We manually reviewed 20 000 randomly selected, potential duplicate record-pairs to identify matches (10 000 training set, 10 000 test set). We evaluated the probabilistic expectation maximization, simple deterministic and fuzzy inference engine (FIE) algorithms. We used particle swarm to optimize algorithm parameters for a single and for two thresholds. We ran 10 iterations of optimization using the training set and report averaged performance against the test set.

The overall estimated duplicate rate was 6%. FIE and simple deterministic algorithms allowed a lower manual review set compared to the probabilistic method (FIE 1.9%, simple deterministic 2.5%, probabilistic 3.6%; p<0.001). For a single threshold, the simple deterministic algorithm performed better than the probabilistic method (positive predictive value 0.956 vs 0.887, sensitivity 0.985 vs 0.887, p<0.001). ER with FIE classifies 98.1% of record-pairs correctly (1/10 000 error rate), assigning the remainder to manual review.

Optimized deterministic algorithms outperform the probabilistic method. There is a strong case for considering optimized deterministic methods for ER.

Based on the undesirable effects section from the summary of product characteristics (SPC) of 7446 drugs, we have built a Danish ADE dictionary. Starting from this dictionary we have developed a pipeline for identifying possible ADEs in unstructured clinical narrative text. We use a named entity recognition (NER) tagger to identify dictionary matches in the text and post-coordination rules to construct ADE compound terms. Finally, we apply post-processing rules and filters to handle, for example, negations and sentences about subjects other than the patient. Moreover, this method allows synonyms to be identified and anatomical location descriptions can be merged to allow appropriate grouping of effects in the same location.

The method identified 1 970 731 (35 477 unique) possible ADEs in a large corpus of 6011 psychiatric hospital patient records. Validation was performed through manual inspection of possible ADEs, resulting in precision of 89% and recall of 75%.

The presented dictionary-building method could be used to construct other ADE dictionaries. The complication of compound words in Germanic languages was addressed. Additionally, the synonym and anatomical location collapse improve the method.

The developed dictionary and method can be used to identify possible ADEs in Danish clinical narratives.

We retrospectively analyzed 7322 messages (6430 secure messages and 892 eVisits). To assess the overall risk associated with the messages, we looked for deaths within 30 days of the message and hospitalizations and emergency department (ED) visits within 7 days following the message. We also examined message content for symptoms of chest pain, breathing concerns, and other symptoms associated with high risk.

Two deaths occurred within 30 days of a patient-generated message, but were not related to the message. There were six hospitalizations related to a previous secure message (0.09% of secure messages), and two hospitalizations related to a previous eVisit (0.22% of eVisits). High-risk symptoms were present in 3.5% of messages but a subject line search to identify these high-risk messages had a sensitivity of only 15% and a positive predictive value of 29%.

Patients use portal messages 3.5% of the time for potentially high-risk symptoms of chest pain, breathing concerns, abdominal pain, palpitations, lightheadedness, and vomiting. Death, hospitalization, or an ED visit was an infrequent outcome following a secure message or eVisit. Screening the message subject line for high-risk symptoms was not successful in identifying high-risk message content.

We searched multiple databases for articles evaluating handoff tools published between 1 February 1983 and 15 June 2012. The selected articles were categorized along the following dimensions: handoff tool characteristics, standardization initiatives, methodological framework, and theoretical perspectives underlying the evaluation.

Thirty-six articles met our inclusion criteria. Handoff evaluations were conducted primarily on electronic tools (64%), with a more recent focus on electronic medical record-integrated tools (36% since 2008). Most evaluations centered on intra-departmental tools (95%). Evaluation studies were quasi-experimental (42%) or observational (50%), with a major focus on handoff-related outcome measures (94%) using predominantly survey-based tools (70%) with user satisfaction metrics (53%). Most of the studies (81%) based their evaluation on aspects of standardization that included continuity of care and patient safety.

The nature, methodological, and theoretical foundations of handoff tool evaluations varied significantly in terms of their quality and rigor, thereby limiting their ability to inform strategic standardization initiatives. Future research should utilize rigorous, multi-method qualitative and quantitative approaches that capture the contextual nuances of handoffs, and evaluate their effect on patient-related outcomes.

Claims data from the dominant health plan in Mesa County, Colorado, from 1 April 2005 to 31 December 2010 were matched to HIE adoption data on the provider level. Using mixed effects regression models with the quarter as the unit of analysis, the effect of HIE adoption on testing rates and associated charges was assessed.

Claims submitted by 306 providers in 69 practices for 34 818 patients were analyzed. The rate of testing per provider was expressed as tests per 1000 patients per quarter. For primary care providers, the rate of laboratory testing increased over the time span (baseline 1041 tests/1000 patients/quarter, increasing by 13.9 each quarter) and shifted downward with HIE adoption (downward shift of 83, p<0.01). A similar effect was found for specialist providers (baseline 718 tests/1000 patients/quarter, increasing by 19.1 each quarter, with HIE adoption associated with a downward shift of 119, p<0.01). Even so, imputed charges for laboratory tests did not shift downward significantly in either provider group, possibly due to the skewed nature of these data. For radiology testing, HIE adoption was not associated with significant changes in rates or imputed charges in either provider group.

Ambulatory HIE adoption is unlikely to produce significant direct savings through reductions in rates of testing. The economic benefits of HIE may reside instead in other downstream outcomes of better informed, higher quality care.

Qualitative study using semi-structured interviews and direct observation of physicians (n=28) working in a multispecialty outpatient care organization.

We identified physicians' perceptions of uncertainty as an important variable in understanding differences in EHR use patterns. Drawing on theories from the medical and organizational literatures, we identified three categories of perceptions of uncertainty: reduction, absorption, and hybrid. We used an existing model of EHR use to categorize physician EHR use patterns as high, medium, and low based on degree of feature use, level of EHR-enabled communication, and frequency that EHR use patterns change.

Physicians' perceptions of uncertainty were distinctly associated with their EHR use patterns. Uncertainty reductionists tended to exhibit high levels of EHR use, uncertainty absorbers tended to exhibit low levels of EHR use, and physicians demonstrating both perspectives of uncertainty (hybrids) tended to exhibit medium levels of EHR use.

We find evidence linking physicians' perceptions of uncertainty with EHR use patterns. Study findings have implications for health IT research, practice, and policy, particularly in terms of impacting health IT design and implementation efforts in ways that consider differences in physicians' perceptions of uncertainty.

A combination of supervised, unsupervised, and rule-based methods were used. Supervised methods include conditional random fields and support vector machines. A flexible automated feature selection technique was used to select the best subset of features for each supervised task. Unsupervised methods include Brown clustering on several corpora, which result in our method being considered semisupervised.

On the 2012 Informatics for Integrating Biology and the Bedside (i2b2) shared task data, we achieved an overall event F1-measure of 0.8045, an overall temporal expression F1-measure of 0.6154, an overall temporal link detection F1-measure of 0.5594, and an end-to-end temporal link detection F1-measure of 0.5258. The most competitive system was our event recognition method, which ranked third out of the 14 participants in the event task.

Analysis reveals the event recognition method has difficulty determining which modifiers to include/exclude in the event span. The temporal expression recognition method requires significantly more normalization rules, although many of these rules apply only to a small number of cases. Finally, the temporal relation recognition method requires more advanced medical knowledge and could be improved by separating the single discourse relation classifier into multiple, more targeted component classifiers.

Recognizing events and temporal expressions can be achieved accurately by combining supervised and unsupervised methods, even when only minimal medical knowledge is available. Temporal normalization and temporal relation recognition, however, are far more dependent on the modeling of medical knowledge.

We have developed an epilepsy and seizure ontology (EpSO) using a four-dimensional epilepsy classification system that integrates the latest International League Against Epilepsy terminology recommendations and National Institute of Neurological Disorders and Stroke (NINDS) common data elements. It imports concepts from existing ontologies, including the Neural ElectroMagnetic Ontologies, and uses formal concept analysis to create a taxonomy of epilepsy syndromes based on their seizure semiology and anatomical location.

EpSO is used in a suite of informatics tools for (a) patient data entry, (b) epilepsy focused clinical free text processing, and (c) patient cohort identification as part of the multi-center NINDS-funded study on sudden unexpected death in epilepsy. EpSO is available for download at http://prism.case.edu/prism/index.php/EpilepsyOntology.

An epilepsy ontology consortium is being created for community-driven extension, review, and adoption of EpSO. We are in the process of submitting EpSO to the BioPortal repository.

