Machine-Guided Polymer Knowledge Extraction Using Natural Language Processing: The Example of Named Entity Normalization.

A rich body of literature has emerged in recent years that discusses the extraction of structured information from materials science text through named entity recognition models. Relatively little work has been done to address the "normalization" of extracted entities, that is, recognizing that two or more seemingly different entities actually refer to the same entity in reality. In this work, we address the normalization of polymer named entities, polymers being a class of materials that often have a variety of common names for the same material in addition to the IUPAC name. We have trained supervised clustering models using Word2Vec and fastText word embeddings reported in previous work so that named entities referring to the same polymer are categorized within the same cluster in the word embedding space. We report the use of parameterized cosine distance functions to cluster and normalize textually derived entities, achieving an F1 score of 0.85. Furthermore, a labeled data set of polymer names was utilized to train our model and to infer the true total number of unique polymers that are actively reported in the literature. For ∼15,500 polymer named entities extracted from our corpus of 0.5 million papers, we detected 6734 unique clusters (i.e., unique polymers), 632 of which were manually curated to train the normalization model. This work will serve as a critical ingredient in a natural language processing-based pipeline for the automatic and efficient extraction of knowledge from the polymer literature.

Derivation and Internal Validation of the Ebola Prediction Score for Risk Stratification of Patients With Suspected Ebola Virus Disease.

STUDY OBJECTIVE: The current outbreak of Ebola virus disease in West Africa is the largest on record and has overwhelmed the capacity of local health systems and the international community to provide sufficient isolation and treatment of all suspected cases. The goal of this study is to develop a clinical prediction model that can help clinicians risk-stratify patients with suspected Ebola virus disease in the context of such an epidemic. METHODS: A retrospective analysis was performed of patient data collected during routine clinical care at the Bong County Ebola Treatment Unit in Liberia during its first 16 weeks of operation. The predictive power of 14 clinical and epidemiologic variables was measured against the primary outcome of laboratory-confirmed Ebola virus disease, using logistic regression to develop a final prediction model. Bootstrap sampling was used to assess the internal validity of the model and estimate its performance in a simulated validation cohort. RESULTS: Ebola virus disease testing results were available for 382 (97%) of 395 patients admitted to the Ebola treatment unit during the study period. A total of 160 patients (42%) tested positive for Ebola virus disease. Logistic regression analysis identified 6 variables independently predictive of laboratory-confirmed Ebola virus disease, including sick contact, diarrhea, loss of appetite, muscle pains, difficulty swallowing, and absence of abdominal pain. The Ebola Prediction Score, constructed with these 6 variables, had an area under the receiver operator characteristic curve of 0.75 (95% confidence interval 0.70 to 0.80) for the prediction of laboratory-confirmed Ebola virus disease. Patients with higher Ebola Prediction Scores had higher likelihoods of laboratory-confirmed Ebola virus disease. CONCLUSION: The Ebola Prediction Score can be used by clinicians as an adjunct to current Ebola virus disease case definitions to risk-stratify patients with suspected Ebola virus disease. Clinicians can use this new tool for the purpose of cohorting patients within the suspected-disease ward of an Ebola treatment unit or community-based isolation center to prevent nosocomial infection or as a triage tool when patient numbers overwhelm available capacity. Given the inherent limitations of clinical prediction models, however, a low-cost, point-of-care test that can rapidly and definitively exclude Ebola virus disease in patients should be a research priority.

Efficacy of emergency department calcium administration in cardiac arrest: A 9-year retrospective evaluation.

