Comparing Medical Term Usage Patterns of Professionals and Search Engine and Community Question Answering Service Users in Japan: Log Analysis.

BACKGROUND: Despite increasing opportunities for acquiring health information online, discussion of the specific words used in searches has been limited. OBJECTIVE: The aim of this study was to clarify the medical information gap between medical professionals and the general public in Japan through health information-seeking activities on the internet. METHODS: Search and posting data were analyzed from one of the most popular domestic search engines in Japan (Yahoo! JAPAN Search) and the most popular Japanese community question answering service (Yahoo! Chiebukuro). We compared the frequency of 100 clinical words appearing in the clinical case reports of medical professionals (clinical frequency) with their frequency in Yahoo! JAPAN Search (search frequency) logs and questions posted to Yahoo! Chiebukuro (question frequency). The Spearman correlation coefficient was used to quantify association patterns among the three information sources. Additionally, user information (gender and age) in the search frequency associated with each registered user was extracted. RESULTS: Significant correlations were observed between clinical and search frequencies (r=0.29, P=.003), clinical and question frequencies (r=0.34, P=.001), and search and question frequencies (r=0.57, P<.001). Low-frequency words in clinical frequency (eg, "hypothyroidism," "ulcerative colitis") highly ranked in search frequency. Similarly, "pain," "slight fever," and "numbness" were highly ranked only in question frequency. The weighted average of ages was 34.5 (SD 2.7) years, and the weighted average of gender (man -1, woman +1) was 0.1 (SD 0.1) in search frequency. Some words were specifically extracted from the search frequency of certain age groups, including "abdominal pain" (10-20 years), "plasma cells" and "inflammatory findings" (20-30 years), "DM" (diabetes mellitus; 30-40 years), "abnormal shadow" and "inflammatory findings" (40-50 years), "hypertension" and "abnormal shadow" (50-60 years), and "lung cancer" and "gastric cancer" (60-70 years). CONCLUSIONS: Search and question frequencies showed similar tendencies, whereas search and clinical frequencies showed discrepancy. Low-clinical frequency words related to diseases such as "hypothyroidism" and "ulcerative colitis" had high search frequencies, whereas those related to symptoms such as "pain," "slight fever," and "numbness" had high question frequencies. Moreover, high search frequency words included designated intractable diseases such as "ulcerative colitis," which has an incidence of less than 0.1% in the Japanese population. Therefore, it is generally worthwhile to pay attention not only to major diseases but also to minor diseases that users frequently seek information on, and more words will need to be analyzed in the future. Some characteristic words for certain age groups were observed (eg, 20-40 years: "cancer"; 40-60 years: diagnoses and diseases identified in health examinations; 60-70 years: diseases with late adulthood onset and "death"). Overall, this analysis demonstrates that medical professionals as information providers should be aware of clinical frequency, and medical information gaps between professionals and the general public should be bridged.

The CLASSE GATOR (CLinical Acronym SenSE disambiGuATOR): A Method for predicting acronym sense from neonatal clinical notes.

OBJECTIVE: To develop an algorithm for identifying acronym 'sense' from clinical notes without requiring a clinically annotated training set. MATERIALS AND METHODS: Our algorithm is called CLASSE GATOR: Clinical Acronym SenSE disambiGuATOR. CLASSE GATOR extracts acronyms and definitions from PubMed Central (PMC). A logistic regression model is trained using words associated with specific acronym-definition pairs from PMC. CLASSE GATOR uses this library of acronym-definitions and their corresponding word feature vectors to predict the acronym 'sense' from Beth Israel Deaconess (MIMIC-III) neonatal notes. RESULTS: We identified 1,257 acronyms and 8,287 definitions including a random definition from 31,764 PMC articles on prenatal exposures and 2,227,674 PMC open access articles. The average number of senses (definitions) per acronym was 6.6 (min = 2, max = 50). The average internal 5-fold cross validation was 87.9 % (on PMC). We found 727 unique acronyms (57.29 %) from PMC were present in 105,044 neonatal notes (MIMIC-III). We evaluated the performance of acronym prediction using 245 manually annotated clinical notes with 9 distinct acronyms. CLASSE GATOR achieved an overall accuracy of 63.04 % and outperformed random for 8/9 acronyms (88.89 %) when applied to clinical notes. We also compared our algorithm with UMN's acronym set, and found that CLASSE GATOR outperformed random for 63.46 % of 52 acronyms when using logistic regression, 75.00 % when using Bert and 76.92 % when using BioBert as the prediction algorithm within CLASSE GATOR. CONCLUSIONS: CLASSE GATOR is the first automated acronym sense disambiguation method for clinical notes. Importantly, CLASSE GATOR does not require an expensive manually annotated acronym-definition corpus for training.

