The Common Data Elements for cancer research: remarks on functions and structure.

OBJECTIVES: The National Cancer Institute (NCI) has developed the Common Data Elements (CDE) to serve as a controlled vocabulary of data descriptors for cancer research, to facilitate data interchange and inter-operability between cancer research centers. We evaluated CDE's structure to see whether it could represent the elements necessary to support its intended purpose, and whether it could prevent errors and inconsistencies from being accidentally introduced. We also performed automated checks for certain types of content errors that provided a rough measure of curation quality. METHODS: Evaluation was performed on CDE content downloaded via the NCI's CDE Browser, and transformed into relational database form. Evaluation was performed under three categories: 1) compatibility with the ISO/IEC 11179 metadata model, on which CDE structure is based, 2) features necessary for controlled vocabulary support, and 3) support for a stated NCI goal, set up of data collection forms for cancer research. RESULTS: Various limitations were identified both with respect to content (inconsistency, insufficient definition of elements, redundancy) as well as structure--particularly the need for term and relationship support, as well as the need for metadata supporting the explicit representation of electronic forms that utilize sets of common data elements. CONCLUSIONS: While there are numerous positive aspects to the CDE effort, there is considerable opportunity for improvement. Our recommendations include review of existing content by diverse experts in the cancer community; integration with the NCI thesaurus to take advantage of the latter's links to nationally used controlled vocabularies, and various schema enhancements required for electronic form support.

Automating identification of adverse events related to abnormal lab results using standard vocabularies.

Laboratory data need to be imported automatically into central Clinical Study Data Management Systems (CSDMSs), and abnormal laboratory data need to be linked to clinically related adverse events. This import of laboratory data can be automated through mapping to standard vocabularies with HL7/LOINC mapping to the metadata within a CSDMS. We have designed a system that uses the UMLS metathesaurus as a common source to map or link abnormal laboratory values to adverse event CTCAE coded terms and grades in the metadata of TrialDB, a generic CSDMS.

Metadata driven generation of reports for clinical studies.

We have developed a reporting engine to dynamically create Web-based reports for multiple types of clinical studies in a large-scale clinical study data management system (CSDMS), TrialDB. The engine, currently in production use, generates reports based on metadata that is defined through a GUI, thus eliminating the need to write code for every report.

Integration of Web-based and PC-based clinical research databases.

UNLABELLED: We have created a Web-based repository or data library of information about measurement instruments used in studies of multi-factorial geriatric health conditions (the Geriatrics Research Instrument Library - GRIL) based upon existing features of two separate clinical study data management systems. GRIL allows browsing, searching, and selecting measurement instruments based upon criteria such as keywords and areas of applicability. Measurement instruments selected can be printed and/or included in an automatically generated standalone microcomputer database application, which can be downloaded by investigators for use in data collection and data management. METHODS: Integration of database applications requires the creation of a common semantic model, and mapping from each system to this model. Various database schema conflicts at the table and attribute level must be identified and resolved prior to integration. Using a conflict taxonomy and a mapping schema facilitates this process. RESULTS: Critical conflicts at the table level that required resolution included name and relationship differences. CONCLUSIONS: A major benefit of integration efforts is the sharing of features and cross-fertilization of applications created for similar purposes in different operating environments. Integration of applications mandates some degree of metadata model unification.

Approaches and informatics tools to assist in the integration of similar clinical research questionnaires.

OBJECTIVE: The integration of similar clinical research questionnaires is a complex process that can benefit from informatics approaches and tools that provide a systematic structure for performing mapping and integration. This systematic approach is necessary to address complex issues in integration such as data heterogeneity, differing levels of granularity of questions and responses, and other issues involving semantic differences. Informatics tools and approaches have been successfully applied to various standard clinical vocabulary integration processes but not for questionnaire integration or mapping. METHODS: A systematic approach to questionnaire integration was developed in the context of a collaboration of researchers using Trial/DB, a database designed to support clinical research. This approach was applied to the integration of questionnaires involving breast cancer risk factors from each of three research sites. RESULTS: From 375 questions on the three original questionnaires, we identified 65 concepts that were measured by two or three of the sites. An algorithm was developed and used to formalize the process of mapping questions and answers across the questionnaires. The approach was applied to previously collected data and prospective data in disparate data-base systems to import and merge the data from these three sites into Trial/DB. CONCLUSION: Informatics tools that support a systematic approach to mapping questionnaires can be used throughout the research process from questionnaire integration and creation, legacy data integration to data library maintenance and curation.