EpSO plays a critical role in informatics tools for epilepsy patient care and multi-center clinical research.

We used 21 354 075 MEDLINE records as the text corpus. We extracted drug–SE co-occurrence pairs using a cancer drug lexicon and a clean SE lexicon that we created. We then developed two filtering approaches to remove drug–disease treatment pairs and subsequently a ranking scheme to further prioritize filtered pairs. Finally, we analyzed relationships among SEs, gene targets, and indications.

We extracted 56 602 cancer drug–SE pairs. The filtering algorithms improved the precision of extracted pairs from 0.252 at baseline to 0.426, representing a 69% improvement in precision with no decrease in recall. The ranking algorithm further prioritized filtered pairs and achieved a precision of 0.778 for top-ranked pairs. We showed that cancer drugs that share SEs tend to have overlapping gene targets and overlapping indications.

The relationship extraction approach is effective in extracting many cancer drug–SE pairs from the literature. This unique knowledge base, when combined with existing cancer drug SE knowledge, can facilitate drug target discovery, drug repurposing, and toxicity prediction.

This overview targets medical informatics researchers who are unfamiliar with the problems and applications of temporal reasoning over clinical text.

We review the major applications of text-based temporal reasoning, describe the challenges for software systems handling temporal information in clinical text, and give an overview of the state of the art. Finally, we present some perspectives on future research directions that emerged during the recent community-wide challenge on text-based temporal reasoning in the clinical domain.

We developed two models to test our hypothesis. The first based on supervised machine learning was created using Apache's OpenNLP Maxent and the second based on information retrieval was created using Apache's Lucene. The models were validated by a random subsampling method that was repeated 20 times and that used 80/20 splits for training and testing, respectively. We also evaluated the performance of these models on all laboratory terms from three test institutions.

For the 20 iterations used for validation of our 80/20 splits Maxent and Lucene ranked the correct LOINC code first for between 70.5% and 71.4% and between 63.7% and 65.0% of local terms, respectively. For all laboratory terms from the three test institutions Maxent ranked the correct LOINC code first for between 73.5% and 84.6% (mean 78.9%) of local terms, whereas Lucene's performance was between 66.5% and 76.6% (mean 71.9%). Using a cut-off score of 0.46 Maxent always ranked the correct LOINC code first for over 57% of local terms.

This study showed that a rich corpus of local terms mapped to LOINC contains collective knowledge that can help map terms from other institutions. Using freely available software tools, we developed a data-driven automated approach that operates on term descriptions from existing mappings in the corpus. Accurate and efficient automated mapping methods can help to accelerate adoption of vocabulary standards and promote widespread health information exchange.

To analyze the effect of a WLAN-based real-time patient localization system for tracking outpatients in our level I trauma center.

Outpatients from April to August 2009 were included in the study, which was performed in two different stages. In phase I, patient tracking was performed with the real-time location system, but acquired data were not displayed to the personnel. In phase II tracking, the acquired data were automatically collected and displayed. Total treatment time was the primary outcome parameter. Statistical analysis was performed using multiple linear regression, with the significance level set at 0.05. Covariates included sex, age, type of encounter, prioritization, treatment team, number of residents, and radiographic imaging.

1045 patients were included in our study (540 in phase I and 505 in phase 2). An overall improvement of efficiency, as determined by a significantly decreased total treatment time (23.7%) from phase I to phase II, was noted. Additionally, significantly lower treatment times were noted for phase II patients even when other factors were considered (increased numbers of residents, the addition of imaging diagnostics, and comparison among various localization zones).

WLAN-based real-time patient localization systems can reduce process inefficiencies associated with manual patient identification and tracking.

Concept mapping, a mixed-methods statistical research tool, was used in the construction of the framework. Overall, 40 participants brainstormed or sorted and rated 83 statements concerning an IMS for older adults with VI. Multidimensional scaling and hierarchical cluster analysis were employed to construct the framework. A questionnaire yielded further insights into the views of a wider sample of older adults with VI (n=78) and caregivers (n=25) regarding IMS.

Concept mapping revealed a nine-cluster model of IMS-related aspects including affordability, awareness of system capabilities, simplicity of installation, operation and maintenance, system integrity and reliability, fall detection and safe movement, user customization, user preferences regarding information delivery, and safety alerts for patients and caregivers. From the questionnaire, independence, safety and fall detection were the most commonly reported reasons for older adults and caregivers to accept an IMS. Concerns included cost, privacy, security of the information obtained through monitoring, system accuracy, and ease of use.

Older adults with VI, caregivers and professionals are receptive to in-home monitoring, mainly for fall detection and safety monitoring, but have concerns that must be addressed when developing an IMS.

Our study provides a novel conceptual framework for the design of an IMS that will be maximally acceptable and beneficial to our ageing and vision-impaired population.

We analyzed data collected from logs of online searches gathered from consenting users of a browser toolbar from Microsoft (N=9740). We employed a complementary survey (N=489) to seek a deeper understanding of information-gathering, reflection, and action on the pursuit of professional healthcare.

We provide insights about HU through the survey, breaking out its findings by different respondent marginalizations as appropriate. Observations made from search logs may be explained by trends observed in our survey responses, even though the user populations differ.

The results provide insights about how users decide if and when to utilize healthcare resources, and how online health information seeking transitions to in-world HU. The findings from both the survey and the logs reveal behavioral patterns and suggest a strong relationship between search behavior and HU. Although the diversity of our survey respondents is limited and we cannot be certain that users visited medical facilities, we demonstrate that it may be possible to infer HU from long-term search behavior by the apparent influence that health concerns and professional advice have on search activity.

Our findings highlight different phases of online activities around queries pursuing professional healthcare facilities and services. We also show that it may be possible to infer HU from logs without tracking people's physical location, based on the effect of HU on pre- and post-HU search behavior. This allows search providers and others to develop more robust models of interests and preferences by modeling utilization rather than simply the intention to utilize that is expressed in search queries.

This mixed-method systematic review is based on a comprehensive search of Medline, Embase, PsychInfo and ABI/Inform Global. Study selection and coding were performed independently by two researchers and were followed by a narrative synthesis. To move beyond a simple description of the technologies, we employed and adapted the diffusion of innovation theory (DOI).

112 papers were included. Analysis revealed five main types of GGHIT: (1) telecare technologies (representing half of the studies); (2) electronic health records; (3) decision support systems; (4) web-based packages for patients and/or family caregivers; and (5) assistive information technologies. On aggregate, the most consistent finding proves to be the positive outcomes of GGHIT in terms of clinical processes. Although less frequently studied, positive impacts were found on patients’ health, productivity, efficiency and costs, clinicians’ satisfaction, patients’ satisfaction and patients’ empowerment.

Further efforts should focus on improving the characteristics of such technologies in terms of compatibility and simplicity. Implementation strategies also should be improved as trialability and observability are insufficient.

Our results will help organizations in making decisions regarding the choice, planning and diffusion of GGHIT implemented for the care of older adults.

We used a subset of 570 272 incidents including 1534 HIT incidents reported to MAUDE between 1 January 2008 and 1 July 2010. Text classifiers using regularized logistic regression were evaluated with both ‘balanced’ (50% HIT) and ‘stratified’ (0.297% HIT) datasets for training, validation, and testing. Dataset preparation, feature extraction, feature selection, cross-validation, classification, performance evaluation, and error analysis were performed iteratively to further improve the classifiers. Feature-selection techniques such as removing short words and stop words, stemming, lemmatization, and principal component analysis were examined.

statistic, F1 score, precision and recall.

Classification performance was similar on both the stratified (0.954 F1 score) and balanced (0.995 F1 score) datasets. Stemming was the most effective technique, reducing the feature set size to 79% while maintaining comparable performance. Training with balanced datasets improved recall (0.989) but reduced precision (0.165).

Statistical text classification appears to be a feasible method for identifying HIT reports within large databases of incidents. Automated identification should enable more HIT problems to be detected, analyzed, and addressed in a timely manner. Semi-supervised learning may be necessary when applying machine learning to big data analysis of patient safety incidents and requires further investigation.

Datasets were the Center for Medicaid and Medicare Services ICD-9-CM to ICD-10-CM mapping files, general equivalence mappings, and statewide Medicaid emergency department billing. Diagnoses were represented as nodes and their mappings as directional relationships. The complex network was synthesized as an aggregate of simpler motifs and tabulation per clinical specialty.