Mapping the knowledge structure and trends of epilepsy genetics over the past decade: A co-word analysis based on medical subject headings terms.

INTRODUCTION: Over the past 10 years, epilepsy genetics has made dramatic progress. This study aimed to analyze the knowledge structure and the advancement of epilepsy genetics over the past decade based on co-word analysis of medical subject headings (MeSH) terms. METHODS: Scientific publications focusing on epilepsy genetics from the PubMed database (January 2009-December 2018) were retrieved. Bibliometric information was analyzed quantitatively using Bibliographic Item Co-Occurrence Matrix Builder (BICOMB) software. A knowledge social network analysis and publication trend based on the high-frequency MeSH terms was built using VOSviewer. RESULTS: According to the search strategy, a total of 5185 papers were included. Among all the extracted MeSH terms, 86 high-frequency MeSH terms were identified. Hot spots were clustered into 5 categories including: "ion channel diseases," "beyond ion channel diseases," "experimental research & epigenetics," "single nucleotide polymorphism & pharmacogenetics," and "genetic techniques". "Epilepsy," "mutation," and "seizures," were located at the center of the knowledge network. "Ion channel diseases" are typically in the most prominent position of epilepsy genetics research. "Beyond ion channel diseases" and "genetic techniques," however, have gradually grown into research cores and trends, such as "intellectual disability," "infantile spasms," "phenotype," "exome," " deoxyribonucleic acid (DNA) copy number variations," and "application of next-generation sequencing." While ion channel genes such as "SCN1A," "KCNQ2," "SCN2A," "SCN8A" accounted for nearly half of epilepsy genes in MeSH terms, a number of additional beyond ion channel genes like "CDKL5," "STXBP1," "PCDH19," "PRRT2," "LGI1," "ALDH7A1," "MECP2," "EPM2A," "ARX," "SLC2A1," and more were becoming increasingly popular. In contrast, gene therapies, treatment outcome, and genotype-phenotype correlations were still in their early stages of research. CONCLUSION: This co-word analysis provides an overview of epilepsy genetics research over the past decade. The 5 research categories display publication hot spots and trends in epilepsy genetics research which could consequently supply some direction for geneticists and epileptologists when launching new projects.

Search results outliers among MEDLINE platforms.

OBJECTIVE: Hypothetically, content in MEDLINE records is consistent across multiple platforms. Though platforms have different interfaces and requirements for query syntax, results should be similar when the syntax is controlled for across the platforms. The authors investigated how search result counts varied when searching records among five MEDLINE platforms. METHODS: We created 29 sets of search queries targeting various metadata fields and operators. Within search sets, we adapted 5 distinct, compatible queries to search 5 MEDLINE platforms (PubMed, ProQuest, EBSCOhost, Web of Science, and Ovid), totaling 145 final queries. The 5 queries were designed to be logically and semantically equivalent and were modified only to match platform syntax requirements. We analyzed the result counts and compared PubMed's MEDLINE result counts to result counts from the other platforms. We identified outliers by measuring the result count deviations using modified z-scores centered around PubMed's MEDLINE results. RESULTS: Web of Science and ProQuest searches were the most likely to deviate from the equivalent PubMed searches. EBSCOhost and Ovid were less likely to deviate from PubMed searches. Ovid's results were the most consistent with PubMed's but appeared to apply an indexing algorithm that resulted in lower retrieval sets among equivalent searches in PubMed. Web of Science exhibited problems with exploding or not exploding Medical Subject Headings (MeSH) terms. CONCLUSION: Platform enhancements among interfaces affect record retrieval and challenge the expectation that MEDLINE platforms should, by default, be treated as MEDLINE. Substantial inconsistencies in search result counts, as demonstrated here, should raise concerns about the impact of platform-specific influences on search results.

A MeSH-based text mining method for identifying novel prebiotics.

