and percent body fat were measured. Saliva samples were collected and methylation of promoter regions of four candidate genes or sequences (LEP, ICAM-1, CRH, and LINE-1) were measured in 3-4 CpG sites each. Linear regression was used to identify associations of methylation with obesity-related outcomes. RESULTS: After adjusting for age, in sex-stratified analysis, the three obesity-related outcomes were negatively associated with LEP methylation in obese boys only. There were no associations of methylation of the other genes or sequences and the obesity-related outcomes. CONCLUSIONS: Our results are consistent with prior studies that reported sex differences in associations of obesity-related outcomes with LEP methylation, and also as would be expected in adipose tissue, the source of circulating leptin. The findings suggest that saliva might be an acceptable tissue for epigenetics studies in adolescents.

Construcción de recursos de texto para la identificación automática de información clínica en narrativas no estructuradas / Construction of text resources for automatic identification of clinical information in unstructured narratives

Background: A significant proportion of the clinical record is in free text format, making it difficult to extract key information and make secondary use of patient data. Automatic detection of information within narratives initially requires humans, following specific protocols and rules, to identify medical entities of interest. Aim: To build a linguistic resource of annotated medical entities on texts produced in Chilean hospitals. Material and Methods: A clinical corpus was constructed using 150 referrals in public hospitals. Three annotators identified six medical entities: clinical findings, diagnoses, body parts, medications, abbreviations, and family members. An annotation scheme was designed, and an iterative approach to train the annotators was applied. The F1-Score metric was used to assess the progress of the annotator's agreement during their training. Results: An average F1-Score of 0.73 was observed at the beginning of the project. After the training period, it increased to 0.87. Annotation of clinical findings and body parts showed significant discrepancy, while abbreviations, medications, and family members showed high agreement. Conclusions: A linguistic resource with annotated medical entities on texts produced in Chilean hospitals was built and made available, working with annotators related to medicine. The iterative annotation approach allowed us to improve performance metrics. The corpus and annotation protocols will be released to the research community.

Adaptación de una metodología de especificaciones en la calificación y retroalimentación en la enseñanza de cursos de ciencias básicas en una carrera de la salud para mejorar los aprendizajes y la motivación / Adaptation of a specifications gradings and feedback methodology for basic sciences courses teaching in a health science career to improving learning and motivation

Introducción: la enseñanza de cursos de ciencias básicas en carreras de la salud es un desafío por no estar directa e inmediatamente rela-cionada con el ámbito profesional. Por otra parte, las condiciones de estrés que ha impuesto el trabajo a distancia requiere de metodologías motivantes, y, que a su vez permitan una evaluación significativa. Objetivos: reportar las adaptaciones metodológicas y los resultados de una adaptación local de la metodología de especificaciones de las calificaciones y retroalimentación del trabajo. Métodos: se aplica una metodología de formación basada en la retroalimentación en el curso de Física para estudiantes de Tecnología Médica (N=106) durante un semestre. Las calificaciones promedio de los estudiantes fueron comparadas con las obtenidas en años anteriores. Para evaluar el desempeño docente se realizaron 2 encuestas a los estudiantes. El cumplimiento de los logros de aprendizaje se midió mediante auto-evaluación (escala likert 1 a 5) al inicio y término de cada uno de los cuatro capítulos. Resultados: las reprobaciones y eliminaciones de estudiantes en el curso fueron menores a años anteriores, siendo las notas significativamente mayores subiendo desde 4,89 a 6,29 (escala de 1 a 7, p<0,001). Los estudiantes se mostraron en un 95% satisfechos con el desempeño docente y finalmente, la auto-evaluación de logros de aprendizaje mostró un aumento promedio de 1 punto. Conclusiones: la metodología de evaluación basada en especificaciones adaptada a dos entregas y con evaluaciones en una escala no-binaria mejoró el rendimiento, los logros de los aprendizajes esperados y la motivación de los estudiantes.

Background: Teaching basic science courses in health careers is a challenge because these courses are not directly linked to professional practice. On the other hand, the stressful conditions imposed by distance work require motivating methodologies and a meaningful evaluation. Objectives: To report the methodological adaptations and the results of a local adaptation of the specifications grading and feedback methodology. Methods: A training methodology based on feedback is applied in the Physics course for Medical Technology students (N = 106) during one semester. We compared the students' average grades to those obtained in previous years with the same topics. To evaluate the teaching performance, we conducted two student surveys. We measure compliance with learning achievements by self-assessment (Likert scale 1 to 5) at the beginning and end of each of the four chapters. Results: Failures and eliminations of students in the course were lower than previous years, with significantly higher grades from 4.89 to 6.29 (p <0.001). The students were 95% satisfied with the teaching performance, and finally, the self-evaluation of learning achievements showed an average increase of 1 point. Conclusion: The evaluation methodology based on specifications adapted to two deliveries and evaluations on a non-binary scale improved the performance, expected learning achievements, and students' motivation.

Obtención automática de palabras clave en textos clínicos: una aplicación de procesamiento del lenguaje natural a datos masivos de sospecha diagnóstica en Chile / Automatic keyword retrieval from clinical texts: an application of natural language processing to massive data of Chilean suspected diagnosis

Background: Free-text imposes a challenge in health data analysis since the lack of structure makes the extraction and integration of information difficult, particularly in the case of massive data. An appropriate machine-interpretation of electronic health records in Chile can unleash knowledge contained in large volumes of clinical texts, expanding clinical management and national research capabilities. Aim: To illustrate the use of a weighted frequency algorithm to find keywords. This finding was carried out in the diagnostic suspicion field of the Chilean specialty consultation waiting list, for diseases not covered by the Chilean Explicit Health Guarantees plan. Material and Methods: The waiting lists for a first specialty consultation for the period 2008-2018 were obtained from 17 out of 29 Chilean health services, and total of 2,592,925 diagnostic suspicions were identified. A natural language processing technique called Term Frequency-Inverse Document Frequency was used for the retrieval of diagnostic suspicion keywords. Results: For each specialty, four key words with the highest weighted frequency were determined. Word clouds showing words weighted by their importance were created to obtain a visual representation. These are available at cimt.uchile.cl/lechile/. Conclusions: The algorithm allowed to summarize unstructured clinical free-text data, improving its usefulness and accessibility.