EHRs & CVD: The Good, the Bad & the Future

Early EHR adopters are showing how they can use health IT to improve the outcomes of patients with or at risk for cardiovascular disease (CVD). There is potential for big savings and improved health, but interoperability issues and payment structures currently hold them back.

The good

While conducting a phased EHR implementation between 2005 and 2008, Kaiser Permanente Northern California (KPNC) wanted to ascertain whether digitization was having a neutral, positive or negative impact on patients’ health outcomes. Researchers examined blood glucose and cholesterol levels of approximately 170,000 patients listed in the health system’s diabetes clinical registry by the end of 2003. They chose those two measures because high levels of glycated hemoglobin (HbA1C) and low-density lipoprotein cholesterol (LDL-C) are risk factors for CVD.

“There have been a lot of questions about whether an EHR would make a difference in patients’ lives and, if it did, whether it would be helpful, harmful or neutral,” says study co-author Marc Jaffe, MD, clinical leader of the CV risk reduction program at KPNC in San Francisco. “Our study demonstrates that it is useful.”

Among KPNC diabetics, a group that already demonstrated strong baseline performance against CV care quality measures, patients with EHRs were more likely to see greater reductions in HbA1C and LDL-C than those without (Ann Intern Med 2012;157[7]:482-489). On average, EHRs were associated with a 0.14 percent greater reduction in HbA1C values among patients with a baseline value higher than 9 percent and a 0.06 mmol/L greater reduction in LDL-C levels among patients with a baseline value higher than 3.4 mmol/L.

Patients also were more likely to return to KPNC for retesting in a shorter time frame after EHR implementation. For instance, 83.5 percent of patients with LDL-C levels greater than 3.4 mmol/L returned for testing within one year pre-implementation, and 86.7 returned within one year post-implementation.

EHRs also helped patients with HbA1C and LDL-C levels below 9 percent and 3.4 mmol/L perform better against the selected outcomes. But the most significant impact was among the group with levels above those thresholds, which Jaffe and his colleagues considered indicators of especially high-risk CVD. EHRs “focused the organizational energy on those farthest from meeting their goals,” Jaffe says.

He attributes the increased rates of retesting and medication adjustments to EHRs, which allowed clinicians to better identify patients overdue for retesting and to track the efficacy of medications. This, in turn, led to better patient performance against established quality measures.

The reductions in HbA1C and LDL-C blood glucose and cholesterol levels are modest, “but when you magnify this over a population of 170,000 patients, it’s going to translate into a meaningful reduction in cardiovascular disease,” Jaffe says.

Similar results have been achieved elsewhere. Patients of 204 physicians using EHRs from the Fishkill, N.Y.-based Taconic Independent Practice Association (TIPA) met a Healthcare Effectiveness Data and Information Set measure for HbA1C testing 90 percent of the time compared with patients of 262 physicians without EHRs who met the measure 84 percent of the time (J Gen Intern Med online Oct. 3, 2012).

“If you have an EHR, you have an opportunity to hit some of the screening recommendations that your practice has decided are key to improving outcomes,” says Susan Stuard, MBA, executive director of the Taconic Health Information Network and Community (THINC), a Fishkill nonprofit organization that sponsors regional health information exchange (HIE) and promotes EHR adoption.

The bad

While some organizations have utilized EHRs to improve patient outcomes on intermediate measures, one criticism often aimed at the costly systems is that they should be capable of more than automated reminders.

In a 2012 working paper, “Using EHRs to Monitor and Improve Adherence to Medication,” researchers from the Mayo Clinic in Rochester, Minn., and North Carolina State University in Raleigh estimated that a nationwide system of fully integrated EHRs with real-time medication monitoring capabilities would each year save $1.41 billion while adding 131,000 quality-adjusted life years (QALY) to patients recently diagnosed with diabetes.

There are barriers to achieving these kinds of results. It would require provider access to uninterrupted, longitudinal streams of patient data from multiple sources and the capability to rapidly collect and combine such data, says the paper’s co-author, Nilay D. Shah, PhD, an associate professor of health services research at Mayo. The paper was written in part to demonstrate the value proposition of real-time medication monitoring programs to payers.

Misaligned financial incentives currently contribute to the low level of data exchange. For instance, payers may be deterred from investing in sharing of claims data to improve the long-term outcomes of patients who may leave their health plan in the next several years. “For patients, the goal is to maximize QALY,” Shah says. “But for payers, the goal is to minimize costs. Payers have an approximate  20 percent member turnover rate every year, so you’re not only talking about saving costs, but saving costs in a short period of time.”

There are legal challenges and technical challenges. Even if payers were willing to make the investment, Shah says, “The question becomes, how do you get data back to providers in a meaningful and secure way?” A provider organization would have to deal with the logistical nightmare of constructing electronic communication routes with all of its patients’ health plans. They would have to maintain compliance with HIPAA and overcome interoperability issues associated with disparate EHR and information systems.

These challenges likely will prevent the full exchange of claims data anytime soon, leaving providers like TIPA to make the best of what they have. “You need to find the best and most appropriate way for your community and set of providers to move forward,” Stuard says. “In the absence of a single payer system, this is where we are.”

The future

The healthcare community is still learning how to best use EHRs. Payment reforms encouraging multidisciplinary partnerships to improve interoperability standards hopefully will help stakeholders move up the learning curve.

With the help of THINC, TIPA is currently entering a partnership with six health plans to form a patient-centered medical home and become eligible for the Centers for Medicare & Medicaid Services’ shared savings program, which allows providers to receive higher payments for delivering high-quality healthcare. Stuard believes it will take some time to develop best practices for using EHRs and sharing electronic health data, but she’s confident that progress will be made. “We’ll be wrangling with harmonizing measurements between disparate practices and systems. We’ll learn a lot and it will be difficult at times, but it’s an opportunity for everyone.”

To facilitate data sharing among these partners, Shah believes that requirements supporting data sharing between providers and payers could be written into Meaningful Use Stage 3 rules.

Current research also focuses on populations rather than individuals. In the future, smart EHRs could determine appropriate courses of care for patients based on their individual needs rather than what has worked across a population.

“That’s the $1 million question,” Jaffe says. “Can the medical record really know me and my unique needs? We are really working hard to have info reflect what your unique needs are as an individual.”

Early EHR adopters like KPNC show how integrated health systems and integrated health IT systems can improve the quality of care provided to patients with or at risk for CVD. As EHRs mature, more positive outcomes are likely to be achieved and huge cost savings could follow. “We’re trying a lot of different things and learning a lot from our implementation,” Jaffe says. “We’ve come a long way and have a long way to go.”