Prebiotics contribute to the well-being of their host by altering the composition of the gut microbiota. Discovering new prebiotics is a challenging and arduous task due to strict inclusion criteria; thus, highly limited numbers of prebiotic candidates have been identified. Notably, the large numbers of published studies may contain substantial information attached to various features of known prebiotics that can be used to predict new candidates. In this paper, we propose a medical subject headings (MeSH)-based text mining method for identifying new prebiotics with structured texts obtained from PubMed. We defined an optimal feature set for prebiotics prediction using a systematic feature-ranking algorithm with which a variety of carbohydrates can be accurately classified into different clusters in accordance with their chemical and biological attributes. The optimal feature set was used to separate positive prebiotics from other carbohydrates, and a cross-validation procedure was employed to assess the prediction accuracy of the model. Our method achieved a specificity of 0.876 and a sensitivity of 0.838. Finally, we identified a high-confidence list of candidates of prebiotics that are strongly supported by the literature. Our study demonstrates that text mining from high-volume biomedical literature is a promising approach in searching for new prebiotics.

Use of scientific social networking to improve the research strategies of PubMed readers.

BACKGROUND: Keeping up with journal articles on a daily basis is an important activity of scientists engaged in biomedical research. Usually, journal articles and papers in the field of biomedicine are accessed through the Medline/PubMed electronic library. In the process of navigating PubMed, researchers unknowingly generate user-specific reading profiles that can be shared within a social networking environment. This paper examines the structure of the social networking environment generated by PubMed users. METHODS: A web browser plugin was developed to map [in Medical Subject Headings (MeSH) terms] the reading patterns of individual PubMed users. RESULTS: We developed a scientific social network based on the personal research profiles of readers of biomedical articles. A browser plugin is used to record the digital object identifier or PubMed ID of web pages. Recorded items are posted on the activity feed and automatically mapped to PubMed abstract. Within the activity feed a user can trace back previously browsed articles and insert comments. By calculating the frequency with which specific MeSH occur, the research interests of PubMed users can be visually represented with a tag cloud. Finally, research profiles can be searched for matches between network users. CONCLUSIONS: A social networking environment was created using MeSH terms to map articles accessed through the Medline/PubMed online library system. In-network social communication is supported by the recommendation of articles and by matching users with similar scientific interests. The system is available at http://bioknol.org/en/.

pubmed.mineR: an R package with text-mining algorithms to analyse PubMed abstracts.

The PubMed literature database is a valuable source of information for scientific research. It is rich in biomedical literature with more than 24 million citations. Data-mining of voluminous literature is a challenging task. Although several text-mining algorithms have been developed in recent years with focus on data visualization, they have limitations such as speed, are rigid and are not available in the open source. We have developed an R package, pubmed.mineR, wherein we have combined the advantages of existing algorithms, overcome their limitations, and offer user flexibility and link with other packages in Bioconductor and the Comprehensive R Network (CRAN) in order to expand the user capabilities for executing multifaceted approaches. Three case studies are presented, namely, 'Evolving role of diabetes educators', 'Cancer risk assessment' and 'Dynamic concepts on disease and comorbidity' to illustrate the use of pubmed.mineR. The package generally runs fast with small elapsed times in regular workstations even on large corpus sizes and with compute intensive functions. The pubmed.mineR is available at http://cran.rproject. org/web/packages/pubmed.mineR.

The support of human genetic evidence for approved drug indications.

Over a quarter of drugs that enter clinical development fail because they are ineffective. Growing insight into genes that influence human disease may affect how drug targets and indications are selected. However, there is little guidance about how much weight should be given to genetic evidence in making these key decisions. To answer this question, we investigated how well the current archive of genetic evidence predicts drug mechanisms. We found that, among well-studied indications, the proportion of drug mechanisms with direct genetic support increases significantly across the drug development pipeline, from 2.0% at the preclinical stage to 8.2% among mechanisms for approved drugs, and varies dramatically among disease areas. We estimate that selecting genetically supported targets could double the success rate in clinical development. Therefore, using the growing wealth of human genetic data to select the best targets and indications should have a measurable impact on the successful development of new drugs.

Medical Subject Headings (MeSH) for indexing and retrieving open-source healthcare data.

The US federal government initiated the Open Government Directive where federal agencies are required to publish high value datasets so that they are available to the public. Data.gov and the community site Healthdata.gov were initiated to disperse such datasets. However, data searches and retrieval for these sites are keyword driven and severely limited in performance. The purpose of this paper is to address the issue of extracting relevant open-source data by proposing a method of adopting the MeSH framework for indexing and data retrieval. A pilot study was conducted to compare the performance of traditional keywords to MeSH terms for retrieving relevant open-source datasets related to "mortality". The MeSH framework resulted in greater sensitivity with comparable specificity to the keyword search. MeSH showed promise as a method for indexing and retrieving data, yet future research should conduct a larger scale evaluation of the performance of the MeSH framework for retrieving relevant open-source healthcare datasets.

