The collaborative effort will use Archimedes' ARCHeS Simulation and Analytics software to analyze the recently published Medicare Data Entrepreneurs Public Use File (DE-SynPUF). This is a set of free downloadable files containing a subset of claims data representing a "synthetic" 5% sample of Medicare claims from 2008 to 2010.
CMS calls the data synthetic because it has not only been de-identified, but it has also been altered to safeguard the privacy of the patients involved, said Mika Newton, VP, ARCHeS and consulting business leader for Archimedes. For example, it does not contain information about geographical location, and it mixes certain kinds of clinical data, such as blood pressure and cholesterol level, among different patients. In the aggregate, however, it represents what the patient population looks like and retains the data structure of the original claims information, he said.
[ Electronic health record (EHR) companies are beginning to partner with outside software developers. Read EHR Vendors Tap Software Developers To Expand Reach. ]
Software developers and researchers can use ARCHeS to query and analyze the rates of various outcomes for specific Medicare populations, such as those with a particular health condition, age criterion, medication use, or a combination of these factors. This can be very powerful, because the dataset represents millions of patients, Newton pointed out.
Health IT firms and researchers can run ARCHeS Population Explorer against the DE-SynPUF, which is freely available on the CMS website, to determine the viability of particular solutions or research projects under development, he said. This can be useful in applying for a grant or deciding whether an idea is worth pursuing before a lot of time and money is spent on a pilot.
For example, he said, the average inpatient cost for patients with diabetes exceeds that for patients without the disease or with a less severe form of it by a certain amount. If a solution is designed to make it easier to control diabetes, one can infer from the difference between these cohorts how much that application might save.
Archimedes built the ARCHeS software with funding from the Robert Wood Johnson Foundation. Its original goal was to help the government or guideline makers decide whether to pursue certain healthcare policies. For example, Newton noted, if policy makers are considering whether to reduce the amount of calories in restaurant food to reduce obesity, they could use ARCHeS to model how much money or resources that would save. If the Joint National Committee that sets guidelines for treating hypertension decides to change its recommendations, it can use ARCHeS to see how much that would cost and how it might affect outcomes.
Software developers that want to use ARCHeS Population Explorer can download the basic version for free to view the synthetic Medicare claims data. If they want to use the analytics within their own application or build a solution around it, they must pay a license fee, Newton said. Archimedes charges government agencies only a fraction of what the ARCHeS license costs commercial firms, he added.
Another Archimedes application, indiGO, is being used by a number of healthcare organizations to predict the chances of patients developing a serious chronic condition. It also suggests to patients and clinicians the best interventions for warding off heart attacks, strokes, or other serious adverse events. Among the organizations that have adopted indiGO are Kaiser Permanente, Intermountain, the Marshfield Clinic, Fairview Health Services, and the Colorado Beacon Consortium.
Archimedes will demonstrate ARCHeS Population Explorer immediately preceding the fourth annual Health Datapalooza, scheduled for June 3-4 at the Omni Shoreham Hotel in Washington, D.C.