|Data Warehouse and Enterprise (reported by Howard Shang)
DEDUCE (Duke Enterprise Data Unified Content Explorer)
Users will benefit from:
Work also proceeds on integrating data from Epic into DEDUCE, which will continue to contain historic data as far back as 1996 and add new availability of Maestro Care data. Epic Healthcare is the software company providing the platform for the integrated electronic health record (EHR) being deployed within Duke Medicine. The Duke-specific implementation of the Epic platform is called Maestro Care. The first implementation of Maestro Care occurred in July with the Wave 1 ambulatory go-live. The next milestone will occur on 10 October with the Wave 2 go-live.
Analytic Workspace Server (AWS)
Since mid-summer, the Data Warehouse has been hosting a new service called the Analytic Workspace Server (AWS), a virtual desktop that provides access to analytic software and storage of data. The AWS allows investigators and analysts to perform their analyses within a secured environment, and prevents the need to copy PHI or other sensitive data to potentially unsecure environments outside of the DHTS firewall. The AWS has been fully configured with installations of analytical software (including SAS, R, and ArcGIS), and is actively being used.
HDWA Abstract Acceptances
Presentation: “Geospatially Enabling Duke’s Enterprise Data Warehouse Tool”
Poster: “Connecting the Dots Between Text and Data”
Poster: “Conceptual Model for Research-Driven Data Marts”
In addition, Information Architect Stephanie Brinson was invited to join a roundtable discussion on “Research Data Warehousing” and Howard Shang was invited to lead a roundtable discussion on “New Data Delivery Methods.” Four members of the Data Warehouse team will travel to Ottawa, Ontario to participate in the conference.
See appendices for usage statistics for DEDUCE/Discern.
|The Human Studies Database (HSDB) Project (reported by Swati Chakraborty)
The HSDB project is a federated database for the design and results of human studies from multiple institutions. HSDB uses Ontology of Clinical Research (OCRe) and common clinical vocabularies to standardize information collection and maintenance. The participants are from CTSAs from the University of California-San Francisco, Duke University, Johns Hopkins, Stanford, The Rockefeller University, Mayo Clinic, University of California- Davis, University of Texas Southwestern, University of Texas Health Science Center-San Antonio, University of Washington, among others. For further details see HSDB wiki – http://hsdbwiki.org/index.php/HSDB_Wiki.
Duke University has led the administrative subgroup to produce a list of about 135 common administrative data elements identified from the CTSA participants’ Institutional Review Boards (IRBs) and Clinical Trial Management Systems (CTMS). Duke also led the harmonization with Biomedical Research Integrated Domain Group (BRIDG) model of the semantics of protocol-driven clinical research. Release 3.1 of the model is publicly available for download at http://bridgmodel.nci.nih.gov/download_model. Along with other leaders like Johns Hopkins University and The Rockefeller University, Duke has participated in developing strategy for local implementation.
Work from the HSDB Project has been presented at national meetings including the TBI-CRI joint summit, AMIA annual symposium, and SCTS/ACRT meeting; all the published materials are available from the HSDB wiki. An upcoming session in November at the AMIA 2012 Annual Symposium in Chicago titled “Ontology-Based Federated Data Access to Human Studies Information” will be presented. Swati Chakraborty from Duke is a co-author along with Ida Sim (University of San Francisco) and other CTSA participants.
|Data Standards Projects (Controlled Vocabulary) (reported by Anita Walden)
Tuberculosis Data Standards
The ORISE (Oak Ridge Institute for Science and Education) has identified data elements for pediatric pulmonary and extra pulmonary TB. An ORISE fellow, the Duke project leader and the FDA project officer are meeting bi-weekly to review data elements and discuss overlap with adult TB data elements. A group of clinical stakeholders have agreed to participate on the Clinical Review Committee and will hold their first meeting in early October.
With help from an intern from the Masters of Management Clinical Informatics program (MMCi), work has been completed on resolution of the critical errors for the model submitted to the National Cancer Institute (NCI) data element repository. The model has been submitted to the Enterprise Vocabulary Services team to create new data elements not currently in the thesaurus.
Anesthesiology Preoperative Domain Analysis Model (DAM)
The Anesthesiology working group had to withdraw their intent to ballot the preoperative data standards due to lack of resources to complete the changes to the data elements and the class model in time for the final content submission. The majority of the changes have now been made to the DAM; the goal is to ballot in January 2013.