TrialDB: A web-based Clinical Study Data Management System.

Clinical Study Data Management Systems (CSDMSs) are a class of software that support centralized management of data generated during the conduct of clinical studies. Commercial CSDMSs include Oracle Clinical, ClinTrial and MetaTrial. Such systems, which are typically deployed at an institutional or organizational level, must accommodate diverse types of data from different clinical domains that is generated by different groups of clinical investigators. Large-scale CSDMSs typically employ a high-end database engine that is usually accessed over an intranet or the Internet using Web-based technologies. CSDMSs in institution-wide use for a variety of clinical domains are best served by entity-attribute-value (EAV) modeling for the clinical data: all the commercial CSDMSs that we are aware of use EAV design. However, de novo development of EAV databases for data management is a challenging task. A large body of generic metadata-driven code must be developed before a basic EAV application can be written. Clearly, the availability of pre-existing software with the requisite functionality would be very valuable. We will discuss the benefits of such software being in open-source form.

An introduction to information retrieval: applications in genomics.

Information retrieval (IR) is the field of computer science that deals with the processing of documents containing free text, so that they can be rapidly retrieved based on keywords specified in a user's query. IR technology is the basis of Web-based search engines, and plays a vital role in biomedical research, because it is the foundation of software that supports literature search. Documents can be indexed by both the words they contain, as well as the concepts that can be matched to domain-specific thesauri; concept matching, however, poses several practical difficulties that make it unsuitable for use by itself. This article provides an introduction to IR and summarizes various applications of IR and related technologies to genomics.

Use of general-purpose negation detection to augment concept indexing of medical documents: a quantitative study using the UMLS.

OBJECTIVES: To test the hypothesis that most instances of negated concepts in dictated medical documents can be detected by a strategy that relies on tools developed for the parsing of formal (computer) languages-specifically, a lexical scanner ("lexer") that uses regular expressions to generate a finite state machine, and a parser that relies on a restricted subset of context-free grammars, known as LALR(1) grammars. METHODS: A diverse training set of 40 medical documents from a variety of specialties was manually inspected and used to develop a program (Negfinder) that contained rules to recognize a large set of negated patterns occurring in the text. Negfinder's lexer and parser were developed using tools normally used to generate programming language compilers. The input to Negfinder consisted of medical narrative that was preprocessed to recognize UMLS concepts: the text of a recognized concept had been replaced with a coded representation that included its UMLS concept ID. The program generated an index with one entry per instance of a concept in the document, where the presence or absence of negation of that concept was recorded. This information was used to mark up the text of each document by color-coding it to make it easier to inspect. The parser was then evaluated in two ways: 1) a test set of 60 documents (30 discharge summaries, 30 surgical notes) marked-up by Negfinder was inspected visually to quantify false-positive and false-negative results; and 2) a different test set of 10 documents was independently examined for negatives by a human observer and by Negfinder, and the results were compared. RESULTS: In the first evaluation using marked-up documents, 8,358 instances of UMLS concepts were detected in the 60 documents, of which 544 were negations detected by the program and verified by human observation (true-positive results, or TPs). Thirteen instances were wrongly flagged as negated (false-positive results, or FPs), and the program missed 27 instances of negation (false-negative results, or FNs), yielding a sensitivity of 95.3 percent and a specificity of 97.7 percent. In the second evaluation using independent negation detection, 1,869 concepts were detected in 10 documents, with 135 TPs, 12 FPs, and 6 FNs, yielding a sensitivity of 95.7 percent and a specificity of 91.8 percent. One of the words "no," "denies/denied," "not," or "without" was present in 92.5 percent of all negations. CONCLUSIONS: Negation of most concepts in medical narrative can be reliably detected by a simple strategy. The reliability of detection depends on several factors, the most important being the accuracy of concept matching.

