One of the more interesting recent health policy events took place at the Brookings Institution’s Engelberg Center for Health Care Reform. The December 2nd event did not focus on the reform bills in the House and Senate or the horse trading and sausage making that has gone into them.  The panels and discussion, titled “Using Data to Support Better Health Care: One Infrastructure with Many Uses,” chaired by Mark McClellan, focused explicitly on health information technology (IT). 

Rather than another dreary recitation of how poorly the United States measures up to other developed countries or to systems such as Kaiser, Geisinger, or Mayo, the first round of panelists–L. Allen Dobson of Carolinas Care, Richard Platt of Harvard Medical School and Harvard Pilgrim, and Alan Garber of the VA and Stanford University–focused their comments on specific areas where health IT is being used and holds great promise.

Dobson’s presentation featured data from Community Care of North Carolina, one of the more exciting primary care programs ongoing.  It started as a Medicaid project but has now extended statewide to provide primary care services in a variety of settings, urban and rural, doctors’ offices, clinics, and hospital systems. Dobson pointed out that in most primary care settings, the United States is under-resourced. Primary care is largely a cottage industry, and creating medical homes or putting in an IT system alone won’t address the shortcomings.  What primary care providers need from IT systems is data that are actionable–that is, provider performance summary data that can identify gaps in care and spur efforts to address them in a timely manner. Dobson found that once data were available in North Carolina, providers engaged in a kind of competition team sport to achieve high quality.

Richard Platt addressed the multiple systems that are now in place that provide data and are getting closer to real-time data in population health, such as adverse events for vaccines and immunizations, notifications of disease, flu surveillance, and disease prevalence. IT systems have to pull together these reporting points of interest from a variety of electronic reporting systems, claims data from defined populations, electronic medical records (EMRs), registries, and public health reporting.  For Platt, what is needed for these systems to work and to continue to grow are flexible and reusable data that are stored accessibly, as he put it, in the “virtual warehouse” like a Toys “R” Us store.  If you enter a Toys “R” Us facility in any city or state, the same type of merchandise is always displayed in the same aisle.  A flu surveillance researcher should be able to access the same type of data from different systems in the same way.

Alan Garber highlighted the great promise of health IT, particularly with data from EMRs to conduct top-flight observational studies to aid comparative effectiveness research. As is well known, not every condition or intervention is suitable for a randomized clinical trial even if we could finance them all.  Well-conducted observational studies with standardized data on real patients with varying degrees of co-morbidities, persistence, and compliance and with practitioners in settings outside of academic medical centers can enhance our evidence base for comparative effectiveness research and complement randomized clinical trials. Many more questions can be addressed this way, and the research can be more directly tied to quality improvements. He also pointed out something that should warm the hearts of all bureaucrats–that with health IT systems in place, observational studies offer the promise of rapid implementation, and costs are likely to fall.  Garber did point out that there are still some regulatory bumps in the road, some tinkering with HIPAA regulations that would still guarantee patients’ privacy but allow researchers to get the data elements they need for population studies.

Perspectives From The Federal Government

The second panel offered input from the federal government’s perspective with Janet Woodcock, director, Center for Drug Evaluation and Research of the FDA; Carolyn Clancy, director of AHRQ; and David Blumenthal, the National Coordinator for Health Information Technology at HHS. The federal representatives focused on what are the new standards we need to get to for methods as we build a robust health IT system.

Woodcock outlined ongoing efforts at the FDA on the reporting of adverse events for vaccines and immunizations through the PRISM system and a new pilot, the Observational Medical Outcomes Pilot (OMOP), to find the best methods, data, and models for recognition of cues and for the signaling of adverse drug events.

Clancy talked about the importance of data aggregation. With apologies to the audience for the “coma-inducing” nature of the subject, Clancy discussed AHRQ’s efforts to build and maintain distributive data networks, AHRQ’s continuing investment in registries, and its discussions about creating a registry of registries. She also challenged the journal community to update its peer-review methods, as the health IT community continually looks for the best methods to collect, store, and translate data.

Blumenthal’s remarks suggested that the vision of health IT held by the administration far surpasses the notion of catching up to the other developed countries. Blumenthal called for laying the groundwork for “learning health care systems,” so that at every point of contact–patient, provider, community, state, and federal government–there are opportunities to learn. He noted that no foreign model exists for this type of learning system; other countries may have more-robust electronic records systems than the United States has, but the United States has a unique opportunity to build a learning system with robust capabilities to capture data, aggregate it into common data elements, and use it for feedback and learning.

To extend Richard Platt’s metaphor, it would be like going to school at Toys “R” Us, and the price scanners would actually work. Furthermore, the scanners and bar codes could tell us if the toys survive for longer than a week and if they are actually suitable for the ages recommended. Happy Holidays!