Characterization of the mechanism of drug-drug interactions from PubMed using MeSH terms

MeSH2Matrix: combining MeSH keywords and machine learning for biomedical relation classification based on PubMed

Biomedical relation classification has been significantly improved by the application of advanced machine learning techniques on the raw texts of scholarly publications. Despite this improvement, the reliance on large chunks of raw text makes these algorithms suffer in terms of generalization, precision, and reliability. The use of the distinctive characteristics of bibliographic metadata can prove effective in achieving better performance for this challenging task. In this research paper, we introduce an approach for biomedical relation classification using the qualifiers of co-occurring Medical Subject Headings (MeSH). First of all, we introduce MeSH2Matrix, our dataset consisting of 46,469 biomedical relations curated from PubMed publications using our approach. Our dataset includes a matrix that maps associations between the qualifiers of subject MeSH keywords and those of object MeSH keywords. It also specifies the corresponding Wikidata relation type and the superclass of semantic relations for each relation. Using MeSH2Matrix, we build and train three machine learning models (Support Vector Machine [SVM], a dense model [D-Model], and a convolutional neural network [C-Net]) to evaluate the efficiency of our approach for biomedical relation classification. Our best model achieves an accuracy of 70.78% for 195 classes and 83.09% for five superclasses. Finally, we provide confusion matrix and extensive feature analyses to better examine the relationship between the MeSH qualifiers and the biomedical relations being classified. Our results will hopefully shed light on developing better algorithms for biomedical ontology classification based on the MeSH keywords of PubMed publications. For reproducibility purposes, MeSH2Matrix, as well as all our source codes, are made publicly accessible at https://github.com/SisonkeBiotik-Africa/MeSH2Matrix .

Global epidemiology of breast cancer based on risk factors: a systematic review

Background

Numerous reviews of the epidemiology and risk factors for breast cancer have been published previously which heighted different directions of breast cancer.

Aim

The present review examined the likelihood that incidence, prevalence, and particular risk factors might vary by geographic region and possibly by food and cultural practices as well.

Methods

A systematic review (2017-2022) was conducted following Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines, reporting on epidemiological and risk factor reports from different world regions. Medical Subject Heading (MeSH) terms: "Breast neoplasm" "AND" country terms such as "Pakistan/epidemiology", "India/epidemiology", "North America/epidemiology", "South Africa/epidemiology" were used to retrieve 2068 articles from PubMed. After applying inclusion and exclusion terms, 49 papers were selected for systematic review.

Results

Results of selected articles were summarized based on risk factors, world regions and study type. Risk factors were classified into five categories: demographic, genetic and lifestyle risk factors varied among countries. This review article covers a variety of topics, including regions, main findings, and associated risk factors such as genetic factors, and lifestyle. Several studies revealed that lifestyle choices including diet and exercise could affect a person's chance of developing breast cancer. Breast cancer risk has also been linked to genetic variables, including DNA repair gene polymorphisms and mutations in the breast cancer gene (BRCA). It has been found that most of the genetic variability links to the population of Asia while the cause of breast cancer due to lifestyle modifications has been found in American and British people, indicating that demographic, genetic, and, lifestyle risk factors varied among countries.

Conclusion

There are many risk factors for breast cancer, which vary in their importance depending on the world region. However, further investigation is required to better comprehend the particular causes of breast cancer in these areas as well as to create efficient prevention and treatment plans that cater to the local population.

The association between warfarin usage and international normalized ratio increase: systematic analysis of FDA Adverse Event Reporting System (FAERS)

Introduction

Elevated international normalized ratio (INR) has been commonly reported as an adverse drug event (ADE) for patients taking warfarin for anticoagulant therapy.

Aim

The purpose of this study was to determine the association between increased INR and the usage of warfarin by using the pharmacovigilance data from the FDA Adverse Event Reporting System (FAERS).

Methods

The ADEs in patients who took warfarin (N = 77,010) were analyzed using FAERS data. Association rule mining was applied to identify warfarin-related ADEs that were most associated with elevated INR (n = 15,091) as well as possible drug-drug interactions (DDIs) associated with increased INR. Lift values were used to identify ADEs that were most commonly reported alongside elevated INR based on the correlation between both item sets. In addition, this study sought to determine if the increased INR risk was influenced by sex, age, temporal distribution, and geographic distribution and were reported as reporting odds ratios (RORs).

