For the last four to five years, the Centers for Medicare & Medicaid Services (CMS) has been very focused on releasing data sets that protect the privacy of Medicare beneficiaries while giving outside entrepreneurs and other interested data users an opportunity to use the data to enhance their understanding of the Medicare program.
Enter Archimedes, Inc., a San Francisco-based healthcare modeling and analytics company, which recently announced it has begun collaborating with the U.S. Department of Health and Human Services (HHS) and CMS to provide unprecedented access to synthetic CMS claims data.
Archimedes will facilitate the analysis of the Medicare Data Entrepreneurs’ Synthetic Public Use File, DE-SynPUF, dataset through its ARCHeS Population Explorer, a web portal that will help developers analyze the potential effectiveness of particular software solutions. Access to CMS Linkable 2008-2010 Medicare DE-SynPUF provides researchers and data entrepreneurs a glimpse into the wealth of information contained in the CMS claims data. All data has been completely de-identified to maintain patient privacy.
In an exclusive interview with Healthcare Informatics, Niall Brennan, director of information products and data analytics at CMS, says these files are Medicare synthetic public use files that have been a culmination of CMS’ efforts over that four to five year period. The data set is built on a five-percent random sample of Medicare beneficiaries and their claims from 2008 to 2010. “These files look like Medicare data, but have been synthetically seeded so they don’t represent any one person’s healthcare experience. They are not real people, but instead an accumulation of different real people. You take one claim from one person and another claim from a different person, Brennan explains, and you create a synthetic person who looks like a Medicare beneficiary with a Medicare claims history but actually isn’t a real person.”
David Eddy, M.D., Ph.D., founder of Archimedes, said in a statement that the company’s modeling and analytics tools can help researchers, life-science organizations, government agencies, payers, and providers use the data to make informed decisions for the benefit of their constituents. The beauty of ARCHeS PE is that you do not need to be a statistician in order for you to benefit from the data, said Eddy.
The idea, according to Brennan, is that data entrepreneurs and researchers will be able to use the synthetic data to develop research protocols that will ultimately help vendors make informed decisions based on the data. This will provide free and easy access to the agency’s big datasets for any software developer (commercial, private, or individual) interested in assessing the viability of a healthcare solution in a given patient population (heart or diabetes care, for example). The objective of the program is to help expedite the successful development of healthcare technology software and solutions that have the potential to improve an individual’s quality of care while lowering costs through the use of analytics.
DE-SynPUF is a set of free downloadable files containing a subset of the data fields contained in inpatient, outpatient, health insurer and prescription drug event claims and beneficiary summary files. Delivering big data sets such as DE-SynPUF has long been a challenge, but Brennan says this gives outside entrepreneurs a really unique and easily accessible way to better understand and use the data. “It is difficult to access our patient identifiable data, and one of the problems is that when users finally get access to all our protected data, it’s very complex and overwhelming, and can take months to understand all of it. This is a way to play with our data, understand it, and develop emerging tools and products that can help both vendors and providers if they were to actually get access to the real Medicare data,” Brennan says.
This data gives developers and entrepreneurs a way for them to detect solutions, which can be quickly implemented and become real once they get access to actual Medicare claims data. A real life example, says Brennan, might involve an innovative analytics company that wants to support accountable care organizations (ACOs) in managing populations from a population health perspective. “A smart company would be downloading this file, building models off the file, and if they do enter into a business agreement with an ACO and get access to the real data, they will be able to hit the ground running.”
At CMS and HHS, and across the healthcare system, most people have come to conclusion that data is fundamental to improving care and lowering cost, Brennan says. “[Now], folks can track these synthetic beneficiaries over time to see where hot spots may occur. The goal for this data is to educate and enable folks to be able to properly leverage actual Medicare claims data once they are able to access it.”
