Relation extraction from DailyMed structured product labels by optimally combining crowd, experts and machines.

The effectiveness of machine learning models to provide accurate and consistent results in drug discovery and clinical decision support is strongly dependent on the quality of the data used. However, substantive amounts of open data that drive drug discovery suffer from a number of issues including inconsistent representation, inaccurate reporting, and incomplete context. For example, databases of FDA-approved drug indications used in computational drug repositioning studies do not distinguish between treatments that simply offer symptomatic relief from those that target the underlying pathology. Moreover, drug indication sources often lack proper provenance and have little overlap. Consequently, new predictions can be of poor quality as they offer little in the way of new insights. Hence, work remains to be done to establish higher quality databases of drug indications that are suitable for use in drug discovery and repositioning studies. Here, we report on the combination of weak supervision (i.e., programmatic labeling and crowdsourcing) and deep learning methods for relation extraction from DailyMed text to create a higher quality drug-disease relation dataset. The generated drug-disease relation data shows a high overlap with DrugCentral, a manually curated dataset. Using this dataset, we constructed a machine learning model to classify relations between drugs and diseases from text into four categories; treatment, symptomatic relief, contradiction, and effect, exhibiting an improvement of 15.5% with Bi-LSTM (F1 score of 71.8%) over the best performing discrete method. Access to high quality data is crucial to building accurate and reliable drug repurposing prediction models. Our work suggests how the combination of crowds, experts, and machine learning methods can go hand-in-hand to improve datasets and predictive models.

OncoPDSS: an evidence-based clinical decision support system for oncology pharmacotherapy at the individual level.

BACKGROUND: Precision oncology pharmacotherapy relies on precise patient-specific alterations that impact drug responses. Due to rapid advances in clinical tumor sequencing, an urgent need exists for a clinical support tool that automatically interprets sequencing results based on a structured knowledge base of alteration events associated with clinical implications. RESULTS: Here, we introduced the Oncology Pharmacotherapy Decision Support System (OncoPDSS), a web server that systematically annotates the effects of alterations on drug responses. The platform integrates actionable evidence from several well-known resources, distills drug indications from anti-cancer drug labels, and extracts cancer clinical trial data from the ClinicalTrials.gov database. A therapy-centric classification strategy was used to identify potentially effective and non-effective pharmacotherapies from user-uploaded alterations of multi-omics based on integrative evidence. For each potentially effective therapy, clinical trials with faculty information were listed to help patients and their health care providers find the most suitable one. CONCLUSIONS: OncoPDSS can serve as both an integrative knowledge base on cancer precision medicine, as well as a clinical decision support system for cancer researchers and clinical oncologists. It receives multi-omics alterations as input and interprets them into pharmacotherapy-centered information, thus helping clinicians to make clinical pharmacotherapy decisions. The OncoPDSS web server is freely accessible at https://oncopdss.capitalbiobigdata.com .

LabeledIn: cataloging labeled indications for human drugs.

