Use of statewide administrative database in assessing a regional trauma system.

Abstract

Listen

Translate article

I have read the article authored by Dr Reilly and colleagues comparing outcomes at trauma centers and nontrauma centers in New York City. I am surprised to see such an important conclusion, “discharge from a trauma center exerted an unfavorable impact on adjusted mortality risk,” derived from an administrative New York Statewide Planning and Research Cooperative System (SPARCS) database. The much-heralded New York State Cardiac Surgery Cardiac Reporting System (CSRS), the inventor of “report card medicine,” discarded the unreliable and nonaudited administrative data contained in the SPARCS database 14 years ago. In reference to the SPARCS system, the seminal report concluded, “it appears that the CSRS (Cardiac Surgery Reporting System) data provide a significant improvement in comparison to typical discharge or claims data.” Adoption of a clinical database marked the end of the use of federal or administrative data in reporting of cardiac surgical results in New York State. All reports generated from the New York State Department of Health about cardiac surgical outcomes have relied on a clinical database, collected by clinicians and audited by the Department of Health. In the article by Reilly and colleagues, the claim that SPARCS data were used to formulate the 1993 to 1995 Coronary Artery Surgery in New York State Report is false. Cardiac surgeons have realized that important conclusions about operative survival require careful analysis of a clinical database, whether it be the New York State CSRS, the Northern New England Cardiovascular Disease Study Group, the Veterans Administration Cooperative study, or the flagship Society of Thoracic Surgeons database, with over 2 million patients entered. Despite the disclaimer in the title and the body of the article, I am concerned that the important outcomes conclusions about the treatment of trauma patients are drawn from unreliable data. Might it be more cautious to do the hard work of creating and maintaining a valid clinical database before significant conclusions are developed?