bims-librar 2018-07-15 papers

Issue of 2018–07–15
six papers selected by
Thomas Krichel, Open Library Society

J Clin Epidemiol. 2018 Jul 05. pii: S0895-4356(18)30085-4. [Epub ahead of print]

Extending PubMed Searches to ClinicalTrials.gov Through a Machine Learning Approach for Systematic Reviews.

Corrado Lanera, Clara Minto, Abhinav Sharma, Dario Gregori, Paola Berchialla, Ileana Baldi.

AIMS: Despite their essential role in collecting and organizing published medical literature, indexed search engines are unable to cover all relevant knowledge. Hence, current literature recommends the inclusion of clinical trial registries in systematic reviews. This study aims to provide an automated approach to extend a search on PubMed to the ClinicalTrials.gov database, relying on text mining and machine learning techniques.
STUDY DESIGN AND SETTING: The procedure starts from a literature search on PubMed. Next, it considers the training of a classifier that can identify documents with a comparable word characterization in the ClinicalTrials.gov clinical trial repository. Fourteen systematic reviews, covering a broad range of health conditions, are used as case studies for external validation. A cross-validated support-vector machine model was used as the classifier.
RESULTS: The sensitivity was 100% in all systematic reviews except one (87.5%), and the specificity ranged from 97.2 to 99.9%. The ability of the instrument to distinguish on-topic from off-topic articles ranged from an AUC of 93.4 to 99.9%.
CONCLUSION: The proposed machine learning instrument has the potential to help researchers identify relevant studies in the systematic review process by reducing workload, without losing sensitivity and at a small price in terms of specificity.

Keywords: Clinical Trial Registry; Indexed Search Engine; Machine Learning; Meta-Analysis; Systematic Review; Text Mining

DOI: https://doi.org/10.1016/j.jclinepi.2018.06.015

Transl Behav Med. 2018 Jul 04.

Bibliometric analysis of literature in AIDS-related stigma and discrimination.

Waleed M Sweileh.

Stigma and discrimination are major barriers in the global fight against human immunodeficiency virus (HIV) and acquired immunodeficiency syndrome (AIDS). The aim of this study was to create an analytical inventory of worldwide research output in AIDS-related stigma and discrimination. SciVerse Scopus was used for the study period from 1980 to 2017 to retrieve literature in AIDS-related stigma and discrimination. Results were presented as bibliometric tables and maps. In total, 2,509 documents were retrieved. Approximately 40% (n = 990) of the retrieved documents were published in the last 5 years (2013-2017). Retrieved documents received an average of 19.8 citations per article and had an average of 3.2 authors per article. The Hirsh index of the retrieved documents was 94. Most frequently encountered topics were mental health, adherence, adolescents, women, disclosure, and Africa. The USA contributed to 1,226 (48.9%) documents while the African region contributed to 531 (21.2%) documents. Research collaboration among most active countries was relatively low. Authors and institutions from the USA dominated this field. AIDS Care was the most active journal in publishing documents in this field with 307 (13.4%) documents while documents published in Social Medicine journal received the highest citations. Research in AIDS-related stigma and discrimination had witnessed a noticeable increase in the past decade, but the overall number of publications is considered insignificant relative to the size of the problem and the global number of infected people. There was a relative underpresentation of literature from African region despite the fact that more than two-thirds of HIV-infected people in the world are living in Africa.

DOI: https://doi.org/10.1093/tbm/iby072

Scand J Work Environ Health. 2018 Jul 08. pii: 3750. [Epub ahead of print]

Trends and topics in occupational diseases over the last 60 years from PubMed.

Jean-Francois Gehanno, Aurore Postel, Mathieu Schuers, Laetitia Rollin.

Objectives The purpose of this study was to provide an analysis of scientific production on occupational diseases (OD) during the period 1945-2015 in order to describe publication trends on that topic and identify the major diseases as well as the predominant actors (journals, countries) involved in this field. Methods A PubMed search was carried out to extract articles related to occupational diseases during the period 1 January 1945 to 31 December 2015 using a specific query. Data were downloaded from PubMed in Extensible Markup Language (XML) and processed through a dedicated parser. Results A total of 160 025 articles were retrieved from 7127 journals. One third of these articles were published in 39 journals: the core journals according to Bradford's law. Following exponential growth, OD publications reached a plateau in 2007. The overall dynamics of the OD field are heterogeneous with differences between subfields: psychological diseases emerged in the 1990s while "traditional" OD are less studied nowadays. Despite a sharp decrease in the proportion of publications, the most productive country remains the USA with 14.5% of the OD publications over the period but Scandinavian countries are, proportionally, the most active in research and publication on OD. Conclusions The proportion of publications on OD is decreasing in Medline, except for specific subfields of OD. This is discrepant with the global burden of occupational diseases.

DOI: https://doi.org/10.5271/sjweh.3750

Database (Oxford). 2018 Jan 01. 2018

Improved biomedical term selection in pseudo relevance feedback.