Drug-disease treatment relationships, i.e., which drug(s) are indicated to treat which disease(s), are among the most frequently sought information in PubMed®. Such information is useful for feeding the Google Knowledge Graph, designing computational methods to predict novel drug indications, and validating clinical information in EMRs. Given the importance and utility of this information, there have been several efforts to create repositories of drugs and their indications. However, existing resources are incomplete. Furthermore, they neither label indications in a structured way nor differentiate them by drug-specific properties such as dosage form, and thus do not support computer processing or semantic interoperability. More recently, several studies have proposed automatic methods to extract structured indications from drug descriptions; however, their performance is limited by natural language challenges in disease named entity recognition and indication selection. In response, we report LabeledIn: a human-reviewed, machine-readable and source-linked catalog of labeled indications for human drugs. More specifically, we describe our semi-automatic approach to derive LabeledIn from drug descriptions through human annotations with aids from automatic methods. As the data source, we use the drug labels (or package inserts) submitted to the FDA by drug manufacturers and made available in DailyMed. Our machine-assisted human annotation workflow comprises: (i) a grouping method to remove redundancy and identify representative drug labels to be used for human annotation, (ii) an automatic method to recognize and normalize mentions of diseases in drug labels as candidate indications, and (iii) a two-round annotation workflow for human experts to judge the pre-computed candidates and deliver the final gold standard. In this study, we focused on 250 highly accessed drugs in PubMed Health, a newly developed public web resource for consumers and clinicians on prevention and treatment of diseases. These 250 drugs corresponded to more than 8000 drug labels (500 unique) in DailyMed in which 2950 candidate indications were pre-tagged by an automatic tool. After being reviewed independently by two experts, 1618 indications were selected, and additional 97 (missed by computer) were manually added, with an inter-annotator agreement of 88.35% as measured by the Kappa coefficient. Our final annotation results in LabeledIn consist of 7805 drug-disease treatment relationships where drugs are represented as a triplet of ingredient, dose form, and strength. A systematic comparison of LabeledIn with an existing computer-derived resource revealed significant discrepancies, confirming the need to involve humans in the creation of such a resource. In addition, LabeledIn is unique in that it contains detailed textual context of the selected indications in drug labels, making it suitable for the development of advanced computational methods for the automatic extraction of indications from free text. Finally, motivated by the studies on drug nomenclature and medication errors in EMRs, we adopted a fine-grained drug representation scheme, which enables the automatic identification of drugs with indications specific to certain dose forms or strengths. Future work includes expanding our coverage to more drugs and integration with other resources. The LabeledIn dataset and the annotation guidelines are available at http://ftp.ncbi.nlm.nih.gov/pub/lu/LabeledIn/.

Predicting new drug indications based on double variational autoencoders.

Experimental drug development is costly, complex, and time-consuming, and the number of drugs that have been put into application treatment is small. The identification of drug-disease correlations can provide important information for drug discovery and drug repurposing. Computational drug repurposing is an important and effective method that can be used to determine novel treatments for diseases. In recent years, an increasing number of large databases have been utilized for biological data research, particularly in the fields of drugs and diseases. Consequently, researchers have begun to explore the application of deep neural networks in biological data development. One particularly promising method for unsupervised learning is the deep generative model, with the variational autoencoder (VAE) being among the mainstream models. Here, we propose a drug indication prediction algorithm called DIDVAE (predicting new drug indications based on double variational autoencoders), which generates new data by learning the latent variable distribution of known data to achieve the goal of predicting drug-disease associations. In the experiment, we compared the DIDVAE algorithm with the BBNR, DrugNet, MBiRW and DRRS algorithms on a unified dataset. The comprehensive experimental results show that, compared with these prediction algorithms, the DIDVAE algorithm provides an overall improved prediction. In addition, further analysis and verification of the predicted unknown drug-disease association also proved the practicality of the method.

Combining biomedical knowledge graphs and text to improve predictions for drug-target interactions and drug-indications.

Biomedical knowledge is represented in structured databases and published in biomedical literature, and different computational approaches have been developed to exploit each type of information in predictive models. However, the information in structured databases and literature is often complementary. We developed a machine learning method that combines information from literature and databases to predict drug targets and indications. To effectively utilize information in published literature, we integrate knowledge graphs and published literature using named entity recognition and normalization before applying a machine learning model that utilizes the combination of graph and literature. We then use supervised machine learning to show the effects of combining features from biomedical knowledge and published literature on the prediction of drug targets and drug indications. We demonstrate that our approach using datasets for drug-target interactions and drug indications is scalable to large graphs and can be used to improve the ranking of targets and indications by exploiting features from either structure or unstructured information alone.

InContext: curation of medical context for drug indications.