Results

The top 5 ADEs most associated with increased INR in patients taking warfarin were decreased hemoglobin (lift = 2.31), drug interactions (lift = 1.88), hematuria (lift = 1.58), asthenia (lift = 1.44), and fall (lift = 1.32). INR risk increased as age increased, with individuals older than 80 having a 63% greater likelihood of elevated INR compared to those younger than 50. Males were 9% more likely to report increased INR as an ADE compared to females. Individuals taking warfarin concomitantly with at least one other drug were 43% more likely to report increased INR. The top 5 most frequently identified DDIs in patients taking warfarin and presenting with elevated INR were acetaminophen (lift = 1.81), ramipril (lift = 1.71), furosemide (lift = 1.64), bisoprolol (lift = 1.58), and simvastatin (lift = 1.58).

Conclusion

The risk of elevated INR increased as patient age increased, particularly among those older than 80. Elevated INR frequently co-presented with decreased hemoglobin, drug interactions, hematuria, asthenia, and fall in patients taking warfarin. This effect may be less pronounced in women due to the procoagulatory effects of estrogen signaling. Multiple possible DDIs were identified, including acetaminophen, ramipril, and furosemide.

Literature-based drug-drug similarity for drug repurposing: impact of Medical Subject Headings term refinement and hierarchical clustering

Background: We describe herein, an improved procedure for drug repurposing based on refined Medical Subject Headings (MeSH) terms and hierarchical clustering method. Materials & methods: In the present study, we have employed MeSH data from MEDLINE (2019), 1669 US FDA approved drugs from Open FDA and a refined set of MeSH terms. Refinement of MeSH terms was performed to include terms related to mechanistic information of drugs and diseases. Results and Conclusions: In-depth analysis of the results obtained, demonstrated greater efficiency of the proposed approach, based on refined MeSH terms and hierarchical clustering, in terms of number of selected drug candidates for repurposing. Further, analysis of misclustering and size of noise clusters suggest that the proposed approach is reliable and can be employed in drug repurposing.

Therapeutic dilemmas in addressing SARS-CoV-2 infection: Favipiravir versus Remdesivir

Coronavirus disease 2019 (COVID-19) represents an unmet clinical need, due to a high mortality rate, rapid mutation rate in the virus, increased chances of reinfection, lack of effectiveness of repurposed drugs and economic damage. COVID-19 pandemic has created an urgent need for effective molecules. Clinically proven efficacy and safety profiles have made favipiravir (FVP) and remdesivir (RDV) promising therapeutic options for use against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection. Even though both are prodrug molecules with an antiviral role based on a similar mechanism of action, differences in pharmacological, pharmacokinetic and pharmacotoxicological mechanisms have been identified. The present study aims to provide a comprehensive comparative assessment of FVP and RDV against SARS-CoV-2 infections, by centralizing medical data provided by significant literature and authorized clinical trials, focusing on the importance of a better understanding of the interactions between drug molecules and infectious agents in order to improve the global management of COVID-19 patients and to reduce the risk of antiviral resistance.

How to catch trends using MeSH terms analysis?

The paper describes a scheme for the comparative analysis of the sets of Pubmed publications. The proposed analysis is based on the comparison of the frequencies of occurrence of keywords—MeSH terms. The purpose of the analysis is to identify MeSH terms that characterize research areas specific to each group of articles, as well as to identify trends—topics on which the number of published works has changed significantly in recent years. The proposed approach was tested by comparing a set of medical publications and a group of articles in the field of personalized medicine. We analyzed about 700 thousand abstracts published in the period 2009–2021 and indexed them with MeSH terms. Topics with increasing research interest have been identified both in the field of medicine in general and specific to personalized medicine. Retrospective analysis of the keywords frequency of occurrence changes has shown the shift of the scientific priorities in this area over the past 10 years. The revealed patterns can be used to predict the relevance and significance of the scientific work direction in the horizon of 3–5 years. The proposed analysis can be scaled in the future for a larger number of groups of publications, as well as adjusted by introducing filters at the stage of sampling (scientific centers, journals, availability of full texts, etc.) or selecting a list of keywords (frequency threshold, use of qualifiers, category of generalizations).

Dysphagia Risk in Patients Prescribed Rivastigmine: A Systematic Analysis of FDA Adverse Event Reporting System

BACKGROUND

Dysphagia has been reported as an adverse event for patients receiving rivastigmine for Alzheimer's disease (AD) treatment.