Mapping the structure and dynamics of genomics-related MeSH terms complex networks.

It has been proposed that the history and evolution of scientific ideas may reflect certain aspects of the underlying socio-cognitive frameworks in which science itself is developing. Systematic analyses of the development of scientific knowledge may help us to construct models of the collective dynamics of science. Aiming at scientific rigor, these models should be built upon solid empirical evidence, analyzed with formal tools leading to ever-improving results that support the related conclusions. Along these lines we studied the dynamics and structure of the development of research in genomics as represented by the entire collection of genomics-related scientific papers contained in the PubMed database. The analyzed corpus consisted in more than 49,000 articles published in the years 1987 (first appearance of the term Genomics) to 2011, categorized by means of the Medical Subheadings (MeSH) content-descriptors. Complex networks were built where two MeSH terms were connected if they are descriptors of the same article(s). The analysis of such networks revealed a complex structure and dynamics that to certain extent resembled small-world networks. The evolution of such networks in time reflected interesting phenomena in the historical development of genomic research, including what seems to be a phase-transition in a period marked by the completion of the first draft of the Human Genome Project. We also found that different disciplinary areas have different dynamic evolution patterns in their MeSH connectivity networks. In the case of areas related to science, changes in topology were somewhat fast while retaining a certain core-structure, whereas in the humanities, the evolution was pretty slow and the structure resulted highly redundant and in the case of technology related issues, the evolution was very fast and the structure remained tree-like with almost no overlapping terms.

Evaluating alignment quality between iconic language and reference terminologies using similarity metrics.

BACKGROUND: Visualization of Concepts in Medicine (VCM) is a compositional iconic language that aims to ease information retrieval in Electronic Health Records (EHR), clinical guidelines or other medical documents. Using VCM language in medical applications requires alignment with medical reference terminologies. Alignment from Medical Subject Headings (MeSH) thesaurus and International Classification of Diseases - tenth revision (ICD10) to VCM are presented here. This study aim was to evaluate alignment quality between VCM and other terminologies using different measures of inter-alignment agreement before integration in EHR. METHODS: For medical literature retrieval purposes and EHR browsing, the MeSH thesaurus and the ICD10, both organized hierarchically, were aligned to VCM language. Some MeSH to VCM alignments were performed automatically but others were performed manually and validated. ICD10 to VCM alignment was entirely manually performed. Inter-alignment agreement was assessed on ICD10 codes and MeSH descriptors, sharing the same Concept Unique Identifiers in the Unified Medical Language System (UMLS). Three metrics were used to compare two VCM icons: binary comparison, crude Dice Similarity Coefficient (DSCcrude), and semantic Dice Similarity Coefficient (DSCsemantic), based on Lin similarity. An analysis of discrepancies was performed. RESULTS: MeSH to VCM alignment resulted in 10,783 relations: 1,830 of which were manually performed and 8,953 were automatically inherited. ICD10 to VCM alignment led to 19,852 relations. UMLS gathered 1,887 alignments between ICD10 and MeSH. Only 1,606 of them were used for this study. Inter-alignment agreement using only validated MeSH to VCM alignment was 74.2% [70.5-78.0]CI95%, DSCcrude was 0.93 [0.91-0.94]CI95%, and DSCsemantic was 0.96 [0.95-0.96]CI95%. Discrepancy analysis revealed that even if two thirds of errors came from the reviewers, UMLS was nevertheless responsible for one third. CONCLUSIONS: This study has shown strong overall inter-alignment agreement between MeSH to VCM and ICD10 to VCM manual alignments. VCM icons have now been integrated into a guideline search engine (http://www.cismef.org) and a health terminologies portal (http://www.hetop.eu).

SIDD: a semantically integrated database towards a global view of human disease.

BACKGROUND: A number of databases have been developed to collect disease-related molecular, phenotypic and environmental features (DR-MPEs), such as genes, non-coding RNAs, genetic variations, drugs, phenotypes and environmental factors. However, each of current databases focused on only one or two DR-MPEs. There is an urgent demand to develop an integrated database, which can establish semantic associations among disease-related databases and link them to provide a global view of human disease at the biological level. This database, once developed, will facilitate researchers to query various DR-MPEs through disease, and investigate disease mechanisms from different types of data. METHODOLOGY: To establish an integrated disease-associated database, disease vocabularies used in different databases are mapped to Disease Ontology (DO) through semantic match. 4,284 and 4,186 disease terms from Medical Subject Headings (MeSH) and Online Mendelian Inheritance in Man (OMIM) respectively are mapped to DO. Then, the relationships between DR-MPEs and diseases are extracted and merged from different source databases for reducing the data redundancy. CONCLUSIONS: A semantically integrated disease-associated database (SIDD) is developed, which integrates 18 disease-associated databases, for researchers to browse multiple types of DR-MPEs in a view. A web interface allows easy navigation for querying information through browsing a disease ontology tree or searching a disease term. Furthermore, a network visualization tool using Cytoscape Web plugin has been implemented in SIDD. It enhances the SIDD usage when viewing the relationships between diseases and DR-MPEs. The current version of SIDD (Jul 2013) documents 4,465,131 entries relating to 139,365 DR-MPEs, and to 3,824 human diseases. The database can be freely accessed from: http://mlg.hit.edu.cn/SIDD.