We identified five mapping motif categories: identity, class-to-subclass, subclass-to-class, convoluted, and no mapping. Convoluted mappings indicate that multiple ICD-9-CM and ICD-10-CM codes share complex, entangled, and non-reciprocal mappings. The proportions of convoluted diagnoses mappings (36% overall) range from 5% (hematology) to 60% (obstetrics and injuries). In a case study of 24 008 patient visits in 217 emergency departments, 27% of the costs are associated with convoluted diagnoses, with ‘abdominal pain’ and ‘gastroenteritis’ accounting for approximately 3.5%.

Previous qualitative studies report that administrators and clinicians are likely to be challenged in understanding and managing their practice because of the ICD-10-CM transition. We substantiate the complexity of this transition with a thorough quantitative summary per clinical specialty, a case study, and the tools to apply this methodology easily to any clinical practice in the form of a web portal and analytic tables.

Post-transition, successful management of frequent diseases with convoluted mapping network patterns is critical. The http://lussierlab.org/transition-to-ICD10CM web portal provides insight in linking onerous diseases to the ICD-10 transition.

To evaluate the viability of using non-expert crowds to label PAT, we compare expert (registered nurses) and non-expert (Amazon Mechanical Turk workers; Turkers) responses to a PAT medical term identification task. Next, we build a crowd-labeled dataset comprising 10 000 sentences from MedHelp. We train two models on this dataset and evaluate their performance, as well as that of MetaMap, Open Biomedical Annotator (OBA), and NaCTeM's TerMINE, against two gold standard datasets: one from MedHelp and the other from CureTogether.

When aggregated according to a corroborative voting policy, Turker responses predict expert responses with an F1 score of 84%. A conditional random field (CRF) trained on 10 000 crowd-labeled MedHelp sentences achieves an F1 score of 78% against the CureTogether gold standard, widely outperforming OBA (47%), TerMINE (43%), and MetaMap (39%). A failure analysis of the CRF suggests that misclassified terms are likely to be either generic or rare.

Our results show that combining statistical models sensitive to sentence-level context with crowd-labeled data is a scalable and effective technique for automatically identifying medical terms in PAT.

AP-HP decided to create a biomedical observation dictionary named AnaBio, to map it to LOINC and to maintain the mapping. A management platform based on methods used for knowledge engineering has been put in place. It aims at integrating AnaBio within the health information system and improving both the quality and stability of the dictionary.

This new management platform is now active in AP-HP. The AnaBio dictionary is shared by 120 laboratories and currently includes 50 000 codes. The mapping implementation to LOINC reaches 40% of the AnaBio entries and uses 26% of LOINC records. The results of our work validate the choice made to develop a local dictionary aligned with LOINC.

This work constitutes a first step towards a wider use of the platform. The next step will support the entire biomedical production chain, from the clinician prescription, through laboratory tests tracking in the laboratory information system to the communication of results and the use for decision support and biomedical research. In addition, the increase in the mapping implementation to LOINC ensures the interoperability allowing communication with other international health institutions.

We developed an image analysis workflow for extracting a large collection of morphological features and deployed the same on histological images from The Cancer Genome Atlas (TCGA) TNBC samples during the discovery phase (n=44). Strong correlations between salient morphological features and gene expression profiles from the same patients were identified. We then evaluated the same morphological features in predicting survival using a local TNBC cohort (n=143). We further tested the predictive power on patient prognosis of correlated gene clusters using two other public gene expression datasets.

Using TCGA data, we identified 48 pairs of significantly correlated morphological features and gene clusters; four morphological features were able to separate the local cohort with significantly different survival outcomes. Gene clusters correlated with these four morphological features further proved to be effective in predicting patient survival using multiple public gene expression datasets. These results suggest the efficacy of our workflow and demonstrate that integrative analysis holds promise for discovering biomarkers of complex diseases.

Quantitative dynamic contrast-enhanced MRI and diffusion-weighted MRI data were acquired on 28 patients before and after one cycle of NAC. A total of 118 semiquantitative and quantitative parameters were derived from these data and combined with 11 clinical variables. We used Bayesian logistic regression in combination with feature selection using a machine learning framework for predictive model building.

The best predictive models using feature selection obtained an area under the curve of 0.86 and an accuracy of 0.86, with a sensitivity of 0.88 and a specificity of 0.82.

With the numerous options for NAC available, development of a method to predict response early in the course of therapy is needed. Unfortunately, by the time most patients are found not to be responding, their disease may no longer be surgically resectable, and this situation could be avoided by the development of techniques to assess response earlier in the treatment regimen. The method outlined here is one possible solution to this important clinical problem.

Predictive modeling approaches based on machine learning using readily available clinical and quantitative MRI data show promise in distinguishing breast cancer responders from non-responders after the first cycle of NAC.

The system combines rule-based and machine learning approaches that rely on morphological, lexical, syntactic, semantic, and domain-specific features. Rule-based components were designed to handle the recognition and normalization of temporal expressions, while conditional random fields models were trained for event and temporal recognition.

The system achieved micro F scores of 90% for the extraction of temporal expressions and 87% for clinical event extraction. The normalization component for temporal expressions achieved accuracies of 84.73% (expression's type), 70.44% (value), and 82.75% (modifier).

Compared to the initial agreement between human annotators (87–89%), the system provided comparable performance for both event and temporal expression mining. While (lenient) identification of such mentions is achievable, finding the exact boundaries proved challenging.

The system provides a state-of-the-art method that can be used to support automated identification of mentions of clinical events and temporal expressions in narratives either to support the manual review process or as a part of a large-scale processing of electronic health databases.

Data from the Blue Cross Blue Shield of Michigan Cardiovascular Consortium (BMC2) multicenter registry for the years 2007 and 2008 (n=41 016) were used to train models to predict 13 different in-laboratory PCI complications using a novel one-plus-class support vector machine (OP-SVM) algorithm. The performance of these models in terms of discrimination and calibration was compared to the performance of models trained using the following classification algorithms on BMC2 data from 2009 (n=20 289): logistic regression (LR), one-class support vector machine classification (OC-SVM), and two-class support vector machine classification (TC-SVM). For the OP-SVM and TC-SVM approaches, variants of the algorithms with cost-sensitive weighting were also considered.

The OP-SVM algorithm and its cost-sensitive variant achieved the highest area under the receiver operating characteristic curve for the majority of the PCI complications studied (eight cases). Similar improvements were observed for the Hosmer–Lemeshow ² value (seven cases) and the mean cross-entropy error (eight cases).

The OP-SVM algorithm based on an augmented one-class learning problem improved discrimination and calibration across different PCI complications relative to LR and traditional support vector machine classification. Such an approach may have value in a broader range of clinical domains.

Here we propose a novel approach to find the minimum regulatory units of genes, co-methylated and co-expressed gene pairs (MEGP) that are highly correlated gene pairs between DNA methylation and gene expression showing the co-regulatory relationship. To evaluate whether our method is applicable to extract disease-associated genes, we applied our method to a large-scale dataset from the Cancer Genome Atlas extracting significantly associated MEGP and analyzed their functional correlation.

We observed that many MEGP physically interacted with each other and showed high semantic similarity with gene ontology terms. Furthermore, we performed gene set enrichment tests to identify how they are correlated in a complex biological process. Our MEGP were highly enriched in the biological pathway associated with ovarian cancers.

Our approach is useful for discovering coordinated epigenetic markers associated with specific diseases.

The term ‘informatics’ derives from the German ‘Informatik’, the Russian ‘Informatika’ and the French ‘informatique’, which combines ‘information’ with ‘automatique’. One concept lies at its core: information. While it is clear how ‘informatics’ is semantically linked to...]]> 2013-04-16T00:00:34-07:00 info:doi/10.1136/amiajnl-2013-001807 hwp:master-id:amiajnl;amiajnl-2013-001807 American Medical Informatics Association 2013-04-16 PostScript http://jamia.bmj.com/cgi/content/short/amiajnl-2012-001355v2?rss=1 Electronic health records (EHRs) are increasingly useful for health services research. For relatively uncommon conditions, such as multiple myeloma (MM) and its treatment-related complications, a combination of multiple EHR sources is essential for such research. The Shared Health Research Information Network (SHRINE) enables queries for aggregate results across participating institutions. Development of a rational search strategy in SHRINE may be augmented through analysis of pre-existing databases. We developed a SHRINE query for likely non-infectious treatment-related complications of MM, based upon an analysis of the Multiparameter Intelligent Monitoring in Intensive Care (MIMIC II) database. Using this query strategy, we found that the rate of likely treatment-related complications significantly increased from 2001 to 2007, by an average of 6% a year (p=0.01), across the participating SHRINE institutions. This finding is in keeping with increasingly aggressive strategies in the treatment of MM. This proof of concept demonstrates that a staged approach to federated queries, using external EHR data, can yield potentially clinically meaningful results.