Easing the transition between attribute-value databases and conventional databases for scientific data.

We have previously developed and described a modeling and development framework called EAV/CR, which is appropriate for designing databases containing highly heterogeneous and evolving data, as in the case of scientific databases for rapidly advancing domains. The use of EAV/CR has been hampered by the lack of generic tools for non-procedurally transferring data into or out of legacy systems or analytical packages: the transfer task is complicated by the different representation of EAV vs. conventional data, which is not addressed by commercial data-transfer programs. We have therefore created such a tool, which works with a wide variety of data sources that are accessible via Microsoft OLE DB technology. The data transfer tool requires minimal programmer intervention to set up, and no programming to use on a regular basis. Current limitations of the tool are also noted.

WebEAV: automatic metadata-driven generation of web interfaces to entity-attribute-value databases.

The task of creating and maintaining a front end to a large institutional entity-attribute-value (EAV) database can be cumbersome when using traditional client-server technology. Switching to Web technology as a delivery vehicle solves some of these problems but introduces others. In particular, Web development environments tend to be primitive, and many features that client-server developers take for granted are missing. WebEAV is a generic framework for Web development that is intended to streamline the process of Web application development for databases having a significant EAV component. It also addresses some challenging user interface issues that arise when any complex system is created. The authors describe the architecture of WebEAV and provide an overview of its features with suitable examples.

Using a computer database to monitor compliance with pharmacotherapeutic guidelines for schizophrenia.

OBJECTIVE: The study examined whether prescription data from a computerized database could be used to measure conformance with treatment recommendations of the Schizophrenia Patient Outcomes Research Team (PORT). METHODS: Records of an academically affiliated Veterans Affairs medical center were reviewed to identify patients who were hospitalized for schizophrenia and later seen for at least two outpatient visits in the six months after discharge (N=353). RESULTS: Conformance with only three of the 18 PORT pharmacotherapeutic recommendations could be measured with the available data. In regard to the recommendation to use antipsychotics other than clozapine as first-line treatments in acute episodes, 77 percent of the sample filled a prescription for an antipsychotic during the acute episode. Of these, only 6 percent received an antipsychotic regimen that included clozapine. In regard to the PORT recommendation on dosage during acute symptom episodes, 42 percent of the patients on conventional antipsychotics received dosages below the recommended range, 5 percent were above the range, and 53 percent were within it. In contrast, of the 53 patients who received clozapine or risperidone, 87 percent received prescriptions within the recommended dosage range. As for the recommendation to offer a trial of clozapine to patients who do not respond to adequate trials of two different classes of conventional drugs, 10 percent of patients who were switched from conventional regimens to clozapine were receiving dosages of conventional medications below the recommended range. CONCLUSIONS: Patient prescription data can provide preliminary measures to cost-effectively assess conformance with treatment. However, the approach has several limitations, and complementary analyses would enhance its usefulness.

Integration of a hematopoietic progenitor cell program using the ACT/DB database system.

Infusion of hematopoietic progenitor cells following high-dose chemotherapy is frequently used to treat patients with hematological malignancies and solid tumors. We have developed a comprehensive software system to monitor these patients once they are entered into an experimental protocol. The captured data encompasses all phases of progenitor cell therapy including progenitor cell mobilization and collection, stem cell processing, as well as cell infusion and engraftment kinetics. Particular attention was paid to the quality assurance and quality control functionality of the software during development of data entry forms and reports. The system was developed using the ACT/DB client-server database, which utilizes Microsoft Access as a front-end and accesses either an Oracle or SQL Server database. ACT/DB has been modified for deployment on the Internet in order to take advantage of Web-based technology. Information technology can help to integrate the diverse data requirements of complex therapeutic trials.

Olfactory receptor database: a sensory chemoreceptor resource.