Accurate and precise information about the therapeutic uses (indications) of a drug is essential for applications in drug repurposing and precision medicine. Leading online drug resources such as DrugCentral and DrugBank provide rich information about various properties of drugs, including their indications. However, because indications in such databases are often partly automatically mined, some may prove to be inaccurate or imprecise. Particularly challenging for text mining methods is the task of distinguishing between general disease mentions in drug product labels and actual indications for the drug. For this, the qualifying medical context of the disease mentions in the text should be studied. Some examples include contraindications, co-prescribed drugs and target patient qualifications. No existing indication curation efforts attempt to capture such information in a precise way. Here we fill this gap by presenting a novel curation protocol for extracting indications and machine processable annotations of contextual information about the therapeutic use of a drug. We implemented the protocol on a reference set of FDA-approved drug product labels on the DailyMed website to curate indications for 150 anti-cancer and cardiovascular drugs. The resulting corpus - InContext - focuses on anti-cancer and cardiovascular drugs because of the heightened societal interest in cancer and heart disease. In order to understand how InContext relates with existing reputable drug indication databases, we analysed it's overlap with a state-of-the-art indications database - LabeledIn - as well as a reputable online drug compendium - DrugCentral. We found that 40% of indications sampled from DrugCentral (and 23% from LabeledIn) respectively, could not be accounted for in InContext. This raises questions about the veracity of indications not appearing in InContext. The additional contextual information curated by InContext about disease mentions in drug SPLs provides a foundation for more precise, structured and formal representations of knowledge related to drug therapeutic use, in order to increase accuracy and agreement of drug indication extraction methods for in silico drug repurposing.

How often do prescribers include indications in drug orders? Analysis of 4 million outpatient prescriptions.

PURPOSE: To examine the extent to which outpatient clinicians currently document drug indications in prescription instructions. METHODS: Free-text sigs were extracted from all outpatient prescriptions generated by the computerized prescriber order entry system of a major academic institution during a 5-year period. Natural language processing was used to identify drug indications. The data set was analyzed to determine the rates at which prescribers included indications. It was stratified by provider specialty, drug class, and specific medications, to determine how often these indications were in prescriptions for as-needed (PRN) versus non-PRN medications. RESULTS: During the study period, 4,356,086 prescriptions were ordered. Indications were included in 322,961 orders (7.41%). From these orders, 249,262 indications (77.18%) were written for PRN orders. Although internal medicine prescribers generated the highest number of medication orders, they included indications in only 6.26% of their prescriptions, whereas orthopedic surgery providers had the highest rate of documenting indications (33.41%). Pain was the most common indication, accounting for 30.35% of all documented indications. The drug class with the highest number of sigs-containing indications was narcotic analgesics. Non-PRN chronic medication prescriptions rarely included the indication. CONCLUSION: Prescribers rarely included drug indications in electronic free-text prescription instructions, and, when they did, it was mostly for PRN uses such as pain.

Incorporating medication indications into the prescribing process.

PURPOSE: The incorporation of medication indications into the prescribing process to improve patient safety is discussed. SUMMARY: Currently, most prescriptions lack a key piece of information needed for safe medication use: the patient-specific drug indication. Integrating indications could pave the way for safer prescribing in multiple ways, including avoiding look-alike/sound-alike errors, facilitating selection of drugs of choice, aiding in communication among the healthcare team, bolstering patient understanding and adherence, and organizing medication lists to facilitate medication reconciliation. Although strongly supported by pharmacists, multiple prior attempts to encourage prescribers to include the indication on prescriptions have not been successful. We convened 6 expert panels to consult high-level stakeholders on system design considerations and requirements necessary for building and implementing an indications-based computerized prescriber order-entry (CPOE) system. We summarize our findings from the 6 expert stakeholder panels, including rationale, literature findings, potential benefits, and challenges of incorporating indications into the prescribing process. Based on this stakeholder input, design requirements for a new CPOE interface and workflow have been identified. CONCLUSION: The emergence of universal electronic prescribing and content knowledge vendors has laid the groundwork for incorporating indications into the CPOE prescribing process. As medication prescribing moves in the direction of inclusion of the indication, it is imperative to design CPOE systems to efficiently and effectively incorporate indications into prescriber workflows and optimize ways this can best be accomplished.

Formalizing drug indications on the road to therapeutic intent.

Therapeutic intent, the reason behind the choice of a therapy and the context in which a given approach should be used, is an important aspect of medical practice. There are unmet needs with respect to current electronic mapping of drug indications. For example, the active ingredient sildenafil has 2 distinct indications, which differ solely on dosage strength. In progressing toward a practice of precision medicine, there is a need to capture and structure therapeutic intent for computational reuse, thus enabling more sophisticated decision-support tools and a possible mechanism for computer-aided drug repurposing. The indications for drugs, such as those expressed in the Structured Product Labels approved by the US Food and Drug Administration, appears to be a tractable area for developing an application ontology of therapeutic intent.