Biorepository Data Elements
The biorepository data element development initiative led by Helena Ellis and the oversight committee consisting of Mary Beth Joshi, Aenoch Lynn and Anita Walden kicked off its cross campus initiative in July. There are four enthusiastic working groups, each made up of a facilitator and an informatics leader, who are developing common data elements for the biorepository. A monthly facilitators-oversight committee meeting was held at the beginning of September to review the working groups’ progress. The informatics leaders also meet monthly to review the data elements identified and promote harmonization across the working groups.
FDA R24 Grants
|Duke Office of Clinical Research (DOCR) (reported by Denise Snyder and Cory Ennis)
The Research Management Team (RMT) and the Clinical Research Support Office (CRSO) have merged into a new office called the Duke Office of Clinical Research (DOCR) now located on the 11th floor of the North Carolina Mutual Life Building. Several staff have been assigned new roles for DOCR. SBRs also have a new name: Clinical Research Units (CRUs). If you have a question or need assistance on a clinical research question, please email DOCR.email@example.com.
DOCR continues to partner with the Information Security Office (ISO) on data security vulnerabilities. The ISO is planning to establish a project governance committee with input from multiple entities. This will enable a transparent process for establishing the policies and workflows used to manage security incidents. A representative from DOCR will participate in this governance process.
DOCR is working with OIT and SOM Finance to investigate an alternative method of securely capturing social security numbers directly into SAP for study participant payments and reporting.
REDCap will be utilized to request retrospective Research Data Security Plans (RDSPs) from studies started before 21 November 2011. Approximately 2,000 surveys have gone out to researchers to date. Additional details regarding the plan will be released via the CRU Research Practice Managers.
REDCap is being migrated to a 3-tier structure to enhance security. A proxy server will be available from the Internet and the REDCap application server will join the database server behind the PHI firewall. Traffic will be monitored and controlled as it flows through the proxy server. In addition, the database server is being migrated to a 64-bit environment to improve performance. Following these changes, the system will be upgraded to the latest stable version of the software.
Results of the Duke Information Security Office (ISO) system penetration test on REDCap have been shared with the CTSA consortium members.
New procedures for clinicaltrials.gov registrations are being implemented for clinical trials where the investigator holds an IND/IDE for the study. Clinicaltrials.gov registrations for these studies will now list the Responsible Party as “Sponsor-Investigator”, meaning the investigator has the authority and responsibility to approve and release the registration record for publishing to clinicaltrials.gov. For current clinical trial registrations where the investigator holds the IND or IDE, the Responsible Party field is being updated and investigators contacted to make them aware of this change. For clinical trials not conducted under an IND or IDE, the Responsible Party field will continue to be listed as “Sponsor” and the Sponsor listed as “Duke University”. DOCR is posting three FTEs this week to assist study teams with results reporting.
DOCR is working with the Office of Corporate Research Collaborations (OCRC) to finalize a contract with ResearchMatch, a national registry to match prospective study volunteers with studies they might qualify for. Denise Snyder will serve as the primary contact for approving the initial registration of a Duke investigator, with backup from Cory Ennis and Shelly Epps. DOCR will work closely with DCRU to integrate ResearchMatch with DCRU subject recruitment.
eIRB 12.1 was released in July. eIRB 12.2 is targeted for release at the end of September. Planned updates include the ability to edit previously entered Research Data Security Plans (RDSPs), a question asking if the study will utilize biospecimens, branding updates to reflect the change of organization to DOCR, and various application fixes. DOCR IT staff has implemented in-depth system monitoring to track and address ongoing performance issues with the application.
In its first step to address the need for verified disaster recovery capabilities, DOCR IT staff now has a staging instance online and has instituted server mirroring to copy the production environment to a backup environment.
For detailed metrics, see Metrics for Data and Research Management Support.
|Velos (reported by Cory Ennis, Denise Snyder, Matt Gardner, Julie Eckstrand)
A new test environment has been created and Velos v9.0 installed for evaluation and testing. Discussion regarding future direction of the CTMS is ongoing. The DOCR IT team is working with Velos developers to address data requirements needed to potentially retire the Subject Billing Registry legacy application. New code has been delivered and tested. Velos is working on issues identified by DOCR IT. Once the technical solution is functioning as expected, DOCR will work with business stakeholders currently utilizing the Subject Billing Registry to address their needs.
