Inside this Issue:

Download the Newsletter

Editors:

Joseph Cummings
Charles Dizard

Informatics Today is a publication of the University of Pittsburgh Department of Biomedical Informatics

The University of Pittsburgh is an affirmative action, equal opportunity institution. Published in cooperation with the Department of University Marketing Communications.

UMC65727-1008

Program Feature: The Clinical Research Informatics Service

Posters from the Clinical and Translational Sciences Institute (CTSI) on display throughout the University of Pittsburgh describe a common research problem: "We need extensive de-identified clinical data from the electronic health records on patients ... for an outcomes study." Health sciences researchers have successfully solved this problem by using the Clinical Research Informatics Service (CRIS), directed by Melissa Saul, MS, who also is an adjunct assistant professor of health information management in the School of Health and Rehabilitation Sciences.

CRIS was established in 2001 by Charles Friedman, PhD, director of the Center for Biomedical Informatics from 1996 to 2003, and Clifford Schold Jr, MD, associate vice chancellor for clinical research from 2001 to 2005. Randall Smith, PhD, senior associate dean of the School of Pharmacy, also was instrumental in developing the business model. CRIS currently is sponsored by the University's Department of Biomedical Informatics (DBMI) and CTSI.

Saul, the founding director of CRIS, has more than 20 years of experience working with health information systems. Saul was an early developer of the Medical Archival Retrieval System (MARS), the UPMC data repository for clinical and financial data. MARS currently contains more than 190 million reports and is indexed on every word in every report, enabling researchers to search millions of reports based on a clinical term or phrase.

"CRIS taps UPMC's enormous data sources and makes the data available to clinicians and health sciences faculty in an integrated and timely fashion," says Saul. "We make this data available in accordance with institutional and governmental regulations."

The University's Institutional Review Board (IRB) has certified CRIS as an "honest broker" of electronic medical record data. An honest broker serves as a disinterested intermediary between the researcher and the patients whose data are being studied. The use of an honest broker ensures that the investigator neither interacts with patients, nor records any identifiable patient information. However, the added value that CRIS provides is the use of De-IDTM, a software application that removes patient-identifying information from clinical reports. Using an automated tool to remove patient identifiers and assign pseudoidentifiers enables CRIS to process thousands of records from multiple sources into one data set for the clinical researcher. This service makes CRIS unique in its product offerings.

CRIS provides health sciences researchers with an extensive range of services, including:

Data Identification: CRIS identifies data sources supporting the researcher's request, examines the limitations of the data, identifies variables of interest, and refines the scope of the project as needed.

Data Collection: CRIS develops extraction methodologies using MARS and other data sources to acquire the data and complete quality assurance testing.

Data De-Identification: De-ID software—developed at the University by Saul and Greg Cooper, MD, PhD, an associate professor of biomedical informatics, computer science, and intelligent systems—allows researchers to use protected health information collected for patient care purposes for studies.

Data analysis: Data analysis is available for post-processing output for summary statistics and other customized reports.

The CRIS team consists of many collaborators from the University and UPMC. Within DBMI, software developer Lena Gemmer provides programming support for the data de-identification tools and other data management projects. Andrew Pople, a senior software architect at DBMI, provides oversight on the software development process for De-ID software.

The Cancer Text Information Extraction System (CATies) group—Rebecca Crowley, MD, MSIS, Girish Chaven, and Kevin Mitchell—work with CRIS to identify new functionality that could be added to the De-ID system.

Steven Handler, MD, MS and Shyam Visweswaran, MD, PhD, both assistant professors of biomedical informatics, collaborate with CRIS in their research into adverse event detection in the nursing home setting.

From the School of Pharmacy, Kim Coley, PharmD, FCCP, director of the Center for Pharmcoinformatics and Outcomes, and Robert Weber, MS, FASHP, chairman of the Department of Pharmacy and Therapeutics, work with CRIS to engage new researchers at the University, as well as pharmaceutical corporations.

CRIS is working closely with the CTSI informatics group to further develop the Clinical and Research Data Informatics Foundation Framework (CARDIFF.) One goal of this project is to provide clinical researchers the ability to come to CRIS for de-identified clinical data and have the data integrated into an existing data model so that the researcher can begin data analysis. This would allow the researcher to use the data sets obtained from CRIS in a more streamlined manner, as well as promote data sharing. In this framework, data that already have been collected and stored within a CARDIFF database are able to be reused for a future research project. The data de-identification process would also be a component of CARDIFF.

For more information about CRIS, visit www.dbmi.pitt.edu/cris and www.ctsi.pitt.edu.