Users Mine Repositories for Better Patient Care

Repositories may be the wave of the future, but there’s a sea of obstacles in the way.

Clinical data repositories (CDRs) have allowed for vast improvements over earlier clinical trial methods, says Steven R. Steinhubl, MD, director of Geisinger Health System’s Cardiovascular Wellness Center, based in Danville, Penn. The organization is a member of the HMO Research Network (HMORN), a consortium of 19 healthcare delivery organizations.

In the days of paper records, researchers had to develop case reports and manually track patients to develop a prospective randomized trial. That resulted in one, inflexible stand-alone report. At a fraction of the cost, researchers now can mine existing EHR data and answer endless, emerging questions, Steinhubl says. Anything entered into a patient’s EHR becomes part of the virtual data warehouse, so all operations are “done completely behind the scenes.”

Through HMORN, “instead of enrolling 10,000 prospective patients, we can get a real-world assessment of 100,000 patients and their treatments and outcomes,” says Steinhubl. Say a researcher plans to look at breast cancer outcomes, but then discovers that some patients developed cardiomyopathy with a certain medication. The repository allows researchers to go back and look at the entire patient population to address medication concerns and any other outlier issues that arise.

Steinhubl currently is working to find out how to treat patients who are statin intolerant. Ten to 20 percent of patients cannot tolerate a statin and “we don’t know how else to treat them. It’s an important question and it can’t be answered in a prospective randomized trial.”

“We’re not aiming for sainthood,” says Steinhubl of HMORN’s goals. But, pharmaceutical companies conduct clinical trials first and foremost to sell more of their drugs or devices, not necessarily to investigate why their products aren’t performing well in all patients.

“Most importantly, we can help practitioners no matter where they are. We certainly don’t need large data warehouses to do it. The data we have available can be translatable at every level of care.”

On the ambulatory side, the American Academy of Family Physicians (AAFP) Center for Health IT launched a CDR pilot project in September 2011. Steven E. Waldren, MD, director of the center, says they partnered with a claims processing vendor to collect all claims data from participating practices. Thus, they were able to provide analytics to the clinicians, such as how much of their patient population is diabetic or hypertensive and, on average, how many visits those patients have had. The practices can find gaps in care, determine their quality level and compare costs of care.

The feedback has been quite positive, Waldren says. Access to the data has helped many practitioners get a better handle on their patients, such as realizing the size of their geriatric population. “They’ve seen that as highly valuable.” The capability to add in lab data and medications will “allow us an even more robust set of analytics.”

Other specialty organizations, such as the Society of Thoracic Surgeons and the American College of Cardiology, have developed repositories as well. Such groups have a focused set of diseases to review while primary care medicine is more varied and comprehensive. “We have at least 20 to 25 diseases that we want to track and probably another 100 to 200 that we want to track,” Waldren says. That’s why AAFP has taken the approach of collecting claims data and blending in clinical data rather than creating a freestanding data-entry tool.

Aside from the sheer volume of information, another challenge is establishing clean data. “Even with some structured data required by Meaningful Use, the problem is that standards used still have some flexibility,” says Waldren. Existing standards also are complex so not all vendors implement them correctly. “Even if we get those two correct, the EHR information is not captured in a structured way. There’s work to be done on the capture piece, which is a huge challenge.”

The main goal, Waldren says, is creating a tool that allows physicians to be accountable from both the clinical and the financial perspective. On the clinical side, he wants to make it easier to identify gaps in care. “How can we make sure we’re doing the best we can reliably for all patients?” On the financial side, the goal is to help practices stay viable and continue to provide good care and access, and help them make the most cost-efficient decisions. For example, CDRs can help a clinician weighing two medication options make the most effective choice for a particular patient’s needs.
The repository also offers the capability to demonstrate primary care outcomes. “People are starting to realize that the more primary care doctors you have, the more quality you have and the more cost effective the care is. We’re trying to determine how to demonstrate that and do even better,” Waldren says.

The shift to value-based payment is inevitable, he adds. “To work in that environment, we need to have good clinical and financial data and tools to understand it, visualize it and take action on it. That’s one reason the academy is pushing forward on the repository.” Another reason is that “we’re moving into an era of data. To have data analytics allow for population-based care management, we need to make sure we have the tools that really work for providers.”

The CDR is the “perfect infrastructure” to run clinical trials, Steinhubl says. The health record plus the genetic data that will be available in coming years is going to be the primary driver of the future of clinical trials. Meanwhile, the repository will drive innovation in medicine. “Networks similar to HMORN are going to be dominating medical research. It’s going to be the dominant way we answer questions in the future.”