Session number: 1023
Abstract title: Development of an Operational Model for Examining the Delivery of Coronary Bypass Surgery
Author(s):
DS Wakefield - University of Iowa
M Jaana - University of Iowa
GE Rosenthal - Iowa City VA Medical Center and University of Iowa
Objectives: Develop an operational model for the delivery of coronary artery bypass surgery (CABG) to VA patients, as a template for examining delivery of specialized services that are not uniformly provided by all VA medical centers.
Methods: We studied 36,407 VA patients who underwent CABG in the 43 VA facilities with cardiac surgery programs or in private sector hospitals on a contract basis during 1994-99. Data were obtained from the VA Patient Treatment File. We developed an operational model that conceptualized four types of service: I) VA local (CABG performed in a VA hospital in the primary service area [PSA] in which the patient resides); II) outsourced local (VA contract hospitalization within patient’s PSA) VA non-local (CABG performed in a VA hospital outside patient’s PSA); and IV) outsourced non-local (VA contract hospitalization outside patient’s PSA).
Results: Of the 36,407 patients undergoing CABG, 49% (n=17,882) were classified as VA local [I], 2% (n=795) as outsourced local [II), 47% (n=17,118) as VA non-local [III], and 2% (n=612) as outsourced non-local [IV]. Several differences (P<.001) were observed among patients in the four service type categories. The proportions of patients admitted with acute myocardial infarction among types I, II, III, and IV were 8%, 24%, 6%, and 25%, respectively (P<.001), while the proportions of patients discharged to other VA facilities were 2%, 52%, 13%, and 37%, respectively (P<.001). In addition, mortality rates in types I, II, III, and IV were 3.5%, 3.1%, 2.7%, and 2.3%, respectively (P<.001), and mean length of stay was 15.4, 9.5, 13.7, and 7.8 days, respectively (P<.001).
Conclusions: The limited availability of CABG in VA hospitals has led to a substantial amount of outsourcing to private sector and non-local VA hospitals. Differences in mortality, length of stay, and the proportions of patients with acute MI or discharged to other VA hospitals suggests that patterns of service delivery are patient (e.g., acuity) and system factors. These differences support the construct validity of this operational service model.
Impact statement: Declines in the Veteran population will require development of new operational models of service delivery. The current model is likely to be applicable to other specialized services.