OBJECTIVE

The purpose of this study was to determine the association between dysphagia and the usage of rivastigmine by using the pharmacovigilance data from the FDA Adverse Event Reporting System (FAERS).

METHODS

The risk of dysphagia in patients who took rivastigmine was compared with those of patients who took other medications. In addition, this study sought to determine if the dysphagia risk was influenced by sex, age, dosage, and medication routes of administration.

RESULTS

When compared to patients prescribed donepezil, galantamine, or memantine, individuals prescribed rivastigmine were almost twice as likely to report dysphagia as an adverse event. The dysphagia risk in individuals prescribed rivastigmine is comparable to individuals prescribed penicillamine but significantly higher than clozapine, drugs of which have been previously shown to be associated with elevated dysphagia likelihood. Individuals older than 80 were 122% more likely to report having dysphagia after being prescribed rivastigmine than patients that were 50-70 years of age. Oral administration of rivastigmine was associated with approximately 2 times greater likelihood of reporting dysphagia relative to users of the transdermal patch. In addition, dysphagia showed higher association with pneumonia than other commonly reported adverse events.

CONCLUSION

Patients prescribed rivastigmine were at greater risk of reporting dysphagia as an adverse event than patients prescribed many other medicines. This increase in dysphagia occurrence may be attributed to the dual inhibition of both acetylcholinesterase and butyrylcholinesterase.

Drug-Drug interaction among admitted patients at primary, district and referral hospitals' medical wards in East Gojjam Zone, Amhara Regional State, Ethiopia

Objectives:

This study was aimed to assess the type, prevalence, characteristics of drug interaction and factors associated from admitted patients in medical wards at primary, district and referral hospitals in East Gojjam Zone, Amhara Regional State, Ethiopia.

Methods:

A facility-based retrospective cross-sectional study design was conducted among admitted patients in medical wards at different hospitals of East Gojjam Zone from September 2019 to February 2020. Patient-specific data were extracted from patient medical prescription papers using a structured data collection tool. Potential drug–drug interaction was identified using www.drugs.com as drug–drug interaction checker. Data were analyzed using SPSS version 23.0. To identify the explanatory predictors of potential drug–drug interaction, logistic regression analysis was done at a statistical significance level of p-value < 0.05.

Results:

Of the total 554 prescriptions, 51.1% were prescribed for females with a mean (±standard deviation) age of 40.85 ± 23.09 years. About 46.4% prescriptions of patients had one or more comorbid conditions, and the most frequent identified comorbid conditions were infectious (18.6%) and cardiac problems (6.3%) with 0.46 ± 0.499 average number of comorbid conditions per patient. Totally, 1516 drugs were prescribed with 2.74 ± 0.848 mean number per patient and range of 2–6. Two hundred and forty-two (43.7%) prescriptions had at least one potential drug–drug interaction, and it was found that 292 drug interactions were presented. Almost half of the drug–drug interaction identified was moderate (50%). Overall, the prevalence rate of drug–drug interaction was 43.7%. Older age (adjusted odds ratio = 8.301; 95% confidence interval (5.51–12.4), p = 0.000), presence of comorbidities (adjusted odds ratio = 1.72; 95% confidence interval (1.10–2.68), p = 0.000) and number of medications greater or equal to 3 (adjusted odds ratio = 2.69; 95% confidence interval (1.42–5.11), p = 0.000) were independent predictors for the occurrence of potential drug–drug interaction.

Conclusion:

The prevalence of potential drug–drug interaction among admitted patients was relatively high. Pharmacodynamic drug–drug interaction was the common mechanism of drug–drug interaction with moderate degree. Therefore, close follow-up of hospitalized patients is highly recommended.

Advancement in predicting interactions between drugs used to treat psoriasis and its comorbidities by integrating molecular and clinical resources

OBJECTIVE

Drug-drug interactions (DDIs) can result in adverse and potentially life-threatening health consequences; however, it is challenging to predict potential DDIs in advance. We introduce a new computational approach to comprehensively assess the drug pairs which may be involved in specific DDI types by combining information from large-scale gene expression (984 transcriptomic datasets), molecular structure (2159 drugs), and medical claims (150 million patients).

