Society of Thoracic Surgeons' National Database Effort Comes of Age

Reed Miller

January 26, 2010

January 26, 2010 (Fort Lauderdale, Florida) — The 21-year-old national database of the Society of cSurgeons (STS) is about 97% complete and ready to support longitudinal research on the outcomes of cardiothoracic surgery, according to results of a study that matched STS database records to records kept by the Centers for Medicare & Medicaid Services (CMS) [1].

On January 25 here at the STS 2010 Annual Meeting, Dr Jeffrey Jacobs (University of South Florida, Tampa) presented results of a study by the Duke Clinical Research Institute that matched records in the STS database and CMS database for coronary bypass graft (CABG) hospitalizations of Medicare patients over 65 from 2000 through 2007. The STS database was launched in 1989 and now has over 1000 participating surgeons and data on almost four million operations, but there are wide geographical variations in participation. In the Midwest, almost 100% of surgical sites participate in the database, while the participation rate is under one-third for sites in parts of northern New England and about 75% in Florida, Jacobs said.

The purpose of the study was to assess the feasibility of linking the STS database to CMS data and to examine the penetration, completeness, and representativeness of STS data on adult cardiac surgery, Jacobs explained.

"The STS database is in the process of being transformed into a platform to provide information about long-term outcomes with cardiothoracic surgery," Jacobs said. "This study was designed to help lay the foundation for us having the ability to [collect that data]."

Researchers have already begun new studies using the linked STS and CMS databases to study coronary disease and heart-valve disease, and there are plans for a study of outcomes of atrial fibrillation to begin soon, Jacobs said. "So this is just the tip of the iceberg of the power we'll have from this linkage to look at long-term outcomes, comparative effectiveness, and health economics." In the future, the STS may also link the STS database with other large databases such as the master Social Security death database or the database of congenital heart disease data within the Pediatric Health Information System maintained by the Agency for Healthcare Research and Quality, he said.

Last Pieces of Puzzle Cause Concern

The study evaluated the completeness of the STS database by dividing the number of CABG cases in the CMS database for which there is a corresponding record in the STS database by the total number of CABG cases in the CMS database. For 2000 through 2007, there were 91 363 CABG surgeries performed in the Medicare population, of which 88 857 are in the STS database, for a completeness rate of about 97%.

The study also linked STS records to CMS records to calculate the center-level penetration, defined as the number of sites in the CMS database that have at least one matching record in the STS database. Center-level penetration increased from 45% in 2000 to 78% in 2007 and is believed to now be over 90%, Jacobs said. In 2007, there were 1101 hospitals doing bypass procedures represented in the CMS database, of which 854 were also in the STS database. Patient-level penetration--the total number of CABG cases in the CMS database with a matching STS record divided by the total number in the CMS database--increased from 51% in 2000 to 84% in 2007, and that number is now probably over 90%, Jacobs said. As of 2007, 94 409 out of 111 967 CABG records in the CMS database matched an STS record.

The researchers evaluated the representativeness of the STS database by comparing data from the CMS records that were not matched to STS records with the CMS records that did have a STS match. They found that the unmatched cases were more likely to be black patients and/or emergency operations. They also discovered that the cases excluded from the STS database were more likely than the included cases to be associated with mortality.

Some of the explanations for these discrepancies, according to Jacobs, could include potential selective underreporting of patients with poor outcomes. Also, Jacobs suggested that "the STS database may fail to capture patients who are not part of routine pathways or routine plans of care, secondary to a lack of systematic protocols to capture these nonpathway patients--for example, patients who die in the operating room or patients who have a very long hospitalization."

Jacobs acknowledged that although these unmatched cases represent only about 3% of the CABG cases conducted at participating hospitals, understanding why they weren't reported in the STS database and how their exclusion may skew conclusions from the STS data is of "extreme importance."

To fill in these gaps, the STS has been ramping up its audits of participating sites to encourage complete and accurate submission of data. In the last year, the organization has begun to increase both the number of sites that will be audited as well as improve the "mechanisms of audit," he said. The audits are an "important mechanism to help sites identify where they're missing cases and why those cases aren't getting enrolled and perhaps to help those sites develop mechanisms to solve the problem and to capture these cases that are not currently being captured."


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.