Development of a technology solution to support electronic acquisition of early phase research data and operational tasks at the Duke Clinical Research Unit (DCRU), which began in 2010, is underway as a Phase 1 suite of applications. A Phase 1 Core product will allow for deconstruction/translation of a protocol (electronic source document) and will interact with volunteer management, scheduling and resource management of research unit space, equipment and personnel, biospecimen chain of custody management, and ultimately a comprehensive electronic study plan/source document. The system will leverage existing functionality and core services built within the eResearch CTMS, and include custom built pieces which cover the functionality gaps. It interacts with larger Duke IT infrastructure such as the Master Patient Index, eIRB, and Duke eResearch. Over time information will be pulled from and pushed to other IT systems as required for operational efficiency, including communicating with the EHR (EPIC Research), increased interactions with the CTMS (eResearch), and many other research IT systems.
DCRU is part of a Global Proof of Concept Network (Singapore IMU and MDRI in India), and expects to share protocol configuration with its global partners. This will allow for real-time access to integrated, multi-site, early phase clinical trial data by sponsors and study teams, global alerting, remote sponsor monitoring, and construction of a global repository of data for secondary research purposes.
|Duke Bioinformatics Shared Resources (DBSR) (reported by Salvatore Mungal)
GenePattern and caArray were deployed in 2010 and are widely used to meet the data storage and data sharing needs of Duke Cancer Institute (DCI) investigators with high-dimensional gene expression datasets. The DCI Bioinformatics unit has launched an aggressive campaign to build GenePattern modules to meet the requirements of reproducible research. Currently, a total of 22 new GenePattern modules have been developed, ranging from data management and preprocessing to data analysis and visualization. The modules have been piloted for analysis of existing mRNA microarray datasets in breast and ovarian cancer from several DCI investigators, and externally from The Cancer Genome Atlas (TCGA) project.
A URL has been established and security clearance obtained to use a special port existing outside the firewall to access the RProteomics software application to caGrid. Implementation and deployment of the software to the training grid is underway. When completed, implementation data, translation, and analytical grid services for proteomics mass spectrometry using provenance and the latest security measures will facilitate reproducible research. It will also extend the functionality of existing grid technology to make sharing data and applications tractable.
These applications use NCI’s caBIG® syntactic, semantic and data standards to facilitate interoperability. They will use the National Cancer Informatics Program’s (NCIP) formal standards when they become available.
Next steps will include evaluation of existing data standards, recommendation for omics-related data elements to be used in the context of the Duke Biobank, and recommendation of standards to be used for the MIDR and other types of initiatives that fall along a spectrum of required structure and rigor.
|Enterprise Biobanking Informatics (reported by Helena Ellis)
The Duke Biobank continues to work towards obtaining institutional funding to support implementation and deployment of Labvantage as Duke’s enterprise-wide Biobank Information Management System to be used by all Duke clinical researchers who use or hold human biological specimens. A detailed 5-year business plan was reviewed by DHTS’s financial group, which includes implementation and migration costs for the major biobanking groups at Duke. The proposal has been reviewed and approved by the DukeQuEST committee (Quality and Excellence in Scientific Translation, formerly TMQF), and a request for funding will be brought to the Chancellor’s office in September 2012. Additional information and links to resources can be found on the Duke Biobank website (http://biobank.duke.edu/).
Index of Biospecimens
The Duke Biobank Index of Biospecimens, also known as “the Index” (https://sites.duke.edu/biobank/), is a simple web-based searchable database for Duke researchers to look for biospecimen collections that may be available for collaboration at Duke. FAQ’s for the Index are available on the Duke Biobank website (https://biobank.duke.edu/resources/duke-index-of-biospecimens-the-index-faq). The Duke Biobank is actively working with many researchers to register more collections in the Index and will assist biobanks or investigators who have collections they wish to register. Please contact the biospecimen-expediter <firstname.lastname@example.org> for details.
Institutional Biobanking Terminology
On 13 July 2012 the Duke Biobank and members of the Biobank Informatics Working Group launched an initiative to establish standard biobanking terminology. This biobanking terminology, which will focus on essential biobanking terms and data elements, will be implemented in the LabVantage system during the deployment of the system and will be made available to all biobanking groups. The standardization of biobanking terminology will help promote semantic interoperability across all biobanking and clinical research groups at Duke. Terminology working groups consisting of individuals with biobanking expertise have been formed and are meeting regularly. The project is expected to last up to one year. Interested parties should contact Helena Ellis <email@example.com> for more information.
|MURDOCK Informatics (reported by Jessie Tenenbaum, Doug Wixted, and Anita Walden)
MIDR i2b2 deployment
A proof-of-concept i2b2 instance is being developed for the MIDR to evaluate i2b2 as a technology solution moving forward. Loading metadata into the i2b2 instance has been completed. Additional progress has been largely on hold while the developer’s effort was needed on other tasks, but as of 5 September, work has re-commenced. The first task will be loading a 1,000-row selected subset of registry data. Next steps after that will include review of the metadata structure by the MIDR team, design and implementation of a streamlined process for future data transfer, and performance testing with a more realistic number of records.
