Prediction of drug indications based on chemical interactions and chemical similarities.

Discovering potential indications of novel or approved drugs is a key step in drug development. Previous computational approaches could be categorized into disease-centric and drug-centric based on the starting point of the issues or small-scaled application and large-scale application according to the diversity of the datasets. Here, a classifier has been constructed to predict the indications of a drug based on the assumption that interactive/associated drugs or drugs with similar structures are more likely to target the same diseases using a large drug indication dataset. To examine the classifier, it was conducted on a dataset with 1,573 drugs retrieved from Comprehensive Medicinal Chemistry database for five times, evaluated by 5-fold cross-validation, yielding five 1st order prediction accuracies that were all approximately 51.48%. Meanwhile, the model yielded an accuracy rate of 50.00% for the 1st order prediction by independent test on a dataset with 32 other drugs in which drug repositioning has been confirmed. Interestingly, some clinically repurposed drug indications that were not included in the datasets are successfully identified by our method. These results suggest that our method may become a useful tool to associate novel molecules with new indications or alternative indications with existing drugs.

Automatic annotation of ICD-to-MedDRA mappings with SKOS predicates.

Robust alignments between ICD and MedDRA are essential to enable the secondary use of clinical data for pharmacovigilance research. UMLS makes available ICD-to-MedDRA mappings, but they are only poorly specified, which introduces difficulties when exploited in an automatic way. SKOS vocabulary can help achieve quality and machine-processable mappings. We have developed an algorithm based on several simple rules which annotates automatically ICD-to-MedDRA mappings with SKOS predicates. The method was tested and evaluated on a sample of ICD-10-to MedDRA mappings extracted from UMLS. The algorithm demonstrated satisfying performances, especially for skos:exactMatch properties, which suggests that automatic methods can be used to improve the quality of terminology mappings.

The effect object paradigm--a means to support medication safety with clinical decision support.

BACKGROUND: In many countries, officially approved drug information known as summary of product characteristics (SPC) is mostly available in text form, which cannot be used for Clinical Decision Support Systems (CDSS). It may be essential however to substantiate CDSS advice with such legally binding text snippets. In an attempt to link various drug data sources including SPC towards a CDSS to support medication safety in psychiatric patients we arrived at the notion of an effect object. METHODS: A requirements analysis revealed data items and data structure which are needed from the patient and from the drug information source for the CDSS functionality. Published drug data modelling approaches were analyzed and found unsuitable. A conceptional database modeling approach using top down and bottom up modeling was performed. RESULTS: The schema based data model implemented within the django framework centered on SPC "effect objects" which comprise all SPC data required for the respective CDSS function such as search for contraindications in the proposed medication. Today six effect objects have been defined for contraindications and warnings, missing indications, adverse effects, drug-drug interactions, dosing and pharmacokinetics. CONCLUSION: The transformation of SPC data to a database-driven "effect objects" structure permits decoupling between the CDSS functions and different underlying data sources and supports the design of reusable, stable and verified CDSS functions.

A hybrid method for prediction and repositioning of drug Anatomical Therapeutic Chemical classes.

In the Anatomical Therapeutic Chemical (ATC) classification system, therapeutic drugs are divided into 14 main classes according to the organ or system on which they act and their chemical, pharmacological and therapeutic properties. This system, recommended by the World Health Organization (WHO), provides a global standard for classifying medical substances and serves as a tool for international drug utilization research to improve quality of drug use. In view of this, it is necessary to develop effective computational prediction methods to identify the ATC-class of a given drug, which thereby could facilitate further analysis of this system. In this study, we initiated an attempt to develop a prediction method and to gain insights from it by utilizing ontology information of drug compounds. Since only about one-fourth of drugs in the ATC classification system have ontology information, a hybrid prediction method combining the ontology information, chemical interaction information and chemical structure information of drug compounds was proposed for the prediction of drug ATC-classes. As a result, by using the Jackknife test, the 1st prediction accuracies for identifying the 14 main ATC-classes in the training dataset, the internal validation dataset and the external validation dataset were 75.90%, 75.70% and 66.36%, respectively. Analysis of some samples with false-positive predictions in the internal and external validation datasets indicated that some of them may even have a relationship with the false-positive predicted ATC-class, suggesting novel uses of these drugs. It was conceivable that the proposed method could be used as an efficient tool to identify ATC-classes of novel drugs or to discover novel uses of known drugs.

Exploring pharmacoepidemiologic groupings of drugs from a clinical perspective.