MATERIALS AND METHODS

Features were integrated using ensemble machine learning techniques, and we evaluated the DDIs predicted with a large hospital-based medical records dataset. Our pipeline integrates information from >30 different resources, including >10 000 drugs and >1.7 million drug-gene pairs. We applied our technique to predict interactions between 37 611 drug pairs used to treat psoriasis and its comorbidities.

RESULTS

Our approach achieves >0.9 area under the receiver operator curve (AUROC) for differentiating 11 861 known DDIs from 25 750 non-DDI drug pairs. Significantly, we demonstrate that the novel DDIs we predict can be confirmed through independent data sources and supported using clinical medical records.

CONCLUSIONS

By applying machine learning and taking advantage of molecular, genomic, and health record data, we are able to accurately predict potential new DDIs that can have an impact on public health.

Evaluating the research domain and achievement for a productive researcher who published 114 sole-author articles: A bibliometric analysis

BACKGROUND

Team science research includes authors from various fields collaborating to publish their work on certain topics. Despite the numerous papers that discussed the ordering of author names and the contributions of authors to an article, no paper evaluatedIn addition, few researchers publish academic articles without co-author collaboration. Whether the bibliometric indexes (eg, h-/x-index) of sole-author researchers are higher than those of other types of multiple authors is required for comparison. We aimed to evaluate a productive author who published 114 sole-author articles with exceptional RA and RD in academics.

METHODS

By searching the PubMed database (Pubmed.com), we used the keyword of (Taiwan[affiliation]) from 2016 to 2017 and downloaded 29,356 articles. One physician (Dr. Tseng from the field of Internal Medicine) who published 12 articles as a single author was selected. His articles and citations were searched in PubMed. A comparison of various types of author ordering placements was conducted using sensitivity analysis to inspect whether this sole author earns the highest metrics in RA. Social network analysis (SNA), Gini coefficient (GC), pyramid plot, and the Kano diagram were applied to gather the following data for visualization: RESULTS:: We observed that CONCLUSIONS:: The metrics on RA are high for the sole author studied. The author's RD can be denoted by the MeSH terms and measured by the GC. The author-weighted scheme is required for quantifying author credits in an article to evaluate the author's RA. Social network analysis incorporating the Kano diagrams provided insights into the relationships between actors (eg, coauthors, MeSH terms, or journals). The methods used in this study can be replicated to evaluate other productive studies on RA and RD in the future.

[Medical subject headings for the scientific groups evolution analysis on the example of academician A.I. Archakov's scientific school]

This paper proposes a method of comparative analysis of scientific trajectories based on bibliographic profiles. The bibliographic profile ("meshprint") is a list of MeSH terms (key terms used to index articles in the PubMed), indicating the relative frequency of occurrence of each term in the scientist's articles. Comparison of personalized bibliographic profiles can be represented in the form of a semantic network, where the nodes are the names of scientists, and the relationships are proportional to the calculated measures of similarity of bibliographic profiles. The proposed method was used to analyze the semantic network of scientists united by the academic school of the academician A.I. Archakov. The results of the work allowed us to show the relationship between the scientific trajectories of one scientific school and to correlate the results with world trends.

Using the Kano model to display the most cited authors and affiliated countries in schizophrenia research

In order to improve individual research achievements (IRA), this study investigates which affiliated countries and authors earn the most cited IRAs and whether those types of articles are associated with the number of cited papers on schizophrenia from a leading journal in the field. The Kano model was used for displaying the IRAs. Clusters of medical subject headings (MeSH) were applied to explore the core concepts of a given journal. This study aimed to apply social network analysis (SNA) and an authorship-weighted scheme (AWS) to inspect the association between MeSH terms and IRA. About 2,008 abstracts published between 2012 and 2016 in the journal Schizophrenia Research were downloaded from Pubmed Central using the keyword (Schizophr Res)[Journal] on September 20, 2018. The MeSH terms were clustered by using SNA to separate the core concepts and compare the differences in bibliometric indices (i.e., h, Ag, x and author impact factor or AIF). Visual dashboards were shown on Google Maps. Results indicate that (1) the US, the UK, and Canada earn the highest x-index; (2) the top one author from the US has the highest x-index (= 5.73 with x-core at cited = 16.44 and citable = 2); (3) the article type of schizophrenic psychology shows distinctly higher frequencies than others; and (4) article types are associated with the number of cited papers. Four approaches of the Kano model, SNA, MeSH terms, and AWS can be accommodated to display IRAs, classify article types, and quantify coauthor contributions in the article byline, respectively, and applied to other scientific disciplines in the future, not just in this specific journal.