BACKGROUND: The efficacy of empiric calcium for patients with undifferentiated cardiac arrest has come under increased scrutiny, including a randomized controlled trial that was stopped early due to a trend towards harm with calcium administration. However, small sample sizes and non-significant findings have hindered precise effect estimates. In this analysis we evaluate the association of calcium administration with survival in a large retrospective cohort of patients with cardiac arrest treated in the emergency department (ED). METHODS: We conducted a retrospective review of medical records from two academic hospitals (one quaternary care center, one county trauma center) in San Francisco between 2011 and 2019. Inclusion criteria were patients aged greater than or equal to 18 years old who received treatment for cardiac arrest during their ED course. Our primary exposure was the administration of calcium while in the ED and the main outcome was survival to hospital admission. The association between calcium and survival to admission was estimated using a multivariable log-binomial regression, and also with two propensity score models. RESULTS: We examined 781 patients with cardiac arrest treated in San Francisco EDs between 2011 and 2019 and found that calcium administration was associated with decreased survival to hospital admission (RR 0.74; 95% CI 0.66-0.82). These findings remained significant after adjustment for patient age, sex, whether the cardiac arrest was witnessed, and including an interaction term for shockable cardiac rhythms (RR 0.60; 95% CI 0.50-0.72) and non-shockable cardiac rhythms (RR 0.87; 95% CI 0.76-0.99). Risk ratios for the association between calcium and survival to hospital admission were also similar between two propensity score-based models: nearest neighbor propensity matching model (RR 0.79; 95% CI 0.68-0.89) and inverse propensity weighted regression adjustment model (RR 0.75; 95% CI 0.67-0.84). CONCLUSIONS: Calcium administration as part of ED-directed treatment for cardiac arrest was associated with lower survival to hospital admission. Given the lack of statistically significant outcomes from smaller, more methodologically robust evaluations on this topic, we believe these findings have an important role to serve in confirming previous results and allowing for more precise effect estimates. Our data adds to the growing body evidence against the empiric use of calcium in cardiac arrest.

A general-purpose material property data extraction pipeline from large polymer corpora using natural language processing.

The ever-increasing number of materials science articles makes it hard to infer chemistry-structure-property relations from literature. We used natural language processing methods to automatically extract material property data from the abstracts of polymer literature. As a component of our pipeline, we trained MaterialsBERT, a language model, using 2.4 million materials science abstracts, which outperforms other baseline models in three out of five named entity recognition datasets. Using this pipeline, we obtained ~300,000 material property records from ~130,000 abstracts in 60 hours. The extracted data was analyzed for a diverse range of applications such as fuel cells, supercapacitors, and polymer solar cells to recover non-trivial insights. The data extracted through our pipeline is made available at polymerscholar.org which can be used to locate material property data recorded in abstracts. This work demonstrates the feasibility of an automatic pipeline that starts from published literature and ends with extracted material property information.

A survey of national physicians working in an active conflict zone: the challenges of emergency medical care in Iraq.

INTRODUCTION: There has been limited research on the perspectives and needs of national caregivers when confronted with large-scale societal violence. In Iraq, although the security situation has improved from its nadir in 2006-2007, intermittent bombings, and other hostilities continue. National workers remain the primary health resource for the affected populace. PROBLEM: To assess the status and challenges of national physicians working in the Emergency Departments of an active conflict area. METHODS: This study was a survey of civilian Iraqi doctors working in Emergency Departments (EDs) across Iraq, via a convenience sample of physicians taking the International Medical Corps (IMC) Doctor Course in Emergency Medicine, given in Baghdad from December 2008 through August 2009. RESULTS: The 148 physician respondents came from 11 provinces and over 50 hospitals in Iraq. They described cardiovascular disease, road traffic injuries, and blast and bullet injuries as the main causes of death and reasons for ED utilization. Eighty percent reported having been assaulted by a patient or their family member at least once within the last year; 38% reported they were threatened with a gun. Doctors reported seeing a median of 7.5 patients per hour, with only 19% indicating that their EDs had adequate physician staffing. Only 19% of respondents were aware of an established triage system for their hospital, and only a minority had taken courses covering ACLS- (16%) or ATLS-related (24%) material. Respondents reported a wide diversity of prior training, with only 3% having some type of specialized emergency medicine degree. CONCLUSIONS: The results of this study describe some of the challenges faced by national health workers providing emergency care to a violence-stricken populace. Study findings demonstrate high levels of violent behavior directed toward doctors in Iraqi Emergency Departments, as well as staffing shortages and a lack of formal training in emergency medical care.

Automated knowledge extraction from polymer literature using natural language processing.

