This past September, HHS launched the Million Hearts initiative, a new public-private partnership aimed at preventing one million heart attacks and strokes across the next five years. Million Hearts will pursue this prevention goal through a focus on ABCS: Aspirin for people at risk, Blood pressure control, Cholesterol management and Smoking cessation.
Early statements and policy decisions indicate that health IT will play a vital role in this initiative. Indeed, HHS' official announcement contained a telling line: "Ongoing HIT improvements will increase focus on cardiovascular prevention and give providers improved tools for their delivery of lifesaving ABCS care."
In a New England Journal of Medicine article accompanying the Million Hearts initiative's launch, CDC Director Thomas Frieden and then-CMS Administrator Donald Berwick noted, "Expanding use of prevention-oriented [electronic health records] will enable providers and health systems to track and improve management of the ABCS."
We agree. The American College of Cardiology has long supported the broad goals outlined in Million Hearts. We welcome the focus on core measures and the recognition of the promise of health IT. But it is equally important to begin with an understanding of the limitations of health IT, as currently deployed (or, often, not deployed) in the U.S., and to view wider EHR adoption as a means to an end, not as an end unto itself.
ACC's PINNACLE Registry is a clinical care database covering outpatient cardiology patients with coronary artery disease, atrial fibrillation, heart failure and hypertension. The registry now boasts more than 2.3 million patient encounter records, the vast majority of which were extracted from EHR systems. Since early 2010, the PINNACLE Registry has required electronic collection and submission of patient records for all new participant practices.
We are familiar with the challenges of integrating disparate data elements, collected across different systems and in different care settings, as well as with the challenges associated with producing harmonized performance reports that are likely to impact progress measurements and, more importantly, quality improvement efforts in Million Hearts. Three traps often emerge when attempting to measure -- and improve -- care delivery using health IT: focusing on too many measures; a lack of standardization; and unrealistic expectations that health IT is the silver bullet.
A Targeted Focus
The first trap when using health IT to measure and improve care quality is the layering of complexity. The PINNACLE Registry regularly reports performance on more than two dozen outpatient measures to our participating providers. Some measures require the extraction of numerous discrete data elements from EHRs, both for assessing patient eligibility for inclusion in the measure reporting denominator and for meeting performance itself.
All of these measures are important and all are based on evidence. While it is important to push the boundaries of our health IT capabilities, it is equally important to offer providers information that is useful and actionable. At times, a constant overemphasis on all measures can impede a focus on core measures.
In this light, linking the Million Hearts prevention goal with quantitative progress on ABCS goals is encouraging. ABCS progress, in turn, will be assessed using Physician Quality Reporting System measures. This focus on a few key measures is laudable. These measures should prove "simple and consistent," to borrow Frieden and Berwick's description; comprehensible across a wide range of care settings; and broadly applicable to wide patient pools.
Further, a focus on these measures should prompt cost-effective interventions. Blood pressure management, aspirin, smoking cessation and statins all are relatively inexpensive, particularly compared with emergency department admissions, hospital stays, interventions and rehabilitation associated with acute cardiac events.
Data Harmonization
The second trap we consistently encounter in EHR data extraction is a lack of standardization, even when e-specifications exist for performance measures. Some differences reflect different understandings of disease states among providers or differently classified borderline or potentially transient cases.
With many performance measures, little room exists for shades of gray. Many performance measures have binary applications (i.e., patients are either included or excluded from eligibility). But differences also can be introduced through distinct EHR systems, different customizations of systems, and varying uses of systems by different providers within the same practice sites and EHR installations.
This is not to say that data harmonization should delay efforts to collect, measure and provide performance feedback, either for individual providers or at the national level. If anything, the need for simple, consistent and clearly articulated measures in Million Hearts bolsters the case to focus on a small set of key PQRS measures, with clear links to patient outcomes.
Using Data To Power Education Interventions
The final trap is a tendency to view health IT itself as a cure-all. Health IT adoption alone will not produce better adherence to care guidelines or better patient outcomes. Rather, health IT is a tool for driving performance improvement. Engagement with core guidelines at the point of care is one way this occurs. Clinical decision support is an often cited buzzword, but it, too, is far from a panacea. Our providers often tell us of pop-up fatigue and colleagues who quickly click through electronic notifications and alerts.
The problem here is not the providers but the alerts themselves. Health care providers welcome information that aids them, but they understandably tend to be less enamored with irrelevant obstacles to entering patient data into EHRs or to closing out a given patient encounter. Balancing evidence-based guidelines with workflow demands is crucial to effective clinical decision support. Here, too, a focus on key, broadly applicable performance measures could help deliver useful information to health care providers when they need it.
Data collected in EHRs and used to calculate performance measures can and should be used to drive the next generation of educational interventions. ACC already offers several Performance Improvement-Continuing Medical Education programs based in providers' own clinical registry data. These types of programs, based on actual data, are increasingly possible when coupled with EHR data collection and extraction, and they can truly move the dial on performance.
Lastly, data on performance at the practice, location and provider levels can be used to identify and create modular best practices, focused on ABCS.
The Million Hearts Opportunity
The Million Hearts initiative has the potential to positively impact hundreds of thousands of patient lives each year. It is a chance for specialty societies and primary care providers to learn from each other. Million Hearts also is an opportunity for a practical demonstration of health IT merits, focused on a set of key measures, standardizing data elements and contributing to data-driven education and practices.
To achieve these goals and demonstrate the value of health IT, we need to avoid overreach, understand limitations up front, guard against unrealistic expectations, and balance data collection against workflow and standardization within the realities of care delivery.