Inferring new relations between medical entities using literature curated term co-occurrences

Objectives

Identifying new relations between medical entities, such as drugs, diseases, and side effects, is typically a resource-intensive task, involving experimentation and clinical trials. The increased availability of related data and curated knowledge enables a computational approach to this task, notably by training models to predict likely relations. Such models rely on meaningful representations of the medical entities being studied. We propose a generic features vector representation that leverages co-occurrences of medical terms, linked with PubMed citations.

Materials and Methods

We demonstrate the usefulness of the proposed representation by inferring two types of relations: a drug causes a side effect and a drug treats an indication. To predict these relations and assess their effectiveness, we applied 2 modeling approaches: multi-task modeling using neural networks and single-task modeling based on gradient boosting machines and logistic regression.

Results

These trained models, which predict either side effects or indications, obtained significantly better results than baseline models that use a single direct co-occurrence feature. The results demonstrate the advantage of a comprehensive representation.

Discussion

Selecting the appropriate representation has an immense impact on the predictive performance of machine learning models. Our proposed representation is powerful, as it spans multiple medical domains and can be used to predict a wide range of relation types.

Conclusion

The discovery of new relations between various medical entities can be translated into meaningful insights, for example, related to drug development or disease understanding. Our representation of medical entities can be used to train models that predict such relations, thus accelerating healthcare-related discoveries.

Constructing a bilingual website with validated database for Herb and Western medicine interactions using Ginseng, Ginkgo and Dong Quai as examples

Background

Concerns have been raised regarding the efficacy and safety resulting from the potential interactions of herbs with Western medications due to the use of both herbs and Western medicine by the general public. Information obtained from the web must be critically evaluated prior to its use in making decisions.

Description

This study aimed to construct an herb-drug interaction (HDI) website (https://drug-herb-interaction.netlify.com) with a critically reviewed database. Node.js was used to store the database by running JavaScript. Vue.js is a front-end framework used for web interface development. A total of 135 sets of information related to the interactions of ginseng, ginkgo and dong quai with Western medicine from the literature identified in Medline were collected, followed by critical reviews to prepare nineteen items of information for each HDI monograph. A total of 80 sets of validated HDIs met all criteria and were further assessed at the individual reliability level (likely, possible, and unevaluable) and labeled with the “interaction” item. This query system of the website can be operated in both the Chinese and English languages to obtain all monographs on HDIs in the database, including bilingual interaction data. The database of HDI monographs can be updated by simply uploading a new version of the information Excel file. The designed “smart search” module, in addition to the “single search”, is convenient for requesting multiple searches. Among the “likely” interactions (n = 26), 50% show negative HDIs. Ten of these can increase the effect of the Western drug, and the others (n = 3) imply that the HDI can be beneficial.

Conclusions

The current study provides a website platform and 80 sets of validated bilingual HDIs involving ginseng, ginkgo and dong quai in an online database. A search of HDI monographs related to these three herbs can be performed with this bilingual, easy-to-use query website, which is feasible for professionals and the general public. The identified reliability level for each HDI may assist readers’ decisions regarding whether taking Western medications concomitant with one of three herbal medicinal foods is safe or whether caution is required due to potentially serious outcomes.

Whether article types of a scholarly journal are different in cited metrics using cluster analysis of MeSH terms to display: A bibliometric analysis

BACKGROUND

Many authors are concerned which types of peer-review articles can be cited most in academics and who were the highest-cited authors in a scientific discipline. The prerequisites are determined by: (1) classifying article types; and (2) quantifying co-author contributions. We aimed to apply Medical Subject Headings (MeSH) with social network analysis (SNA) and an authorship-weighted scheme (AWS) to meet the prerequisites above and then demonstrate the applications for scholars.

METHODS

By searching the PubMed database (pubmed.com), we used the keyword "Medicine" [journal] and downloaded 5,636 articles published from 2012 to 2016. A total number of 9,758 were cited in Pubmed Central (PMC). Ten MeSH terms were separated to represent the journal types of clusters using SNA to compare the difference in bibliometric indices, that is, h, g, and x as well as author impact factor(AIF). The methods of Kendall coefficient of concordance (W) and one-way ANOVA were performed to verify the internal consistency of indices and the difference across MeSH clusters. Visual representations with dashboards were shown on Google Maps.