OBJECTIVES: To investigate the extent to which pharmacoepidemiologic groupings are homogeneous in terms of clinical properties. METHODS: In our analysis, we classified drug subgroups from the pharmacoepidemiologic Anatomical Therapeutic Chemical (ATC) classification system based on clinical drug properties. We established mappings from ATC fifth level drug entities to drug property annotations in the National Drug File Reference Terminology (NDF-RT), including therapeutic categories, mechanisms of action, and physiologic effects. Based on the annotations for the individual drugs we computed homogeneity scores for all ATC groups and analyzed their distribution. CONCLUSIONS: We found ATC groups to be generally homogeneous, more so for mechanisms of action, and physiologic effects than for therapeutic intent. However, only half of all ATC drugs can be analyzed with this approach, in part because of missing properties in NDF-RT.

Stevens-Johnson syndrome/toxic epidermal necrolysis: are drug dictionaries correctly informing physicians regarding the risk?

BACKGROUND: Stevens-Johnson syndrome/toxic epidermal necrolysis (SJS/TEN) are severe drug reactions associated with high mortality and multiple incapacitating sequelae. In the past 20 years, two large multinational case control studies, published in 1995 and 2008, had identified different degrees of drug association with SJS/TEN: 'strongly associated', 'associated', 'suspected' and 'not suspected' medications. OBJECTIVE: The aim of this study was to check the adequacy of mention of risk of SJS/TEN in the drug dictionaries most widely used by physicians in five European countries. STUDY DESIGN: In each country one expert investigator looked at the most widely used drug dictionary (2009 edition) for mentions of risk of SJS/TEN. This was done for a predefined list of medications with a different degree of risk. The presence and clarity or absence of warning was compared with available evidence provided by published results from case-control studies. SETTING: The five countries participating in the RegiSCAR group: Austria, France, Germany, The Netherlands and the UK. RESULTS: A total of 3,268 drug descriptions of medications for systemic use were analysed, including all brands of 14 'strongly associated' drugs, 5 'associated' drugs and 12 widely used drugs with no established association. Discrepancies were found by country, and between descriptions for different brands of the same generic. Among 522 descriptions of 14 'strongly associated' drugs, only 5 did not mention the risk. For the 1,013 descriptions of 'associated' drugs, 3 % did not mention the risk. One-third of 'not suspected' drugs contained a specific or less specific warning (e.g. bullous cutaneous eruption). Warnings for 'strongly associated' medications were often as imprecise as those for 'not suspected' drugs. CONCLUSION: Information on the risk of SJS/TEN in drug dictionaries needs improvement to enhance the quality of advice given by general physicians and to raise the understanding of risk by patients.

Comparison of alphabetical versus categorical display format for medication order entry in a simulated touch screen anesthesia information management system: an experiment in clinician-computer interaction in anesthesia.

BACKGROUND: Anesthesia information management system (AIMS) records should be designed and configured to facilitate the accurate and prompt recording of multiple drugs administered coincidentally or in rapid succession. METHODS: We proposed two touch-screen display formats for use with our department's new EPIC touch-screen AIMS. In one format, medication "buttons" were arranged in alphabetical order (i.e. A-C, D-H etc.). In the other, buttons were arranged in categories (Common, Fluids, Cardiovascular, Coagulation etc.). Both formats were modeled on an iPad screen to resemble the AIMS interface. Anesthesia residents, anesthesiologists, and Certified Registered Nurse Anesthetists (n = 60) were then asked to find and touch the correct buttons for a series of medications whose names were displayed to the side of the entry screen. The number of entries made within 2 minutes was recorded. This was done 3 times for each format, with the 1st format chosen randomly. Data were analyzed from the third trials with each format to minimize differences in learning. RESULTS: The categorical format had a mean of 5.6 more drugs entered using the categorical method in two minutes than the alphabetical format (95% confidence interval [CI] 4.5 to 6.8, P < 0.0001). The findings were the same regardless of the order of testing (i.e. alphabetical-categorical vs. categorical - alphabetical) and participants' years of clinical experience. Most anesthesia providers made no (0) errors for most trials (N = 96/120 trials, lower 95% limit 73%, P < 0.0001). There was no difference in error rates between the two formats (P = 0.53). CONCLUSIONS: The use of touch-screen user interfaces in healthcare is increasingly common. Arrangement of drugs names in a categorical display format in the medication order-entry touch screen of an AIMS can result in faster data entry compared to an alphabetical arrangement of drugs. Results of this quality improvement project were used in our department's design of our final intraoperative electronic anesthesia record. This testing approach using cognitive and usability engineering methods can be used to objectively design and evaluate many aspects of the clinician-computer interaction in electronic health records.

Mapping Partners Master Drug Dictionary to RxNorm using an NLP-based approach.