The Olfactory Receptor Database (ORDB) is a WWW-accessible database that has been expanded from an olfactory receptor resource to a chemoreceptor resource. It stores data on six classes of G-protein-coupled sensory chemoreceptors: (i) olfactory receptor-like proteins, (ii) vomeronasal receptors, (iii) insect olfactory receptors, (iv) worm chemo-receptors, (v) taste papilla receptors and (vi) fungal pheromone receptors. A complementary database of the ligands of these receptors (OdorDB) has been constructed and is publicly available in a pilot mode. The database schema of ORDB has been changed from traditional relational to EAV/CR (Entity-Attribute-Value with Classes and Relationships), which allows the interoperability of ORDB with other related databases as well as the creation of intra-database associations among objects. This inter-operability facilitates users to follow information from odor molecule binding to its putative receptor, to the properties of the neuron expressing the receptor, to a computational model of activity of olfactory bulb neurons. In addition, tools and resources have been added allowing users to access interactive phylogenetic trees and alignments of sensory chemoreceptors. ORDB is available via the WWW at http://ycmi.med. yale.edu/senselab/ordb/

Organization of heterogeneous scientific data using the EAV/CR representation.

Entity-attribute-value (EAV) representation is a means of organizing highly heterogeneous data using a relatively simple physical database schema. EAV representation is widely used in the medical domain, most notably in the storage of data related to clinical patient records. Its potential strengths suggest its use in other biomedical areas, in particular research databases whose schemas are complex as well as constantly changing to reflect evolving knowledge in rapidly advancing scientific domains. When deployed for such purposes, the basic EAV representation needs to be augmented significantly to handle the modeling of complex objects (classes) as well as to manage interobject relationships. The authors refer to their modification of the basic EAV paradigm as EAV/CR (EAV with classes and relationships). They describe EAV/CR representation with examples from two biomedical databases that use it.

Olfactory Receptor Database: a database of the largest eukaryotic gene family.

The Olfactory Receptor Database (ORDB) is a WWW-accessible database that stores data on Olfactory Receptor-like molecules (ORs) and has been open to the public since June 1996. It contains a public and a private area. The public area includes published DNA and protein sequence data for ORs, links to OR models and data on their expression, chromosomal localization and source organism, as well as (i) links to bibliography through PubMed and (ii) interactive WWW-based tools, such as BLAST homology searching. The private area functions as a service to laboratories that are actively cloning receptors. Source laboratories enter the sequences of the receptor clones they have characterized to the private database and can search for identical or near identical OR sequences in both public and private databases. If another laboratory has cloned and deposited an identical or closely matching sequence there are means for communication between the laboratories to help avoid duplication of work. ORDB is available via the WWW at http://crepe.med.yale.edu/ORDB/HTML

CHRONOMERGE: an application for the merging and display of multiple time-stamped data streams.

CHRONOMERGE is a database application that facilitates merging and display of multiple time-stamped data streams. Each stream is a table containing time-stamped values of one or more parameters (such as a panel of laboratory tests) for multiple patients, and is typically created by querying a clinical data respiratory. The data within a single stream therefore represent a pool of multiple time series. The merge operation is complex because of the numerous options to be considered, such as the granularity of the time interval for merge, and the choice of statistical aggregates. CHRONOMERGE combines multiple streams into a single stream based on patient and time, or time alone (if aggregates are to be computed across patients). It allows specification of various options through a graphical user interface and generates appropriate SQL code (or invokes procedural routines) to perform the merge. The resultant stream, or subsets of it, can then be displayed graphically. CHRONOMERGE is intended to facilitate the analysis of time-stamped data that have been extracted from repositories when standard tools (such as the time-series modules of statistics packages) are inadequate.

Data extraction and ad hoc query of an entity-attribute-value database.

Entity--attribute--value (EAV) tables form the major component of several mainstream electronic patient record systems (EPRSs). Such systems have been optimized for real-time retrieval of individual patient data. Data warehousing, on the other hand, involves cross-patient data retrieval based on values of patient attributes, with a focus on ad hoc query. Attribute-centric query is inherently more difficult when data are stored in EAV form than when they are stored conventionally. The authors illustrate their approach to the attribute-centric query problem with ACT/DB, a database for managing clinical trials data. This approach is based on metadata supporting a query front end that essentially hides the EAV/non-EAV nature of individual attributes from the user. The authors' work does not close the query problem, and they identify several complex subproblems that are still to be solved.