We processed four public medication resources, RxNorm, Side Effect Resource (SIDER) 2, MedlinePlus, and Wikipedia, to create MEDI. We applied natural language processing and ontology relationships to extract indications for prescribable, single-ingredient medication concepts and all ingredient concepts as defined by RxNorm. Indications were coded as Unified Medical Language System (UMLS) concepts and International Classification of Diseases, 9th edition (ICD9) codes. A total of 689 extracted indications were randomly selected for manual review for accuracy using dual-physician review. We identified a subset of medication–indication pairs that optimizes recall while maintaining high precision.

MEDI contains 3112 medications and 63 343 medication–indication pairs. Wikipedia was the largest resource, with 2608 medications and 34 911 pairs. For each resource, estimated precision and recall, respectively, were 94% and 20% for RxNorm, 75% and 33% for MedlinePlus, 67% and 31% for SIDER 2, and 56% and 51% for Wikipedia. The MEDI high-precision subset (MEDI-HPS) includes indications found within either RxNorm or at least two of the three other resources. MEDI-HPS contains 13 304 unique indication pairs regarding 2136 medications. The mean±SD number of indications for each medication in MEDI-HPS is 6.22±6.09. The estimated precision of MEDI-HPS is 92%.

MEDI is a publicly available, computable resource that links medications with their indications as represented by concepts and billing codes. MEDI may benefit clinical EMR applications and reuse of EMR data for research.

To detect clinical events, we used rules and Conditional Random Fields. We built Random Forest models to identify event modality and polarity. To identify temporal expressions we built on the HeidelTime system. To detect temporal relations, we systematically studied their breakdown into distinct situations; we designed an oracle method to determine the most prominent situations and the most suitable associated classifiers, and combined their results.

We achieved F-measures of 0.8307 for event identification, based on rules, and 0.8385 for temporal expression identification. In the temporal relation task, we identified nine main situations in three groups, experimentally confirming shared intuitions: within-sentence relations, section-related time, and across-sentence relations. Logistic regression and Naïve Bayes performed best on the first and third groups, and decision trees on the second. We reached a 0.6231 global F-measure, improving by 7.5 points our official submission.

Carefully hand-crafted rules obtained good results for the detection of events and temporal expressions, while a combination of classifiers improved temporal link prediction. The characterization of the oracle recall of situations allowed us to point at directions where further work would be most useful for temporal relation detection: within-sentence relations and linking History of Present Illness events to the admission date. We suggest that the systematic situation breakdown proposed in this paper could also help improve other systems addressing this task.

The 2012 i2b2 NLP challenge organizers manually annotated 310 clinic notes according to a defined annotation guideline: a training set of 190 notes and a test set of 120 notes. All participating systems were developed on the training set and evaluated on the test set. Our system consists of three modules: event extraction, temporal expression extraction, and temporal relation (also called Temporal Link, or ‘TLink’) extraction. The TLink extraction module contains three individual classifiers for TLinks: (1) between events and section times, (2) within a sentence, and (3) across different sentences. The performance of our system was evaluated using scripts provided by the i2b2 organizers. Primary measures were micro-averaged Precision, Recall, and F-measure.

Our system was among the top ranked. It achieved F-measures of 0.8659 for temporal expression extraction (ranked fourth), 0.6278 for end-to-end TLink track (ranked first), and 0.6932 for TLink-only track (ranked first) in the challenge. We subsequently investigated different strategies for TLink extraction, and were able to marginally improve performance with an F-measure of 0.6943 for TLink-only track.

We developed a framework based on an ontology, the general information model (GIM), to unify structural models and terminologies, together with relevant mapping sets. This allowed a uniform access to these resources within LexEVS to facilitate interoperability by various components and data sources from implementing architectures.

Our unified framework has been tested in the context of the EU Framework Program 7 TRANSFoRm project, where it was used to achieve data integration in a retrospective diabetes cohort study. The GIM was successfully instantiated in TRANSFoRm as the clinical data integration model, and necessary mappings were created to support effective information retrieval for software tools in the project.

We present a novel, unifying approach to address interoperability challenges in heterogeneous data sources, by representing structural and semantic models in one framework. Systems using this architecture can rely solely on the GIM that abstracts over both the structure and coding. Information models, terminologies and mappings are all stored in LexEVS and can be accessed in a uniform manner (implementing the HL7 CTS2 service functional model). The system is flexible and should reduce the effort needed from data sources personnel for implementing and managing the integration.

Care providers who are potential end-users of the CDSS were invited to provide their recommendations for a random set of patients that represented diverse decision scenarios. The recommendations of the care providers and those generated by the CDSS were compared. Mismatched recommendations were reviewed by two independent experts.

A total of 25 users participated in this study and provided recommendations for 175 cases. The CDSS had an accuracy of 87% and 12 types of CDSS errors were identified, which were mainly due to deficiencies in the system's guideline rules. When the deficiencies were rectified, the CDSS generated optimal recommendations for all failure cases, except one with incomplete documentation.

The crowd-sourcing approach for construction of the reference set, coupled with the expert review of mismatched recommendations, facilitated an effective evaluation and enhancement of the system, by identifying decision scenarios that were missed by the system's developers. The described methodology will be useful for other researchers who seek rapidly to evaluate and enhance the deployment readiness of complex decision support systems.

The challenge evaluated systems on the information extraction tasks that targeted: (1) clinically significant events, including both clinical concepts such as problems, tests, treatments, and clinical departments, and events relevant to the patient's clinical timeline, such as admissions, transfers between departments, etc; (2) temporal expressions, referring to the dates, times, durations, or frequencies phrases in the clinical text. The values of the extracted temporal expressions had to be normalized to an ISO specification standard; and (3) temporal relations, between the clinical events and temporal expressions. Participants determined pairs of events and temporal expressions that exhibited a temporal relation, and identified the temporal relation between them.

For event detection, statistical machine learning (ML) methods consistently showed superior performance. While ML and rule based methods seemed to detect temporal expressions equally well, the best systems overwhelmingly adopted a rule based approach for value normalization. For temporal relation classification, the systems using hybrid approaches that combined ML and heuristics based methods produced the best results.

To construct automated systems for EVENT/TIMEX3 extraction and temporal link (TLINK) identification from clinical text.

The i2b2 organizers provided 190 annotated discharge summaries as the training set and 120 discharge summaries as the test set. Our Event system used a conditional random field classifier with a variety of features including lexical information, natural language elements, and medical ontology. The TIMEX3 system employed a rule-based method using regular expression pattern match and systematic reasoning to determine normalized values. The TLINK system employed both rule-based reasoning and machine learning. All three systems were built in an Apache Unstructured Information Management Architecture framework.

Our TIMEX3 system performed the best (F-measure of 0.900, value accuracy 0.731) among the challenge teams. The Event system produced an F-measure of 0.870, and the TLINK system an F-measure of 0.537.

Our TIMEX3 system demonstrated good capability of regular expression rules to extract and normalize time information. Event and TLINK machine learning systems required well-defined feature sets to perform well. We could also leverage expert knowledge as part of the machine learning features to further improve TLINK identification performance.

To develop a support vector machine (SVM)-based system for identifying EHR progress notes pertaining to diabetes, and to validate it at two institutions.

We retrieved 2000 EHR progress notes from patients with diabetes at the Brigham and Women's Hospital (1000 for training and 1000 for testing) and another 1000 notes from the University of Texas Physicians (for validation). We manually annotated all notes and trained a SVM using a bag of words approach. We then used the SVM on the testing and validation sets and evaluated its performance with the area under the curve (AUC) and F statistics.

The model accurately identified diabetes-related notes in both the Brigham and Women's Hospital testing set (AUC=0.956, F=0.934) and the external University of Texas Faculty Physicians validation set (AUC=0.947, F=0.935).

Overall, the model we developed was quite accurate. Furthermore, it generalized, without loss of accuracy, to another institution with a different EHR and a distinct patient and provider population.

It is possible to use a SVM-based classifier to identify EHR progress notes pertaining to diabetes, and the model generalizes well.

To design an automatic glaucoma diagnosis architecture automatic glaucoma diagnosis through medical imaging informatics (AGLAIA-MII) that combines patient personal data, medical retinal fundus image, and patient's genome information for screening.