Toward a comprehensive drug ontology: extraction of drug-indication relations from diverse information sources.

BACKGROUND: Drug ontologies could help pharmaceutical researchers overcome information overload and speed the pace of drug discovery, thus benefiting the industry and patients alike. Drug-disease relations, specifically drug-indication relations, are a prime candidate for representation in ontologies. There is a wealth of available drug-indication information, but structuring and integrating it is challenging. RESULTS: We created a drug-indication database (DID) of data from 12 openly available, commercially available, and proprietary information sources, integrated by terminological normalization to UMLS and other authorities. Across sources, there are 29,964 unique raw drug/chemical names, 10,938 unique raw indication "target" terms, and 192,008 unique raw drug-indication pairs. Drug/chemical name normalization to CAS numbers or UMLS concepts reduced the unique name count to 91 or 85% of the raw count, respectively, 84% if combined. Indication "target" normalization to UMLS "phenotypic-type" concepts reduced the unique term count to 57% of the raw count. The 12 sources of raw data varied widely in coverage (numbers of unique drug/chemical and indication concepts and relations) generally consistent with the idiosyncrasies of each source, but had strikingly little overlap, suggesting that we successfully achieved source/raw data diversity. CONCLUSIONS: The DID is a database of structured drug-indication relations intended to facilitate building practical, comprehensive, integrated drug ontologies. The DID itself is not an ontology, but could be converted to one more easily than the contributing raw data. Our methodology could be adapted to the creation of other structured drug-disease databases such as for contraindications, precautions, warnings, and side effects.

[Off-label prescription: practice and problems]. / A prescrição fora das indicações aprovadas (off-label): prática e problemas.

Approval of a drug for clinical use requires production of data on efficacy and safety through submission of results from randomized controlled trials (RCTs), in which the new molecule is usually compared with placebo (or an active comparator) for a set of outcomes that will serve as the basis for the drug's indications. These indications are crucial, because drugs are approved on the basis of their net clinical benefit for specific and well-defined diseases and--importantly--only for these. Once the drug is available for use in tens or hundreds of thousands of patients, physicians may realize that some medications can be effective in diseases for which they were not approved, i.e., no studies have been presented to the regulatory authorities, and therefore they are not formally approved for those indications. Convinced of the benefits for their patients, some physicians prescribe them for unapproved indications--off-label prescription. In this paper we discuss the prevalence of off-label prescription, and its advantages and problems.

Randomized Trials Built on Sand: Examples from COPD, Hormone Therapy, and Cancer.

The randomized controlled trial is the fundamental study design to evaluate the effectiveness of medications and receive regulatory approval. Observational studies, on the other hand, are essential to address post-marketing drug safety issues but have also been used to uncover new indications or new benefits for already marketed drugs. Hormone replacement therapy (HRT) for instance, effective for menopausal symptoms, was reported in several observational studies during the 1980s and 1990s to also significantly reduce the incidence of coronary heart disease. This claim was refuted in 2002 by the large-scale Women's Health Initiative randomized trial. An example of a new indication for an old drug is that of metformin, an anti-diabetic medication, which is being hailed as a potential anti-cancer agent, primarily on the basis of several recent observational studies that reported impressive reductions in cancer incidence and mortality with its use. These observational studies have now sparked the conduct of large-scale randomized controlled trials currently ongoing in cancer. We show in this paper that the spectacular effects on new indications or new outcomes reported in many observational studies in chronic obstructive pulmonary disease (COPD), HRT, and cancer are the result of time-related biases, such as immortal time bias, that tend to seriously exaggerate the benefits of a drug and that eventually disappear with the proper statistical analysis. In all, while observational studies are central to assess the effects of drugs, their proper design and analysis are essential to avoid bias. The scientific evidence on the potential beneficial effects in new indications of existing drugs will need to be more carefully assessed before embarking on long and expensive unsubstantiated trials.