Muhammad Nabeel Asim, Muhammad Wasim, Muhammad Usman Ghani Khan, Waqar Mahmood.

Biomedical information retrieval systems are becoming popular and complex due to massive amount of ever-growing biomedical literature. Users are unable to construct a precise and accurate query that represents the intended information in a clear manner. Therefore, query is expanded with the terms or features that retrieve more relevant information. Selection of appropriate expansion terms plays key role to improve the performance of retrieval task. We propose document frequency chi-square, a newer version of chi-square in pseudo relevance feedback for term selection. The effects of pre-processing on the performance of information retrieval specifically in biomedical domain are also depicted. On average, the proposed algorithm outperformed state-of-the-art term selection algorithms by 88% at pre-defined test points. Our experiments also conclude that, stemming cause a decrease in overall performance of the pseudo relevance feedback based information retrieval system particularly in biomedical domain.Database URL: http://biodb.sdau.edu.cn/gan/.

DOI: https://doi.org/10.1093/database/bay056

J Biomed Inform. 2018 Jul 04. pii: S1532-0464(18)30128-X. [Epub ahead of print]

Relevance Feedback for Enhancing Content Based Image Retrieval and Automatic Prediction of Semantic Image Features: Application to Bone Tumor Radiographs.

Imon Banerjee, Camille Kurtz, Alon Edward Devorah, Bao Do, Daniel L Rubin, Christopher F Beaulieu.

BACKGROUND: The majority of current medical CBIR systems perform retrieval based only on "imaging signatures" generated by extracting pixel-level quantitative features, and only rarely has a feedback mechanism been incorporated to improve retrieval performance. In addition, current medical CBIR approaches do not routinely incorporate semantic terms that model the user's high-level expectations, and this can limit CBIR performance.
METHOD: We propose a retrieval framework that exploits a hybrid feature space (HFS) that is built by integrating low-level image features and high-level semantic terms, through rounds of relevance feedback (RF) and performs similarity-based retrieval to support semi-automatic image interpretation. The novelty of the proposed system is that it can impute the semantic features of the query image by reformulating the query vector representation in the HFS via user feedback. We implemented our framework as a prototype that performs the retrieval over a database of 811 radiographic images that contains 69 unique types of bone tumors.
RESULTS: We evaluated the system performance by conducting independent reading sessions with two subspecialist musculoskeletal radiologists. For the test set, the proposed retrieval system at fourth RF iteration of the sessions conducted with both the radiologists achieved mean average precision (MAP) value ∼ 0.90 where the initial MAP with baseline CBIR was 0.20. In addition, we also achieved high prediction accuracy (>0.8) for the majority of the semantic features automatically predicted by the system.
CONCLUSION: Our proposed framework addresses some limitations of existing CBIR systems by incorporating user feedback and simultaneously predicting the semantic features of the query image. This obviates the need for the user to provide those terms and makes CBIR search more efficient for inexperience users/trainees. Encouraging results achieved in the current study highlight possible new directions in radiological image interpretation employing semantic CBIR combined with relevance feedback of visual similarity.

Keywords: bone tumors; content based image retrieval; pixel-level features; radiography; radiomics; relevance feedback; semantic features

DOI: https://doi.org/10.1016/j.jbi.2018.07.002

J Natl Cancer Inst. 2018 Jun 29.

A Bibliometric Analysis of the Landscape of Cancer Rehabilitation Research (1992-2016).

Nicole L Stout, Catherine M Alfano, Christopher W Belter, Ralph Nitkin, Alison Cernich, Karen Lohmann Siegel, Leighton Chan.

Cancer rehabilitation research has accelerated as great attention has focused on improving survivorship care. Recent expert consensus has attempted to prioritize research needs and suggests greater focus on studying physical functioning of survivors. However, no analysis of the publication landscape has substantiated these proposed needs. This manuscript provides an analysis of PubMed indexed articles related to cancer rehabilitation published between 1992 and 2017. A total of 22 171 publications were analyzed using machine learning and text analysis to assess publication metrics, topic areas of emphasis, and their interrelationships through topic similarity networks. Publications have increased at a rate of 136 articles per year. Approximately 10% of publications were funded by the National Institutes of Health institutes and centers, with the National Cancer Institute being the most prominent funder. The greatest volume and rate of publication increase were in the topics of Cognitive and Behavioral Therapies and Psychological Interventions, followed by Depression and Exercise Therapy. Four research topic similarity networks were identified and provide insight on areas of robust publication and notable deficits. Findings suggest that publication emphasis has strongly supported cognitive, behavioral, and psychological therapies; however, studies of functional morbidity and physical rehabilitation research are lacking. Three areas of publication deficits are noted: research on populations outside of breast, prostate, and lung cancers; methods for integrating physical rehabilitation services with cancer care, specifically regarding functional screening and assessment; and physical rehabilitation interventions. These deficits align with the needs identified by expert consensus and support the supposition that future research should emphasize a focus on physical rehabilitation.

DOI: https://doi.org/10.1093/jnci/djy108