2258 cases from a population study were used to evaluate the screening software. These cases were attributed with patient personal data, retinal images and quality controlled genome data. Utilizing the multiple kernel learning-based classifier, AGLAIA-MII, combined patient personal data, major image features, and important genome single nucleotide polymorphism (SNP) features.

Receiver operating characteristic curves were plotted to compare AGLAIA-MII's performance with classifiers using patient personal data, images, and genome SNP separately. AGLAIA-MII was able to achieve an area under curve value of 0.866, better than 0.551, 0.722 and 0.810 by the individual personal data, image and genome information components, respectively. AGLAIA-MII also demonstrated a substantial improvement over the current glaucoma screening approach based on intraocular pressure.

AGLAIA-MII demonstrates for the first time the capability of integrating patients’ personal data, medical retinal image and genome information for automatic glaucoma diagnosis and screening in a large dataset from a population study. It paves the way for a holistic approach for automatic objective glaucoma diagnosis and screening.

A two-arm comparative study with crossover design was conducted. Subjects, who were resident physicians at the University of Botswana, completed eight scenarios, each with multi-part questions. The primary outcome was a grade for each question. The primary independent variable was the intervention arm and other independent variables included residency and question.

Within each question type there were significant differences in ‘percentage correct’ between Medical Apps and PubMed4Hh for three of the six types of questions: drug-related, diagnosis/definitions, and treatment/management. Within each of these question types, Medical Apps had a higher percentage of fully correct responses than PubMed4Hh (63% vs 13%, 33% vs 12%, and 41% vs 13%, respectively). PubMed4Hh performed better for epidemiologic questions.

While mobile access to primary literature remains important and serves an information niche, mobile applications with condensed content may be more appropriate for point-of-care information needs. Further research is required to examine the specific information needs of clinicians in resource-limited settings and to evaluate the appropriateness of current resources in bridging location- and context-specific information gaps.

The team from the National Research Council Canada (NRC) submitted three system runs to the second track of the challenge: typifying the time-relationship between pre-annotated entities. The NRC system was designed around four specialist modules containing statistical machine learning classifiers. Each specialist targeted distinct sets of relationships: local relationships, ‘sectime’-type relationships, non-local overlap-type relationships, and non-local causal relationships.

The best NRC submission achieved a precision of 0.7499, a recall of 0.6431, and an F1 score of 0.6924, resulting in a statistical tie for first place. Post hoc improvements led to a precision of 0.7537, a recall of 0.6455, and an F1 score of 0.6954, giving the highest scores reported on this task to date.

Methods for general relation extraction extended well to temporal relations, and gave top-ranked state-of-the-art results. Careful ordering of predictions within result sets proved critical to this success.

Both quantitative and qualitative analyses were used to analyze responses to a cross-sectional survey disseminated to American Pharmacists Association pharmacists nationwide. The survey assessed factors related to pharmacists’ attitudinal readiness (ie, perceptions of impact) and pharmacies’ structural readiness (eg, availability of electronic resources, internet access) for the paperless labeling initiative.

We received a total of 436 survey responses (6% response rate) from pharmacists representing 44 US states and territories. Across the spectrum of settings we studied, pharmacists had work access to computers, printers, fax machines and access to the internet or intranet. Approximately 79% of respondents believed that the initiative would improve the adequacy of drug information available in their work site and 95% believed it would either not change (33%) or would improve (62%) communication to patients. Overall, respondents’ comments supported advancing the initiative; however, some comments revealed reservations regarding corporate or pharmacy buy-in, success of implementation, and ease of adoption.

This is the first nationwide study to report about pharmacists’ perspectives on paperless labeling. In general, pharmacists believe they are ready and that their pharmacies are well equipped for the transition to paperless labeling. Further exploration of perspectives from product label manufacturers and corporate pharmacy offices is needed to understand fully what will be necessary to complete this transition.

A public repository of clinical trials information (ClinicalTrials.gov) was explored to determine the state of plant-based interventions under investigation.

The results showed that ~15% of drug interventions in ClinicalTrials.gov were potentially plant related, with about 60% of them clustered within 10 taxonomic families. Further analysis of these plant-based interventions identified ~3.7% of associated plant species as endangered as determined from the International Union for the Conservation of Nature Red List.

The diversity of the plant kingdom has provided human civilization with life-sustaining food and medicine for centuries. There has been renewed interest in the investigation of botanicals as sources of new drugs, building on traditional knowledge about plant-based medicines. However, data about the plant-based biodiversity potential for therapeutics (eg, based on genetic or chemical information) are generally scattered across a range of sources and isolated from contemporary pharmacological resources. This study explored the potential to bridge biodiversity and biomedical knowledge sources.

The findings from this feasibility study suggest that there is an opportunity for developing plant-based drugs and further highlight taxonomic relationships between plants that may be rich sources for bioprospecting.

We conducted 10 focus groups with 75 African-American youth from a city with high HIV/STI prevalence. Data analyses involved coding using qualitative content analysis procedures and memo writing.

Unexpectedly, the majority of participants’ design recommendations concerned trust. Youth expressed distrust towards people and groups, which was amplified within the context of information technology-mediated interactions about HIV/STI. Participants expressed distrust in the reliability of condoms and the accuracy of HIV tests. They questioned the benevolence of many institutions, and some rejected authoritative HIV/STI information. Therefore, reputational information, including rumor, influenced HIV/STI-related decision making. Participants’ design requirements also focused on trust-related concerns. Accordingly, we developed a novel trust-centered design framework to guide intervention design.

Current approaches to online trust for health informatics do not consider group-level trusting patterns. Yet, trust was the central intervention-relevant issue among African-American youth, suggesting an important focus for culturally informed design. Our design framework incorporates: intervention objectives (eg, network embeddedness, participation); functional specifications (eg, decision support, collective action, credible question and answer services); and interaction design (eg, member control, offline network linkages, optional anonymity).

Trust is a critical focus for HIV/STI informatics interventions for young African Americans. Our design framework offers practical, culturally relevant, and systematic guidance to designers to reach this underserved group better.

Two target corpora from independent healthcare institutions were constructed from high frequency clinical narratives. Four leading POS taggers with their out-of-the-box models trained from general English and biomedical abstracts were evaluated against these clinical corpora. A high performing domain adaptation method, Easy Adapt, was compared to our newly proposed method ClinAdapt.

The evaluated POS taggers drop in accuracy by 8.5–15% when tested on clinical narratives. The highest performing tagger reports an accuracy of 88.6%. Domain adaptation with Easy Adapt reports accuracies of 88.3–91.0% on clinical texts. ClinAdapt reports 93.2–93.9%.

ClinAdapt successfully boosts POS tagging performance through domain adaptation requiring a modest amount of annotated clinical data. Improving the performance of critical NLP subtasks is expected to reduce pipeline error propagation leading to better overall results on complex processing tasks.

An end-to-end temporal relation system was developed. It includes three subsystems: an event extraction system (conditional random fields (CRF) name entity extraction and their corresponding attribute classifiers), a temporal extraction system (CRF name entity extraction, their corresponding attribute classifiers, and context-free grammar based normalization system), and a temporal relation system (10 multi-support vector machine (SVM) classifiers and a Markov logic networks inference system) using labeled sequential pattern mining, syntactic structures based on parse trees, and results from a coordination classifier. Micro-averaged precision (P), recall (R), averaged P&R (P&R), and F measure (F) were used to evaluate results.

For event extraction, the system achieved 0.9415 (P), 0.8930 (R), 0.9166 (P&R), and 0.9166 (F). The accuracies of their type, polarity, and modality were 0.8574, 0.8585, and 0.8560, respectively. For timex extraction, the system achieved 0.8818, 0.9489, 0.9141, and 0.9141, respectively. The accuracies of their type, value, and modifier were 0.8929, 0.7170, and 0.8907, respectively. For temporal relation, the system achieved 0.6589, 0.7129, 0.6767, and 0.6849, respectively. For end-to-end temporal relation, it achieved 0.5904, 0.5944, 0.5921, and 0.5924, respectively. With the F measure used for evaluation, we were ranked first out of 14 competing teams (event extraction), first out of 14 teams (timex extraction), third out of 12 teams (temporal relation), and second out of seven teams (end-to-end temporal relation).

The system achieved encouraging results, demonstrating the feasibility of the tasks defined by the i2b2 organizers. The experiment result demonstrates that both global and local information is useful in the 2012 challenge.

