Research

New Biomedical Informatics Tool Allows Self-Service Data Access for Researchers

of

LEXINGTON, Ky. (March 16, 2015) The UK Center for Clinical and Translational Science (CCTS) Biomedical Informatics Core is offering a new self-service informatics tool that lets researchers search UK HealthCare data to generate hypotheses and estimate cohorts available for clinical research studies. Informatics for Integrating Biology and the Bedside, or i2b2, is a searchable database of de-identified clinical data that allows users to select inclusion and exclusion criteria and produce aggregate counts within seconds.

i2b2 is one component of the CCTS Enterprise Data Trust (EDT), a collection of clinical data sets that can be used to expedite research at UK. It was created by the National Center for Biomedical Computing based at Partners HealthCare System and funded by a cooperative agreement with the National Institutes of Health (NIH). According to the NIH, the i2b2 platform has been adopted internationally and aims to create " a scalable informatics framework that will enable clinical researchers to use existing clinical data for discovery research and, when combined with IRB-approved genomic data, facilitate the design of targeted therapies for individual patients with diseases having genetic origins."

“One of the many uses of this is to help with research on less common and sometimes neglected diseases, said Philip Kern, MD, director of the CCTS.  If a scientist wants to study patients with acromegaly, cystic fibrosis, sarcoidosis, muscular dystrophy, scleroderma,   or any of a number of diseases that cause much suffering but which are relatively uncommon, this database will better allow a search for afflicted individuals."

Empowering researchers with a self-service informatics tool improves upon the old process in which they were dependent on data analysts to perform queries, an additional step that slowed the research process and distanced the researchers from the data they needed.

"i2b2 makes data access self-service," said Jeffrey Talbert, Phd, co-director of the CCTS Biomedical Informatics core.  "In the past, you would have to pull charts and count patients by hand. Then with electronic systems, we would run a query for you. But now that it's self-service, you get to explore the data on your own and play with the query until you find a population that meets your needs. It's a quick way to get cohort counts that you couldn't get before."

Steve Leung, MD, an assistant professor of medicine who specializes in cardiovascular imaging and medicine, has used i2b2 to determine populations for at least three projects.

"It allowed us to obtain patient list data much easier than previously," he said.  "For example, instead of going through 10,000+ echocardiograms performed in 2014 and selecting out the ones that had heart failure by hand, we can use i2b2 to quickly obtain this patient list by doing a simple search."


If an appropriate population is identified from an i2b2 search, researchers can then use this information to recruit participants for studies. Alternatively, an i2b2 query could demonstrate that a sufficient population isn't available to support a study, saving a researcher the time, expense, and effort of launching a study that would ultimately be unsuccessful. Leung has run several i2b2 queries that didn't result in projects because the data demonstrated that there wasn't a feasible population.

"The database allowed me to determine whether we have enough patients to perform the research, before we even start.   If the search determined that there were too few patients, then we would have to alter our research method.  On the other hand, if there are sufficient patients in the database, then we would continue planning out our project," Leung said.  "If it was a retrospective study, then we would submit an IRB and obtain the patient list based on the searches.  If it was a prospective study, we can potentially determine whether we can recruit enough patients within a reasonable amount of time for the study to be meaningful."


Researchers can now register for an i2b2 account and begin running their own queries. An online video tutorial and bimonthly online training sessions are available in addition to one-on-one assistance from the CCTS biomedical informatics core.

"If someone has sufficient knowledge of the system and the question is simple, then self-service query is a great advantage.  You can create many queries over a short period of time, and determine feasibility of a study quickly.  However, if there are complex questions, it is always helpful to have data analysts assist in those cases," said Leung.

Currently, i2b2 has several searchable dimensions available from the UK HealthCare inpatient clinical data, including demographics (age, gender, race, marital status); diagnoses (ICD9 standard); visit details (age at visit, length of stay, financial classification, service, visit type, weekend admission); vital signs (height, weight, BMI, direct arterial blood pressure, noninvasive blood pressure, heart rate, pulse oximetry, respiratory rate, temperature, death status, tobacco status); lab results; procedures (inpatient facility and technical procedures); CPT codes; and inpatient medications.

More data points are continually added to the database, with outpatient professional procedures, cost data, surgical data, pathology data, biospecimen availability and radiology data coming soon.

Researchers will not be required to obtain individual Institutional Review Board (IRB) approval for using this research tool. The IRB has approved an umbrella IRB application # 11-0750-F6A for the use of this tool to obtain patient aggregate counts.

The CCTS biomedical informatics team is working to connect UK's informatics infrastructure with that of other institutions to further improve and expedite research.  Data from St. Claire Regional Medical Center will be added to UK's EDT, and eventually data from other community and family practice sites.  The Shared Health Research Informatics NEtwork (SHRINE) will additionally allow UK's i2b2 to link with its counterparts at other universities. The UK team helped Marshall University establish its i2b2 and is currently working on linking the i2b2 platforms of UK and West Virginia University.

i2b2 is also being used in UK classrooms, including in the biomedical informatics program and pharmacy research.

"We're trying to help the next generation of researchers get used to self-service for interactive data analysis," said Talbert.

Media Contact: Mallory Powell, Mallory.powell@uky.edu