RESULTS

We found that Kendall W is 0.97 (χ = 26.22, df = 9, P < .001) congruent with internal consistency on metrics across MeSH clusters. Both article types of methods and therapeutic use show higher frequencies than other 8 counterparts. The author Klaus Lechner (Austria) earns the highest research achievement(the mean of core articles on g = Ag = 15.35, AIF = 21, x = 3.92, h = 1) with one paper (PMID: 22732949, 2012), which was cited 23 times in 2017 and the preceding 5 years.

CONCLUSION

Publishing article type with study methodology and design might lead to a higher IF. Both classifying article types and quantifying co-author contributions can be accommodated to other scientific disciplines. As such, which type of articles and who contributes most to a specific journal can be evaluated in the future.

Public data sources to support systems toxicology applications

Public databases provide a wealth of freely available information about chemicals, genes, proteins, biological networks, phenotypes, diseases, and exposure science that can be integrated to construct pathways for systems toxicology applications. Relating this disparate information from public repositories, however, can be challenging since databases use a variety of ways to represent, describe, and make available their content. The use of standard vocabularies to annotate key data concepts, however, allows the information to be more easily exchanged and combined for discovery of new findings. We explore some of the many public data sources currently available to support systems toxicology, and demonstrate the value of standardizing data to help construct chemical-induced outcome pathways.

Exploring the Chemical Space of Cytochrome P450 Inhibitors Using Integrated Physicochemical Parameters, Drug Efficiency Metrics and Decision Tree Models

The cytochrome P450s (CYPs) play a central role in the metabolism of various endogenous and exogenous compounds including drugs. CYPs are vulnerable to inhibition and induction which can lead to adverse drug reactions. Therefore, insights into the underlying mechanism of CYP450 inhibition and the estimation of overall CYP inhibitor properties might serve as valuable tools during the early phases of drug discovery. Herein, we present a large data set of inhibitors against five major metabolic CYPs (CYP1A2, CYP2C9, CYP2C19, CYP2D6 and CYP3A4) for the evaluation of important physicochemical properties and ligand efficiency metrics to define property trends across various activity levels (active, efficient and inactive). Decision tree models for CYP inhibition were developed with an accuracy >90% for both the training set and 10-folds cross validation. Overall, molecular weight (MW), hydrogen bond acceptors/donors (HBA/HBD) and lipophilicity (clogP/logPo/w) represent important physicochemical descriptors for CYP450 inhibitors. However, highly efficient CYP inhibitors show mean MW, HBA, HBD and logP values between 294.18–482.40,5.0–8.2,1–7.29 and 1.68–2.57, respectively. Our results might help in optimization of toxicological profiles associated with new chemical entities (NCEs), through a better understanding of inhibitor properties leading to CYP-mediated interactions.

CuDDI: A CUDA-Based Application for Extracting Drug-Drug Interaction Related Substance Terms from PubMed Literature

Drug-drug interaction (DDI) is becoming a serious issue in clinical pharmacy as the use of multiple medications is more common. The PubMed database is one of the biggest literature resources for DDI studies. It contains over 150,000 journal articles related to DDI and is still expanding at a rapid pace. The extraction of DDI-related information, including compounds and proteins from PubMed, is an essential step for DDI research. In this paper, we introduce a tool, CuDDI (compute unified device architecture-based DDI searching), for identification of DDI-related terms (including compounds and proteins) from PubMed. There are three modules in this application, including the automatic retrieval of substances from PubMed, the identification of DDI-related terms, and the display of relationship of DDI-related terms. For DDI term identification, a speedup of 30–105 times was observed for the compute unified device architecture (CUDA)-based version compared with the implementation with a CPU-based Python version. CuDDI can be used to discover DDI-related terms and relationships of these terms, which has the potential to help clinicians and pharmacists better understand the mechanism of DDIs. CuDDI is available at: https://github.com/chengusf/CuDDI.

Identifying Common Methods Used by Drug Interaction Experts for Finding Evidence About Potential Drug-Drug Interactions: Web-Based Survey

Background

Preventing drug interactions is an important goal to maximize patient benefit from medications. Summarizing potential drug-drug interactions (PDDIs) for clinical decision support is challenging, and there is no single repository for PDDI evidence. Additionally, inconsistencies across compendia and other sources have been well documented. Standard search strategies for complete and current evidence about PDDIs have not heretofore been developed or validated.

