Medicare Fee for Service (FFS) claims annually cost the country $300 billion. What are the most expensive diagnoses? Costliest hospital revenue centers? In a three-year $150,000 project, the LMI Research Institute developed a database to answer such questions.
LMI’s FFS database contains more than 250 million records, derived from 2007–2010 Medicare Standard Analytic Files (SAFs), which contain a 5 percent longitudinal sample of Parts A and Parts B claims data. A nine-person team spent two years developing processes for quickly obtaining, cleansing, formatting, and loading the SAF data into a Microsoft Excel–accessible database.
Securing Protected Health Information
Building the database wasn’t just about gathering and sanitizing records. As LMI expanded the data collected to include geography, provider-specific information, age, and other demographics, it also invested in processes and infrastructure to securely house that data. “As you move from high-level summary data to more specific information like geography, the data becomes protected,” noted Sam Mallette, LMI Senior Consultant and Health Systems Chief Information Engineer. “We implemented procedures like two-factor authentication to secure the protected health information.”
Identifying High-Cost Areas
LMI can use the database to identify high-cost healthcare areas. “It helps people better understand how to bend the cost curve, so they can identify opportunities to cut costs without sacrificing quality,” said Mallette.
For example, the data show that a coronary bypass costs Medicare $33,347 per claim. “Why does it cost that? Are there ways to provide the same service at less money?” he asked. Such understanding can focus attention on process improvements rather than behavioral changes to lower costs.
Accountable Care Organizations
LMI is also using the database to help the Centers for Medicare & Medicaid Services (CMS) assess performance of Accountable Care Organizations (ACOs). “Each ACO has a different risk profile—patients with different diseases, overall health level, etc. It would be unfair to compare them directly,” explained Mallette. LMI is using the claims data to develop a reliable mathematical model to enable CMS to assess each ACO’s performance given the risk profile of the patients it serves, and compare that performance to known costs from the FFS database.
Development of the Medicare FFS database was natural for LMI, which has been working with CMS’ managed care for nearly 20 years. “We are unique because we have knowledge of more than one type of healthcare delivery model. No one comes close to matching us in our knowledge of managed care data, and now we have fee-for-service claims data expertise,” Mallette said.