These difficulties in the collection of labeled patient data have led researchers to consider semi-supervised learning (SSL), a recent machine learning algorithm, because it is also capable of utilizing unlabeled patient data, which is relatively easier to collect. Therefore, it is regarded as an algorithm that could circumvent the known difficulties. However, the fact is yet valid even on SSL that more labeled data lead to better prediction. To compensate for the lack of labeled patient data, we may consider the concept of tagging virtual labels to unlabeled patient data, that is, ‘pseudo-labels,’ and treating them as if they were labeled.

Our proposed algorithm, ‘SSL Co-training’, implements this concept based on SSL. SSL Co-training was tested using the surveillance, epidemiology, and end results database for breast cancer and it delivered a mean accuracy of 76% and a mean area under the curve of 0.81.

To concurrently examine these phenomena, previous work has successfully considered the application of learning classifier systems (LCSs), a flexible class of evolutionary algorithms that distributes learned associations over a population of rules. Subsequent work dealt with the inherent problems of knowledge discovery and interpretation within these algorithms, allowing for the characterization of heterogeneous patterns of association. Whereas these previous advancements were evaluated using complex simulation studies, this study applied these collective works to a ‘real-world’ genetic epidemiology study of bladder cancer susceptibility.

We replicated the identification of previously characterized factors that modify bladder cancer risk—namely, single nucleotide polymorphisms from a DNA repair gene, and smoking. Furthermore, we identified potentially heterogeneous groups of subjects characterized by distinct patterns of association. Cox proportional hazard models comparing clinical outcome variables between the cases of the two largest groups yielded a significant, meaningful difference in survival time in years (survivorship). A marginally significant difference in recurrence time was also noted. These results support the hypothesis that an LCS approach can offer greater insight into complex patterns of association.

This methodology appears to be well suited to the dissection of disease heterogeneity, a key component in the advancement of personalized medicine.

In this study, we propose a fast, non-parametric, and model-free measure for three-way epistasis.

Such a measure is based on information gain, and is able to separate all lower order effects from pure three-way epistasis.

Our method was verified on synthetic data and applied to real data from a candidate-gene study of tuberculosis in a West African population. In the tuberculosis data, we found a statistically significant pure three-way epistatic interaction effect that was stronger than any lower-order associations.

Our study provides a methodological basis for detecting and characterizing high-order gene-gene interactions in genetic association studies.

We searched databases, previously collected papers, and references for relevant papers published from January 2000 to June 2012. For selection, papers had to: be a primary study; involve a computerized system for health information processing, decision support, or management reporting; and include an economic evaluation. Data on study design and economic evaluation methods were extracted and analyzed.

Forty-two papers were selected and 33 were deemed high quality (scores ≥8/10) for further analysis. These included 12 economic analyses, five input cost analyses, and 16 cost-related outcome analyses. For HIS types, there were seven primary care electronic medical records, six computerized provider order entry systems, five medication management systems, five immunization information systems, four institutional information systems, three disease management systems, two clinical documentation systems, and one health information exchange network. In terms of value for money, 23 papers reported positive findings, eight were inconclusive, and two were negative.

We found a wide range of economic evaluation papers that were based on different assumptions, methods, and metrics. There is some evidence of value for money in selected healthcare organizations and HIS types. However, caution is needed when generalizing these findings. Better reporting of economic evaluation studies is needed to compare findings and build on the existing evidence base we identified.

To reveal the molecular mechanisms of various intellectual disabilities and to suggest small molecules which by binding to the malfunctioning protein can reduce unwanted effects.

Using various in silico methods we reveal the molecular mechanism of XLID in cases involving proteins with known 3D structure. The 3D structures were used to predict the effect of disease-causing missense mutations on the folding free energy, conformational dynamics, hydrogen bond network and, if appropriate, protein-protein binding free energy.

It is shown that the vast majority of XLID mutation sites are outside the active pocket and are accessible from the water phase, thus providing the opportunity to alter their effect by binding appropriate small molecules in the vicinity of the mutation site.

This observation is used to demonstrate, computationally and experimentally, that a particular condition, Snyder-Robinson syndrome caused by the G56S spermine synthase mutation, might be ameliorated by small molecule binding.

Using data on cancer cases and cancer-free controls from Vanderbilt's biorepository, BioVU, we evaluated the utility of ICD-9 tobacco use codes to identify ever-smokers in general and high smoking prevalence (lung cancer) clinic populations. We assessed potential biases in documentation, and performed temporal analysis relating transitions between smoking codes to smoking cessation attempts. We also examined the suitability of these codes for use in genetic association analyses.

ICD-9 tobacco use codes can identify smokers in a general clinic population (specificity of 1, sensitivity of 0.32), and there is little evidence of documentation bias. Frequency of code transitions between ‘current’ and ‘former’ tobacco use was significantly correlated with initial success at smoking cessation (p<0.0001). Finally, code-based smoking status assignment is a comparable covariate to text-based smoking status for genetic association studies.

Our results support the use of ICD-9 tobacco use codes for identifying smokers in a clinical population. Furthermore, with some limitations, these codes are suitable for adjustment of smoking status in genetic studies utilizing electronic health records.

Researchers should not be deterred by the unavailability of full-text records to determine smoking status if they have ICD-9 code histories.

Five terminologies were identified, including the Systemized Nomenclature of Medical Clinical Terms (SNOMED CT), National Drug File–Reference Terminology (NDF-RT), Medication Dictionary for Regulatory Activities (MedDRA), Unique Ingredient Identifier (UNII), and RxNorm. A qualitative analysis was conducted to compare desirable characteristics of each terminology, including content coverage, concept orientation, formal definitions, multiple granularities, vocabulary structure, subset capability, and maintainability. A quantitative analysis was also performed to compare the content coverage of each terminology for (1) common food, drug, and environmental allergens and (2) descriptive concepts for common drug allergies, adverse reactions (AR), and no known allergies.

Our qualitative results show that SNOMED CT fulfilled the greatest number of desirable characteristics, followed by NDF-RT, RxNorm, UNII, and MedDRA. Our quantitative results demonstrate that RxNorm had the highest concept coverage for representing drug allergens, followed by UNII, SNOMED CT, NDF-RT, and MedDRA. For food and environmental allergens, UNII demonstrated the highest concept coverage, followed by SNOMED CT. For representing descriptive allergy concepts and adverse reactions, SNOMED CT and NDF-RT showed the highest coverage. Only SNOMED CT was capable of representing unique concepts for encoding no known allergies.

The proper terminology for encoding a patient's allergy is complex, as multiple elements need to be captured to form a fully structured clinical finding. Our results suggest that while gaps still exist, a combination of SNOMED CT and RxNorm can satisfy most criteria for encoding common allergies and provide sufficient content coverage.

We developed support vector machine (SVM) classifiers to disambiguate 197 ambiguous terms and abbreviations in the MSH WSD collection. Three different uncertainty sampling-based active learning algorithms were implemented with the SVM classifiers and were compared with a passive learner (PL) based on random sampling. For each ambiguous term and each learning algorithm, a learning curve that plots the accuracy computed from the test set as a function of the number of annotated samples used in the model was generated. The area under the learning curve (ALC) was used as the primary metric for evaluation.

Our experiments demonstrated that active learners (ALs) significantly outperformed the PL, showing better performance for 177 out of 197 (89.8%) WSD tasks. Further analysis showed that to achieve an average accuracy of 90%, the PL needed 38 annotated samples, while the ALs needed only 24, a 37% reduction in annotation effort. Moreover, we analyzed cases where active learning algorithms did not achieve superior performance and identified three causes: (1) poor models in the early learning stage; (2) easy WSD cases; and (3) difficult WSD cases, which provide useful insight for future improvements.

This study demonstrated that integrating active learning strategies with supervised WSD methods could effectively reduce annotation cost and improve the disambiguation models.

Two large-scale pharmacogenomic studies recently screened multiple anticancer drugs on over 1000 cell lines. We propose to combine these datasets to build and robustly validate genomic predictors of drug response. We compared five different approaches for building predictors of increasing complexity. We assessed their performance in cross-validation and in two large validation sets, one containing the same cell lines present in the training set and another dataset composed of cell lines that have never been used during the training phase.

Sixteen drugs were found in common between the datasets. We were able to validate multivariate predictors for three out of the 16 tested drugs, namely irinotecan, PD-0325901, and PLX4720. Moreover, we observed that response to 17-AAG, an inhibitor of Hsp90, could be efficiently predicted by the expression level of a single gene, NQO1.