A confluence of separate but related initiatives are coming together at an opportune time for next steps with respect to incorporation of omic data in the MIDR:
A revised version of the MURDOCK Study Governance document has been completed based on input from a number of stakeholders. It was distributed to MURDOCK Leadership and will be discussed in an upcoming Leadership meeting. This document describes in detail the orchestration of people, processes, technology, and policy to leverage, optimize, and maximize the value of both MURDOCK data and other aspects of the Study, including biospecimens, infrastructure, and human expertise. In particular, two key areas the document addresses are data governance- how data is generated/acquired, stored, and accessed using role-based permissions, and MURDOCK Study research proposals- how sub-studies are to be proposed, reviewed, approved, and executed.
In parallel, a distinct IRB Protocol for the MIDR itself (as opposed to references to it in the Horizon 1 and Horizon 1.5 protocols) has been developed and is being submitted for IRB approval. Having this separate protocol is necessary to enable use of the MIDR (which holds a limited data set) for review preparatory to research.
The MURDOCK Registry development team is working to provide Registry participants with the option of completing their enrollment forms online and submitting them to the Kannapolis office prior to their enrollment visit. Completing enrollment forms and any follow-up forms via the web is expected to increase enrollment and improve follow-up compliance.
Progress on the web enrollment functionality completed since the last iteration includes:
It is anticipated that testing the developed items and evaluation of the tools needed to code the medications will take three weeks. Results will be reviewed by the team at their meeting on 18 September.
The project team is reviewing the heuristics test results and the user acceptance test results. The findings will be ranked in order of severity and then discussed with the developers for implementation. The public will also participate in the testing to ensure the tool is user friendly, intuitive, and provides the information needed. The Internal Review Board documents are being reviewed by the Kannapolis office prior to being submitted for approval.
Horizon 2 Data Transfers
Access and procedures have been established to enable investigators/staff for the Severe Acne and Physical Performance Studies to access pdf copies of CRFs that Kannapolis study staff have scanned and reside on a secure FTP site. In addition to CRFs, files from Actigraph accelerometers (used in the Physical Performance Study) are being transferred in this manner.
Systems have been identified and will be evaluated for coding the medications collected in the registry, a step needed for analysis and reporting. Initial investigation has been done on the National Center for Biomedical Ontology (NCBO) tools using the RxNorm terminology for coding both legacy free text data and data entry moving forward. A research side project will evaluate the accuracy rate for coding existing data and characterize the types of entries that cannot be mapped through the tool.
RXNORM, which is produced by the National Library of Medicine, has an annotator specifically used for coding medications and is also being evaluated to determine accuracy and ease of use and implementation.
Electronic Health Record (EHR) Data
MURDOCK Study informaticians are working with the two major healthcare providers in Kannapolis/Cabarrus County, Carolinas Healthcare System (CHS; source of primary care for 55% of current participants) and Novant Health (source of primary care for 10% of participants) to share electronic health record (EHR) data for MURDOCK participants.
The CHS data warehouse and analytics team (DA2, formerly the Dickson Institute) and Duke Data Warehouse group, with support from the MURDOCK Study, are planning a visit to Duke so CHS can learn more about building the Duke enterprise data warehouse [i.e. Decision Support Repository (DSR)].
Novant leadership is interested in using Epic as a common EHR platform between Novant and Duke, leveraging the interoperability between systems to share data securely for common patients. This would facilitate integration of clinical data from Novant for the MURDOCK Study, and may provide a secure method to do so. Doug Wixted, Informatics Project Leader for the MURDOCK Study, is working with Keith Griffin, Novant CMIO (MURDOCK champion and lead contact) and Novant compliance personnel regarding a Data Use Agreement (DUA) and related privacy and security considerations. Evaluation of the Epic Care Everywhere application for the MURDOCK Study is in nascent stages; however, efforts to determine feasibility have been initiated from both sides.
|Subscribe to Our Newsletter!|