Clinical Data Repositories: Boosting Patient Care & Research for a Data-Intensive Future

Clinical data repositories are allowing facilities to improve patient care and move toward an integrated and interoperable health IT environment. The HITECH Act has spurred a need for CDR development in both academic and large-scale healthcare settings, and health IT professionals are getting the technology in place with an eye toward the features that are important to a well-designed CDR.

The CDR development team at Medical University of South Carolina (MUSC) in Charleston tried to create a CDR that would cause as little disruption as possible and be easy to use for all practitioners at the academic medical center, which encompasses four hospitals and 1,200 physicians, and each year has 38,000 in-patient admissions and nearly 1 million outpatient visits. 

“We had the opportunity to create an enterprisewide environment where we aggregate clinical data across care settings,” says Frank Clark, PhD, CIO and vice president of the medical center. “It’s sort of a best-of-breed or an evolution of systems.”

MUSC’s EMR system includes Emergis’ Oacis CDR and Practice Partner, an outpatient documentation system from Physician Micro Systems. The MUSC system also uses Lanvision's AccessAnywhere document imaging system. Clinical information and results are integrated in Oacis from Practice Partner, AccessAnywhere, and other sources such as general laboratory, microbiology, pathology, radiology, respiratory therapy and multiple sources of transcription.

The system currently in place represents six or seven vendor products, Clark says. “We think of it as the center of the clinical IT environment and then we populated it with these other systems including lab, pharmacy, radiology, clinical documentation, medication administration and physician order entry,” he explains. 

MUSC can aggregate clinical data across all its care settings, including inpatient, outpatient and the emergency room, Clark says. In the inpatient setting, for example, the CDR collects data from all of its ancillary systems, including radiology, pharmacy and cardiology. In the outpatient setting, any data captured in the clinics can be aggregated back into the repository, he says. Having a CDR has enabled the center to create longitudinal patient records, providing more complete patient histories to clinicians with the click of a mouse.

A robust system that could handle simultaneous searches by multiple users was a priority for the center, explains Clark.

“We have tried to build in as much redundancy as possible and stay ahead of the demand curve as far as processing, so that we hopefully have a [maximum] response time of two or three seconds when caregivers are trying to pull up vitals, lab results or info on patients,” he says. 

In addition, reliability is a main component of a CDR in any clinical setting. “We have tried to build in safeguards because there’s one thing I can guarantee, there is going to be some downtime,” he explains. “What we’ve tried to do is make data available, and we will post some files out on a stand-alone printer. In case the system goes down, these printers can still print out patient rosters and some of the most recent clinical information.”

“We have an open-source technology that we use for the CDR and for our clinical results viewer, so we aggregate data anywhere we deliver care and provide service, and we capture that data in an electronic format and aggregate that data into the repository,” Clark says.
 

‘Research data warehouse’

The Department of Public Health Sciences at the University of Virginia (UVA) School of Medicine in Charlottesville has a different mission for its clinical data repository than the Medical University of South Carolina, but has deployed an equally effective repository system to meet its research needs. 

The CDR within clinical informatics is a large-scale integrated database of clinical and population data optimized for research and not for clinical use, says Jason Lyman, MD, associate professor of clinical informatics and medical director of the CDR of the Virginia Health System.   

“[Our CDR] is really more of a research data warehouse,” says Lyman. “Our system was set up to serve people who wanted to do research on the patients we take care of, [but] had no way of easily doing a population-based query of our system.”

The clinical informatics department designed and implemented a system from scratch that allows users to conduct an ad-hoc query by selecting patients by disease and set as many other parameters that they want, which will bring up de-indentified reports on those patients, he says.

“Data in the CDR are stored on a server with dual 3.6 GHz Xeon processors with 8 GB of main memory and an 8x36 GB RAID Disk Array,” Lyman says. “At present, we have just under 75 GB of user accessible data.” 

Prior to the CDR implementation, “there was no easy way, for example, for a research faculty member submitting a grant for the study of patients with diabetes to find out how many patients here at UVA with diabetes had been seen in the past year, what the demographics of those patients are and what types of medications they are getting," Lyman says. 

The CDR at UVA has approximately 680 users including students, nurses, physicians and researchers. Several measures are in place to ensure that the data are protected against unauthorized access, says Lyman. All health information is stored separately from any patient identifiers. Patient identifiers are stored on a separate server that is behind a secure UVA network that requires dual-factor authentication to access, including a password and a hardware token, he explains.
 

Research, better care and compliance

Rather than being dedicated to clinical purpose or research, the CDR at Geisinger Health System in Danville, Penn., is used for both. The health system — which is composed of two medical center campuses, three hospitals, a nonprofit health insurance company and the Geisinger Center for Health Research — employs a CDR, referred to as a Clinical Decision Intelligence System (CDIS). The CDIS makes accessible a number of different data, including clinical, financial, insurance and patient satisfaction information.

Although the health system has been collecting patient data since 1996, the CDR is just two years old, says Joseph Scopelliti, program director of clinical decision support systems at Geisinger. 

The CDR, developed in partnership with IBM, is helping to move the health system in a more patient-centered direction, says Gail Riley, director of CDIS analytics and consulting. “While each department is good at identifying their own requirement for high-quality patient care, the goal is to build a patient profile so that our patients are contacted once, but all the care they need is identified.” The system allows clinicians and employees from other departments to utilize the database based on their job roles. “Our financial analysts are using it as an easier way of getting data, while administrators are using it to get a quick count on how many patients have a particular diagnosis,” Riley says.

“About 350 users have been trained on the system and we expect that to grow from hundreds to thousands over the next few years,” explains Riley.  

Users must go through training and submit a request to use the repository. “When you request access, we determine your business need,” Scopelliti says. “This determines whether the data are fully identified or de-identified based on your personal access level.” 
 

The CDR of the future

With recent changes to the Stark Law and state-initiated anti-kickback measures, physicians are now able to align themselves more tightly with hospital organizations, not only in academic settings, but also in community-based organizations, says Clark. “The Office of the National Coordinator (ONC) is to trying to bring those care settings together. It makes sense for large hospitals with the infrastructure and staff to reach out to physicians and offer them outpatient clinical functionality.”

When hospitals host the data that are shared in the physician’s office, a seamless IT environment, including a longitudinal patient record, can be developed. “A lot of the big vendors are scrambling to build a repository strategy that cuts across the care setting. I think we are, and will be, seeing more and more of that,” he says.