These results suggest that genomic predictors could be robustly validated for specific drugs. If successfully validated in patients’ tumor cells, and subsequently in clinical trials, they could act as companion tests for the corresponding drugs and play an important role in personalized medicine.

We conducted 14 focus groups at five Department of Veterans Affairs facilities with 129 participants (54 physicians or practitioners, 34 nurses, and 37 administrators). Investigators qualitatively analyzed resultant transcripts, developed categories linked to the data, and identified emergent themes.

Five major themes related to CPD emerged: communication and coordination; control and limitations in expressivity; information availability and reasoning support; workflow alteration and disruption; and trust and confidence concerns. The results highlight that documentation intertwines tightly with clinical and administrative workflow. Perceptions differed between the three stakeholder groups but remained consistent within groups across facilities.

CPD has dramatically changed documentation processes, impacting clinical understanding, decision-making, and communication across multiple groups. The need for easy and rapid, yet structured and constrained, documentation often conflicts with the need for highly reliable and retrievable information to support clinical reasoning and workflows. Current CPD systems, while better than paper overall, often do not meet the needs of users, partly because they are based on an outdated ‘paper-chart’ paradigm. These findings should inform those implementing CPD systems now and future plans for more effective CPD systems.

To describe the effects of different communication interventions and their problems.

Prospective observational case study using a mixed methods approach of quantitative and qualitative methods.

General internal medicine (GIM) inpatient wards at five tertiary care academic teaching hospitals.

Clinicians consisting of residents, attending physicians, nurses, and allied health (AH) staff working on the GIM wards.

Ethnographic methods and interviews with clinical staff (doctors, nurses, medical students, and AH professionals) were conducted over a 16-month period from 2009 to 2010.

We identified four categories that described the intended and unintended consequences of communication interventions: impacts on senders, receivers, interprofessional collaboration, and the use of informal communication processes. The use of alphanumeric pagers, smartphones, and web-based communication systems had positive effects for senders and receivers, but unintended consequences were seen with all interventions in all four categories.

Interventions that aimed to improve clinical communications solved some but not all problems, and unintended effects were seen with all systems.

We hypothesized that protein interaction modeling of GWAS findings could highlight important disease-associated loci and unveil the role of their network topology in the genetic architecture of diseases with complex inheritance.

Network modeling of proteins associated with the intragenic single nucleotide polymorphisms of the National Human Genome Research Institute catalog of complex trait GWAS revealed that complex trait associated loci are more likely to be hub and bottleneck genes in available, albeit incomplete, networks (OR=1.59, Fisher's exact test p<2.24x10^–12). Network modeling also prioritized novel type 2 diabetes (T2D) genetic variations from the Finland–USA Investigation of Non-Insulin-Dependent Diabetes Mellitus Genetics and the Wellcome Trust GWAS data, and demonstrated the enrichment of hubs and bottlenecks in prioritized T2D GWAS genes. The potential biological relevance of the T2D hub and bottleneck genes was revealed by their increased number of first degree protein interactions with known T2D genes according to several independent sources (p<0.01, probability of being first interactors of known T2D genes).

Virtually all common diseases are complex human traits, and thus the topological centrality in protein networks of complex trait genes has implications in genetics, personal genomics, and therapy.

Manual annotation of a clinical narrative corpus of 127 606 tokens following the Treebank schema for syntactic information, PropBank schema for predicate-argument structures, and the Unified Medical Language System (UMLS) schema for semantic information. NLP components were developed.

The final corpus consists of 13 091 sentences containing 1772 distinct predicate lemmas. Of the 766 newly created PropBank frames, 74 are verbs. There are 28 539 named entity (NE) annotations spread over 15 UMLS semantic groups, one UMLS semantic type, and the Person semantic category. The most frequent annotations belong to the UMLS semantic groups of Procedures (15.71%), Disorders (14.74%), Concepts and Ideas (15.10%), Anatomy (12.80%), Chemicals and Drugs (7.49%), and the UMLS semantic type of Sign or Symptom (12.46%). Inter-annotator agreement results: Treebank (0.926), PropBank (0.891–0.931), NE (0.697–0.750). The part-of-speech tagger, constituency parser, dependency parser, and semantic role labeler are built from the corpus and released open source. A significant limitation uncovered by this project is the need for the NLP community to develop a widely agreed-upon schema for the annotation of clinical concepts and their relations.

This project takes a foundational step towards bringing the field of clinical NLP up to par with NLP in the general domain. The corpus creation and NLP components provide a resource for research and application development that would have been previously impossible.

Between June and November 2010, we conducted a cross-sectional, web-based survey of all primary care practitioners (PCPs) within the Department of Veterans Affairs nationwide. Survey development was guided by a socio-technical model describing multiple inter-related dimensions of EHR use.

Of 5001 PCPs invited, 2590 (51.8%) responded. 55.5% believed that the EHRs did not have convenient features for notifying patients of test results. Over a third (37.9%) reported having staff support needed for notifying patients of test results. Many relied on the patient's next visit to notify them for normal (46.1%) and abnormal results (20.1%). Only 45.7% reported receiving adequate training on using the EHR notification system and 35.1% reported having an assigned contact for technical assistance with the EHR; most received help from colleagues (60.4%). A majority (85.6%) stayed after hours or came in on weekends to address notifications; less than a third reported receiving protected time (30.1%). PCPs strongly endorsed several new features to improve test result management, including better tracking and visualization of result notifications.

Despite an advanced EHR, both social and technical challenges exist in ensuring notification of test results to practitioners and patients. Current EHR technology requires significant improvement in order to avoid similar challenges elsewhere.

We double annotated 3000 clinical trial announcements (CTA) and 1655 clinical notes (CN) for medication named entities and their attributes. A binary support vector machine (SVM) classification method with parsimonious feature sets, and a conditional random fields (CRF)-based multi-layered sequence labeling (MLSL) model were proposed to identify the linkages between the entities and their corresponding attributes. We evaluated the system's performance against the human-generated gold standard.

The experiments showed that the two machine learning approaches performed statistically significantly better than the baseline rule-based approach. The binary SVM classification achieved 0.94 F-measure with individual tokens as features. The SVM model trained on a parsimonious feature set achieved 0.81 F-measure for CN and 0.87 for CTA. The CRF MLSL method achieved 0.80 F-measure on both corpora.

We compared the novel MLSL method with a binary classification and a rule-based method. The MLSL method performed statistically significantly better than the rule-based method. However, the SVM-based binary classification method was statistically significantly better than the MLSL method for both the CTA and CN corpora. Using parsimonious feature sets both the SVM-based binary classification and CRF-based MLSL methods achieved high performance in detecting medication name and attribute linkages in CTA and CN.

Across 30 overnight shifts, questions asked of on-call senior residents, question askers’ roles, and residents’ responses were documented. External sources were noted.

158 of 397 questions (39.8%) related to the plan of care, 53 (13.4%) to medical knowledge, 48 (12.1%) to taskwork knowledge, and 44 (11.1%) to the current condition of patients. For 351 (88.4%) questions residents provided specific, direct answers or visited the patient. For 16 of these, residents modeled or completed the task. For 216 questions, residents used previous knowledge or their own clinical judgment. Residents solicited external information sources for 118 questions and only a single source for 77 (65.3%) of them. For the 118, most questions concerned either the plan of care or the patient's current condition and were asked by interns and nurses (those with direct patient care responsibilities).

Resident physicians serve as an information system and they often specifically answer the question using previous knowledge or their own clinical judgment, suggesting that askers are contacting an appropriately knowledgeable person. However, they do need to access patient information such as the plan of care. They also serve an educator role and answer many knowledge-related questions.

As synchronous verbal communications continue to be important pathways for information flow, informaticians need to consider the relationship between such communications and workflow in the development of healthcare support tools.

2241 patients were selected with a fall-related ICD-9-CM E-code or matched injury diagnosis code while being treated as an outpatient at one of four sites within the Veterans Health Administration. All clinical documents within a 48-h window of the recorded E-code or injury diagnosis code for each patient were obtained (n=26 010; 611 distinct document titles) and annotated for falls. Logistic regression, support vector machine, and cost-sensitive support vector machine (SVM-cost) models were trained on a stratified sample of 70% of documents from one location (dataset A_train) and then applied to the remaining unseen documents (datasets A_test–D).