OBJECTIVE: To develop an automated method based on natural language processing (NLP) to facilitate the creation and maintenance of a mapping between RxNorm and a local medication terminology for interoperability and meaningful use purposes. METHODS: We mapped 5961 terms from Partners Master Drug Dictionary (MDD) and 99 of the top prescribed medications to RxNorm. The mapping was conducted at both term and concept levels using an NLP tool, called MTERMS, followed by a manual review conducted by domain experts who created a gold standard mapping. The gold standard was used to assess the overall mapping between MDD and RxNorm and evaluate the performance of MTERMS. RESULTS: Overall, 74.7% of MDD terms and 82.8% of the top 99 terms had an exact semantic match to RxNorm. Compared to the gold standard, MTERMS achieved a precision of 99.8% and a recall of 73.9% when mapping all MDD terms, and a precision of 100% and a recall of 72.6% when mapping the top prescribed medications. CONCLUSION: The challenges and gaps in mapping MDD to RxNorm are mainly due to unique user or application requirements for representing drug concepts and the different modeling approaches inherent in the two terminologies. An automated approach based on NLP followed by human expert review is an efficient and feasible way for conducting dynamic mapping.

Availability of pharmacogenetic and pharmacogenomic information in anticancer drug monographs in France: personalized cancer therapy.

AIM: To determine the availability of pharmacogenetic and pharmacogenomic information for healthcare professionals in France during 2009 for anticancer drugs. MATERIALS & METHODS: We searched in the informatic version of the VIDAL dictionary which is currently used by healthcare professionals in France. We then compared this with data available in the PubMed database. RESULTS: Among the 109 anticancer molecules available in France during 2009, 13 have pharmacogenomic or pharmacogenetic information in their monographs. In the scientific literature, we found numerous pharmacogenomic and pharmacogenetic biomarkers concerning 43 of the 109 anticancer agents. Some are pharmacogenomic biomarkers related to drug effectiveness, others are pharmacogenetic biomarkers related to drug toxicity. CONCLUSION: We believe that the lack of pharmacogenomic and pharmacogenetic information in drug monographs reflects the relative newness of the discipline. However, pharmacogenetics and pharmacogenomics can offer valuable information for improving the safety of drugs, reducing toxicity and predicting nonresponders. The drugs might then be incorporated into clinical practice through several strategies, including increased drug labeling and better education of healthcare professionals.

Database for crude drugs and Kampo medicine.

A wiki-based repository for crude drugs and Kampo medicine is introduced. It provides taxonomic and chemical information for 158 crude drugs and 348 prescriptions of the traditional Kampo medicine in Japan, which is a variation of ancient Chinese medicine. The system is built on MediaWiki with extensions for inline page search and for sending user-input elements to the server. These functions together realize implementation of word checks and data integration at the user-level. In this scheme, any user can participate in creating an integrated database with controlled vocabularies on the wiki system. Our implementation and data are accessible at http://metabolomics.jp/wiki/.

RxTerms - a drug interface terminology derived from RxNorm.

A good interface terminology is an essential component of any Computerized Provider Order Entry system. RxTerms is a drug interface terminology derived from RxNorm. By reorganizing the drug information into two dimensions as prescribers do when writing prescriptions and by eliminating drug names that are less likely to be needed in a prescribing environment, RxTerms helps the user to efficiently enter complete prescription orders. Preliminary evaluation of RxTerms using a list of most commonly prescribed drugs showed that its coverage was very good (99% for both generic and branded drug names). There was significant efficiency gain compared to using the unprocessed RxNorm names. RxTerms fills the gap for a free, up-to-date drug interface terminology that is linked to RxNorm, the U.S. designated standard for clinical drugs.

Using the RxNorm web services API for quality assurance purposes.

Auditing large, rapidly evolving terminological systems is still a challenge. In the case of RxNorm, a standardized nomenclature for clinical drugs, we argue that quality assurance processes can benefit from the recently released application programming interface (API) provided by RxNav. We demonstrate the usefulness of the API by performing a systematic comparison of alternative paths in the RxNorm graph, over several thousands of drug entities. This study revealed potential errors in RxNorm, currently under review. The results also prompted us to modify the implementation of RxNav to navigate the RxNorm graph more accurately. The RxNav web services API used in this experiment is robust and fast.

Identification of inactive medications in narrative medical text.

Discontinued medications are frequently not removed from EMR medication lists - a patient safety risk. We developed an algorithm to identify inactive medications using in the text of narrative notes in the EMR. The algorithm was evaluated against manual review of 297 randomly selected notes. One in five notes documented inactive medications. Sensitivity and precision of 87.7% and 80.7%, respectively, on per-note basis and 66.3% and 80.0%, respectively, on per-medication basis. When medication names missing from the dictionary were excluded, the algorithm achieved sensitivity of 91.4%. Using real clinical data, the algorithm identified inactive medications documented in the note but still listed as active on the patients medication list in more than one in ten notes. Documentation of inactive medications is common in narrative provider notes and can be computationally extracted. This technology could be employed in real-time patient care as well as for research and quality of care monitoring.

A novel system for rapidly identifying toxic chemicals.

First-responders have a critical need to rapidly identify toxic chemicals during emergencies. However, current systems such as WISER require a large number of inputs before a chemical can be identified. Here we present a novel system which significantly reduces the number of inputs required to identify a toxic chemical.