A study of the influence of task familiarity on user behaviors and performance with a MeSH term suggestion interface for PubMed bibliographic search.

PURPOSE: Previous research has shown that information seekers in biomedical domain need more support in formulating their queries. A user study was conducted to evaluate the effectiveness of a metadata based query suggestion interface for PubMed bibliographic search. The study also investigated the impact of search task familiarity on search behaviors and the effectiveness of the interface. METHODS: A real user, user search request and real system approach was used for the study. Unlike tradition IR evaluation, where assigned tasks were used, the participants were asked to search requests of their own. Forty-four researchers in Health Sciences participated in the evaluation - each conducted two research requests of their own, alternately with the proposed interface and the PubMed baseline. Several performance criteria were measured to assess the potential benefits of the experimental interface, including users' assessment of their original and eventual queries, the perceived usefulness of the interfaces, satisfaction with the search results, and the average relevance score of the saved records. RESULTS: The results show that, when searching for an unfamiliar topic, users were more likely to change their queries, indicating the effect of familiarity on search behaviors. The results also show that the interface scored higher on several of the performance criteria, such as the "goodness" of the queries, perceived usefulness, and user satisfaction. Furthermore, in line with our hypothesis, the proposed interface was relatively more effective when less familiar search requests were attempted. CONCLUSIONS: Results indicate that there is a selective compatibility between search familiarity and search interface. One implication of the research for system evaluation is the importance of taking into consideration task familiarity when assessing the effectiveness of interactive IR systems.

Improving information retrieval using Medical Subject Headings Concepts: a test case on rare and chronic diseases.

BACKGROUND: As more scientific work is published, it is important to improve access to the biomedical literature. Since 2000, when Medical Subject Headings (MeSH) Concepts were introduced, the MeSH Thesaurus has been concept based. Nevertheless, information retrieval is still performed at the MeSH Descriptor or Supplementary Concept level. OBJECTIVE: The study assesses the benefit of using MeSH Concepts for indexing and information retrieval. METHODS: Three sets of queries were built for thirty-two rare diseases and twenty-two chronic diseases: (1) using PubMed Automatic Term Mapping (ATM), (2) using Catalog and Index of French-language Health Internet (CISMeF) ATM, and (3) extrapolating the MEDLINE citations that should be indexed with a MeSH Concept. RESULTS: Type 3 queries retrieve significantly fewer results than type 1 or type 2 queries (about 18,000 citations versus 200,000 for rare diseases; about 300,000 citations versus 2,000,000 for chronic diseases). CISMeF ATM also provides better precision than PubMed ATM for both disease categories. DISCUSSION: Using MeSH Concept indexing instead of ATM is theoretically possible to improve retrieval performance with the current indexing policy. However, using MeSH Concept information retrieval and indexing rules would be a fundamentally better approach. These modifications have already been implemented in the CISMeF search engine.

Evaluation of the key words used in articles of the Acta Cirurgica Brasileira from 1997 to 2012.

PURPOSE: To evaluate the key words used in Acta Cirurgica Brasileira from 1997 to 2012. METHODS: All the key words of all articles published in regular issues between 1997 and 2012 were analyzed, ensuring that these key words were in the MeSH database (Medical Subjects Headings) and the most used subject headings and most wrong repeated key words were ranked. RESULTS: > 4230 key words used in 990 articles were analyzed. Only 579 key words (13.68%) were not in the MeSH database, considering that there was a statistically significant decrease over the years (p<0.001). The three most used key words were Rats, Dogs and Wound healing. Among the wrong ones, the key words were Adhesions, Experimental surgery and Anatomosis. CONCLUSION: There was a gradual improvement in the amount of key words used that belonged to the MeSH database, and there were 618 articles (62.42%) with all key words correct.