Objective

This study aimed to identify common methods for conducting PDDI literature searches used by experts who routinely evaluate such evidence.

Methods

We invited a convenience sample of 70 drug information experts, including compendia editors, knowledge-base vendors, and clinicians, via emails to complete a survey on identifying PDDI evidence. We created a Web-based survey that included questions regarding the (1) development and conduct of searches; (2) resources used, for example, databases, compendia, search engines, etc; (3) types of keywords used to search for the specific PDDI information; (4) study types included and excluded in searches; and (5) search terms used. Search strategy questions focused on 6 topics of the PDDI information—(1) that a PDDI exists; (2) seriousness; (3) clinical consequences; (4) management options; (5) mechanism; and (6) health outcomes.

Results

Twenty participants (response rate, 20/70, 29%) completed the survey. The majority (17/20, 85%) were drug information specialists, drug interaction researchers, compendia editors, or clinical pharmacists, with 60% (12/20) having >10 years’ experience. Over half (11/20, 55%) worked for clinical solutions vendors or knowledge-base vendors. Most participants developed (18/20, 90%) and conducted (19/20, 95%) search strategies without librarian assistance. PubMed (20/20, 100%) and Google Scholar (11/20, 55%) were most commonly searched for papers, followed by Google Web Search (7/20, 35%) and EMBASE (3/20, 15%). No respondents reported using Scopus. A variety of subscription and open-access databases were used, most commonly Lexicomp (9/20, 45%), Micromedex (8/20, 40%), Drugs@FDA (17/20, 85%), and DailyMed (13/20, 65%). Facts and Comparisons was the most commonly used compendia (8/20, 40%). Across the 6 attributes of interest, generic drug name was the most common keyword used. Respondents reported using more types of keywords when searching to identify the existence of PDDIs and determine their mechanism than when searching for the other 4 attributes (seriousness, consequences, management, and health outcomes). Regarding the types of evidence useful for evaluating a PDDI, clinical trials, case reports, and systematic reviews were considered relevant, while animal and in vitro data studies were not.

Conclusions

This study suggests that drug interaction experts use various keyword strategies and various database and Web resources depending on the PDDI evidence they are seeking. Greater automation and standardization across search strategies could improve one’s ability to identify PDDI evidence. Hence, future research focused on enhancing the existing search tools and designing recommended standards is needed.

Finding medication doses in the liteature

Medication doses, one of the determining factors in medication safety and effectiveness, are present in the literature, but only in free-text form. We set out to determine if the systems developed for extracting drug prescription information from clinical text would yield comparable results on scientific literature and if sequence-to-sequence learning with neural networks could improve over the current state-of-the-art. We developed a collection of 694 PubMed Central documents annotated with drug dose information using the i2b2 schema. We found that less than half of the drug doses are present in the MEDLINE/PubMed abstracts, and full-text is needed to identify the other half. We identified the differences in the scope and formatting of drug dose information in the literature and clinical text, which require developing new dose extraction approaches. Finally, we achieved 83.9% recall, 87.2% precision and 85.5% F1 score in extracting complete drug prescription information from the literature.

Predicting potential drug-drug interactions on topological and semantic similarity features using statistical learning

Drug-drug interaction (DDI) is a change in the effect of a drug when patient takes another drug. Characterizing DDIs is extremely important to avoid potential adverse drug reactions. We represent DDIs as a complex network in which nodes refer to drugs and links refer to their potential interactions. Recently, the problem of link prediction has attracted much consideration in scientific community. We represent the process of link prediction as a binary classification task on networks of potential DDIs. We use link prediction techniques for predicting unknown interactions between drugs in five arbitrary chosen large-scale DDI databases, namely DrugBank, KEGG, NDF-RT, SemMedDB, and Twosides. We estimated the performance of link prediction using a series of experiments on DDI networks. We performed link prediction using unsupervised and supervised approach including classification tree, k-nearest neighbors, support vector machine, random forest, and gradient boosting machine classifiers based on topological and semantic similarity features. Supervised approach clearly outperforms unsupervised approach. The Twosides network gained the best prediction performance regarding the area under the precision-recall curve (0.93 for both random forests and gradient boosting machine). The applied methodology can be used as a tool to help researchers to identify potential DDIs. The supervised link prediction approach proved to be promising for potential DDIs prediction and may facilitate the identification of potential DDIs in clinical research.