Learning by Doing in CABG Surgery

This study provides empiric confirmation of the widely held belief that coronary artery bypass graft (CABG) surgery has become quicker over time as clinical and organizational improvements diffuse across hospitals. Based on a two-stage cluster sample of 74 hospitals performing CABG surgery in 1984, data on skin-to-skin procedure and pump times were analyzed for 2,784 Medicare- eligible cases over the 1972-84 period. Using multivariate regression analysis to adjust for patient characteristics and number of grafts, we found that CABG skin-to-skin times fell an average of 36.5% during the first 12 years of hospital experience while cardiopulmonary pump times fell 28.2%. Rates of decline averaged 4.5-5.4% annually initially, but tapered off to 1.5% by the twelfth year of experience. Extrapolating over a 20-year period, gains in skinto- skin times are projected to reach 42%, resulting in nearly a halving of surgical times. Our findings have implications both for physician reimbursement and patient outcomes. Medicare CABG prevailing fees do not reflect these gains in efficiency and could be reduced substantially. Alternatively, efficient, high-volume hospitals could be identified for inclusion in a Medicare “CABG PPO” so that society might share in the productivity gains. Finally, recent research showing better outcomes in high-volume institutions may be partially explained by their shorter procedure times, which should reduce operative complications.