Materials science literature has grown exponentially in recent years making it difficult for individuals to master all of this information. This constrains the formulation of new hypotheses that scientists can come up with. In this work, we explore whether materials science knowledge can be automatically inferred from textual information contained in journal papers. Using a data set of 0.5 million polymer papers, we show, using natural language processing methods that vector representations trained for every word in our corpus can indeed capture this knowledge in a completely unsupervised manner. We perform time-based studies through which we track popularity of various polymers for different applications and predict new polymers for novel applications based solely on the domain knowledge contained in our data set. Using co-relations detected automatically from literature in this manner thus, opens up a new paradigm for materials discovery.

Dielectric Polymers Tolerant to Electric Field and Temperature Extremes: Integration of Phenomenology, Informatics, and Experimental Validation.

Flexible polymer dielectrics tolerant to electric field and temperature extremes are urgently needed for a spectrum of electrical and electronic applications. Given the complexity of the dielectric breakdown mechanism and the vast chemical space of polymers, the discovery of suitable candidates is nontrivial. We have laid the foundation for a systematic search of the polymer chemical space, which starts with "gold-standard" experimental measurements and data on the temperature-dependent breakdown strength (Ebd) for a benchmark set of commercial dielectric polymer films. Phenomenological guidelines are derived from this data set on easily accessible properties (or "proxies") that are correlated with Ebd. Screening criteria based on these proxy properties (e.g., band gap, charge injection barrier, and cohesive energy density) and other necessary characteristics (e.g., a high glass transition temperature to maintain the thermal stability and a high dielectric constant for high energy density) were then setup. These criteria, along with machine learning models of these properties, were used to screen polymers candidates from a candidate list of more than 13 000 previously synthesized polymers, followed by experimental validation of some of the screened candidates. These efforts have led to the creation of a consistent and high-quality data set of temperature-dependent Ebd, and the identification of screening criteria, chemical design rules, and a list of optimal polymer candidates for high-temperature and high-energy-density capacitor applications, thus demonstrating the power of an integrated and informatics-based philosophy for rational materials design.

Surviving Ebola: A historical cohort study of Ebola mortality and survival in Sierra Leone 2014-2015.

BACKGROUND: While a number of predictors for Ebola mortality have been identified, less is known about post-viral symptoms. The identification of acute-illness predictors for post-viral symptoms could allow the selection of patients for more active follow up in the future, and those in whom early interventions may be beneficial in the long term. Studying predictors of both mortality and post-viral symptoms within a single cohort of patients could also further our understanding of the pathophysiology of survivor sequelae. METHODS/PRINCIPAL FINDINGS: We performed a historical cohort study using data collected as part of routine clinical care from an Ebola Treatment Centre (ETC) in Kerry Town, Sierra Leone, in order to identify predictors of mortality and of post-viral symptoms. Variables included as potential predictors were sex, age, date of admission, first recorded viral load at the ETC and symptoms (recorded upon presentation at the ETC). Multivariable logistic regression was used to identify predictors. Of 263 Ebola-confirmed patients admitted between November 2014 and March 2015, 151 (57%) survived to ETC discharge. Viral load was the strongest predictor of mortality (adjusted OR comparing high with low viral load: 84.97, 95% CI 30.87-345.94). We did not find evidence that a high viral load predicted post-viral symptoms (ocular: 1.17, 95% CI 0.35-3.97; musculoskeletal: 1.07, 95% CI 0.28-4.08). Ocular post-viral symptoms were more common in females (2.31, 95% CI 0.98-5.43) and in those who had experienced hiccups during the acute phase (4.73, 95% CI 0.90-24.73). CONCLUSIONS/SIGNIFICANCE: These findings may add epidemiological support to the hypothesis that post-viral symptoms have an immune-mediated aspect and may not only be a consequence of high viral load and disease severity.

Successful Implementation of a Multicountry Clinical Surveillance and Data Collection System for Ebola Virus Disease in West Africa: Findings and Lessons Learned.