All three STM models obtained area under the receiver operating characteristic curve (AUC) scores above 0.950 on the four test datasets (A_test–D). The SVM-cost model obtained the highest AUC scores, ranging from 0.953 to 0.978. The SVM-cost model also achieved F-measure values ranging from 0.745 to 0.853, sensitivity from 0.890 to 0.931, and specificity from 0.877 to 0.944.

The STM models performed well across a large heterogeneous collection of document titles. In addition, the models also generalized across other sites, including a traditionally bilingual site that had distinctly different grammatical patterns.

The results of this study suggest STM-based models have the potential to improve surveillance of falls. Furthermore, the encouraging evidence shown here that STM is a robust technique for mining clinical documents bodes well for other surveillance-related topics.

Text from 4409 clinical encounters for 33 men enrolled in a 15-year longitudinal clinical/molecular autopsy study of metastatic prostate cancer (Project to ELIminate lethal CANcer) was subjected to natural language processing (NLP) using Unified Medical Language System-based terms. A four-tiered pain scale was developed, and logistic regression analysis identified factors that correlated with experience of severe pain during each month.

NLP identified 6387 pain and 13 827 drug mentions in the text. Graphical displays revealed the pain ‘landscape’ described in the textual records and confirmed dramatically increasing levels of pain in the last years of life in all but two patients, all of whom died from metastatic cancer. Severe pain was associated with receipt of opioids (OR=6.6, p<0.0001) and palliative radiation (OR=3.4, p=0.0002). Surprisingly, no severe or controlled pain was detected in two of 33 subjects’ clinical records. Additionally, the NLP algorithm proved generalizable in an evaluation using a separate data source (889 Informatics for Integrating Biology and the Bedside (i2b2) discharge summaries).

Patterns in the pain experience, undetectable without the use of NLP to mine the longitudinal clinical record, were consistent with clinical expectations, suggesting that meaningful NLP-based pain status monitoring is feasible. Findings in this initial cohort suggest that ‘outlier’ pain phenotypes useful for probing the molecular basis of cancer pain may exist.

The results are limited by a small cohort size and use of proprietary NLP software.

We have established the feasibility of tracking longitudinal patterns of pain by text mining of free text clinical records. These methods may be useful for monitoring pain management and identifying novel cancer phenotypes.

We evaluated three areas of impact: length of stay, transcription time, and laboratory use. We collected data on expenditures in these categories under the paper-based (pre-EMR) system, and then estimated reductions in each category based on findings from EMR systems in the USA and backed by ambulatory data from low-income settings. We compared these potential savings accrued over a period of 5 years with the costs of implementing the touchscreen point-of-care EMR system at that site.

Estimated cost savings in length of stay, transcription time, and laboratory use totaled US$284 395 annually. When compared with the costs of installing and sustaining the EMR system, there is a net financial gain by the third year of operation. Over 5 years the estimated net benefit was US$613 681.

Despite considering only three categories of savings, this analysis demonstrates the potential financial benefits of EMR systems in low-income settings. The results are robust to higher discount rates, and a net benefit is realized even under more conservative assumptions.

This model demonstrates that financial benefits could be realized with an EMR system in a low-income setting. Further studies will examine these and other categories in greater detail, study the financial effects at different levels of organization, and benefit from post-implementation data. This model will be further improved by substituting its assumptions for evidence as we conduct more detailed studies.

The proposed approach combines information retrieval and semantic information extraction techniques to identify relevant sentences from Medline abstracts. We assessed this approach in two case studies on the treatment alternatives for depression and Alzheimer's disease.

A total of 515 of 564 (91.3%) sentences retrieved in the two case studies were relevant to the topic of interest. About one-third of the relevant sentences described factual knowledge or a study conclusion that can be used for supporting information needs at the point of care.

The high rate of relevant sentences is desirable, given that clinicians’ lack of time is one of the main barriers to using knowledge resources at the point of care. Sentence rank was not significantly associated with relevancy, possibly due to most sentences being highly relevant. Sentences located closer to the end of the abstract and sentences with treatment and comparative predications were likely to be conclusive sentences. Our proposed technical approach to helping clinicians meet their information needs is promising. The approach can be extended for other knowledge resources and information need types.

We evaluated our system on biomedical WSD datasets and determined the contribution of our WSD system to clinical document classification on the 2007 Computational Medicine Challenge corpus.

Our system compared favorably with other knowledge-based methods. Machine learning classifiers trained on disambiguated concepts significantly outperformed those trained using all concepts.

We developed a WSD system that achieves high disambiguation accuracy on standard biomedical WSD datasets and showed that our WSD system improves clinical document classification.

We integrated our WSD system with MetaMap and the clinical Text Analysis and Knowledge Extraction System, two popular biomedical natural language processing systems. All codes required to reproduce our results and all tools developed as part of this study are released as open source, available under http://code.google.com/p/ytex.

Experimental materials used in this study were provided by the 2011 i2b2 Natural Language Processing Challenge. The 2011 i2b2 challenge involves co-reference resolution in medical documents. Concept mentions have been annotated in clinical texts, and the mentions that co-refer in each document are linked by co-reference chains. Normally, there are two ways of constructing a system to automatically discover co-referent links. One is to manually build rules for co-reference resolution; the other is to use machine learning systems to learn automatically from training datasets and then perform the resolution task on testing datasets.

The existing co-reference resolution systems are able to find some of the co-referent links; our rule based system performs well, finding the majority of the co-referent links. Our system achieved 89.6% overall performance on multiple medical datasets.

Manually crafted rules based on observation of training data is a valid way to accomplish high performance in this co-reference resolution task for the critical biomedical domain.

We compared the performance of two biomedical concept normalization systems, MetaMap and Peregrine, on the Arizona Disease Corpus, with and without the use of a rule-based NLP module. Performance was assessed for exact and inexact boundary matching of the system annotations with those of the gold standard and for concept identifier matching.

Without the NLP module, MetaMap and Peregrine attained F-scores of 61.0% and 63.9%, respectively, for exact boundary matching, and 55.1% and 56.9% for concept identifier matching. With the aid of the NLP module, the F-scores of MetaMap and Peregrine improved to 73.3% and 78.0% for boundary matching, and to 66.2% and 69.8% for concept identifier matching. For inexact boundary matching, performances further increased to 85.5% and 85.4%, and to 73.6% and 73.3% for concept identifier matching.

We have shown the added value of NLP for the recognition and normalization of diseases with MetaMap and Peregrine. The NLP module is general and can be applied in combination with any concept normalization system. Whether its use for concept types other than disease is equally advantageous remains to be investigated.

In this work, we developed a novel network-based approach to identify putative causal module biomarkers of complex diseases by integrating heterogeneous information, for example, epigenomic data, gene expression data, and protein–protein interaction network. We first formulated the identification of modules as a mathematical programming problem, which can be solved efficiently and effectively in an accurate manner. Then, we applied our approach to colorectal cancer (CRC) and identified several network modules that can serve as potential module biomarkers for characterizing CRC. Further validations using three additional gene expression datasets verified their candidate biomarker properties and the effectiveness of the method. Functional enrichment analysis also revealed that the identified modules are strongly related to hallmarks of cancer, and the enriched functions, such as inflammatory response, receptor and signaling pathways, are specific to CRC.

Through constructing a transcription factor (TF)-module network, we found that aberrant DNA methylation of genes encoding TF considerably contributes to the activity change of some genes, which may function as causal genes of CRC, and that can also be exploited to develop efficient therapies or effective drugs.

Our method can potentially be extended to the study of other complex diseases and the multiclassification problem.

To perform the expansions we utilized the hierarchical structure by which the MeSH descriptors are organized. Two strategies for selecting the terms to be expanded in each query were studied. The first consisted of identifying the medical concepts using the unified medical language system metathesaurus. In the second strategy the text of the query was divided into n-grams, resulting in sequences corresponding to MeSH descriptors.

For the evaluation of the system, we used the collection made available by the ImageCLEF organization in its 2011 medical image retrieval task. The main measure of efficiency employed for evaluating the techniques developed was the mean average precision (MAP).

Both strategies exceeded the average MAP score in the ImageCLEF 2011 competition (0.1644). The n-gram expansion strategy achieved a MAP of 0.2004, which represents an improvement of 21.89% over the average MAP score in the competition. On the other hand, the medical concepts expansion strategy scored 0.2172 in the MAP, representing a 32.11% improvement. This run won the text-based medical image retrieval task in 2011.

Query expansion exploiting the hierarchical structure of the MeSH descriptors achieved a significant improvement in image retrieval systems.