BACKGROUND: The 2014 outbreak of Ebola virus disease (EVD) in West Africa was the largest ever recorded. Starting in September 2014, International Medical Corps (IMC) managed 5 Ebola treatment units (ETUs) in Liberia and Sierra Leone, which cumulatively cared for about 2,500 patients. We conducted a retrospective cohort study of patient data collected at the 5 ETUs over 1 year of operations. METHODS: To collect clinical and epidemiological data from the patient care areas, each chart was either manually copied across the fence between the high-risk zone and low-risk zone, imaged across the fence, or imaged in the high-risk zone. Each ETU's data were entered into a separate electronic database, and these were later combined into a single relational database. Lot quality assurance sampling was used to ensure data quality, with reentry of data with high error rates from imaged records. RESULTS: The IMC database contains records on 2,768 patient presentations, including 2,351 patient admissions with full follow-up data. Of the patients admitted, 470 (20.0%) tested positive for EVD, with an overall case fatality ratio (CFR) of 57.0% for EVD-positive patients and 8.1% for EVD-negative patients. Although more men were admitted than women (53.4% vs. 46.6%), a larger proportion of women were diagnosed EVD positive (25.6% vs. 15.2%). Diarrhea, red eyes, contact with an ill person, and funeral attendance were significantly more common in patients with EVD than in those with other diagnoses. Among EVD-positive patients, age was a significant predictor of mortality: the highest CFRs were among children under 5 (89.1%) and adults over 55 (71.4%). DISCUSSION: While several prior reports have documented the experiences of individual ETUs, this study is the first to present data from multiple ETUs across 2 countries run by the same organization with similar clinical protocols. Our experience demonstrates that even in austere settings under difficult conditions, it is possible for humanitarian organizations to collect high-quality clinical and epidemiologic data during a major infectious disease outbreak.

Whisker plucking alters responses of rat trigeminal ganglion neurons.

Whisker plucking in developing and adult rats provides a convenient method of temporarily altering tactile input for the purposes of studying experience-dependent plasticity in the somatosensory cortex. Yet, a comprehensive examination of the effect of whisker plucking on the response properties of whisker follicle-innervating trigeminal ganglion (NVg) neurons is lacking. We used extracellular single unit recordings to examine responses of NVg neurons to controlled whisker stimuli in three groups of animals: (1) rats whose whiskers were plucked from birth for 21 days; (2) rats whose whiskers were plucked once at 21 days of age; and (3) control animals. After at least 3 weeks of whisker re-growth, NVg neurons in plucked rats displayed normal, single whisker receptive fields and could be characterized as slowly (SA) or rapidly adapting (RA). The proportion of SA and RA neurons was unaffected by whisker plucking. Both SA and RA NVg neurons in plucked rats displayed normal response latencies and angular tuning but abnormally large responses to whisker movement onsets and offsets. SA neurons were affected to a greater extent than RA neurons. The effect of whisker plucking was more pronounced in animals whose whiskers were plucked repeatedly during development than in rats whose whiskers were plucked once. Individual neurons in plucked animals displayed abnormal periods of prolonged rhythmic firing following deflection onsets and aberrant bursts of activity during the plateau phase of the stimulus. These results indicate that whisker plucking exerts a long-term effect on responses of trigeminal ganglion neurons to peripheral stimulation.

Protracted development of responses to whisker deflection in rat trigeminal ganglion neurons.

The rodent whisker-to-barrel pathway constitutes a major model system for studying experience-dependent brain development. Yet little is known about responses of neurons to whisker stimulation in young animals. Response properties of trigeminal ganglion (NV) neurons in 2-, 3-, and 4-week-old and adult rats were examined using extracellular single-unit recordings and controlled whisker stimuli. We found that the receptive field size of NV neurons is mature in 2-week-old animals while response latencies, magnitudes, and angular tuning continue to develop between 2 weeks of age and adulthood. At the earliest time recorded, NV neurons respond to stimulation of only one whisker and can be characterized as slowly or rapidly adapting (SA, RA). The proportion of SA and RA neurons remains constant during development. Consistent with known on-going myelination of NV axons, response latencies decrease with age, becoming adult-like during the third and fourth postnatal weeks for RA and SA neurons, respectively. Unexpectedly, we found that evoked response magnitudes increase several-fold during development becoming adult-like only during the fourth postnatal week. In addition, RA neurons become less selective for whisker deflection angle with age. Maturation of response magnitude and angular tuning is consistent with developmental changes in the mechanical properties of the whisker, the whisker follicle, and the surrounding tissues. The findings indicate that whisker-derived tactile inputs mature during the first postnatal month when whisker-related cortical circuits are susceptible to long-term modification by sensory experience. Thus normal developmental changes in sensory input may influence functional development of cortical circuits.