Heart disease and hospital deaths: an empirical study.

  • 1 August 1987
    • journal article
    • Vol. 22  (3) , 369-95
Abstract
This study examines the effects of selected characteristics of hospitals and physicians on the mortality rates of heart patients who survive their first day in the hospital. Separate multivariate regression analyses are conducted for three groups: (1) patients who undergo a direct heart revascularization or coronary artery bypass graft (CABG) operation; (2) patients who undergo a cardiac catheterization and do not undergo a CABG operation; and (3) patients with a principal diagnosis of acute myocardial infarction (AMI) who do not undergo surgery. The number of patients in each group treated by specific physicians, and the number treated in specific hospitals, measure provider experience with similar patients. Other hypothesized determinants of in-hospital mortality include: (1) patient severity of illness, age, sex, and the presence of comorbidities; (2) hospital ownership, size, location, teaching status, resources expended, and the presence of a coronary care unit; and (3) board certification status of the attending physician or surgeon who operated. Empirical results show that presence of a coronary care unit decreases the chance that CABG patients will die in the hospital but is not significant for other heart patients included in this study. Patients with atherosclerosis who receive a CABG or a cardiac catheterization procedure are more likely to survive in hospitals with high volumes of these procedures. However, hospital volume of AMI admissions was not a factor in survival; AMI patients are more likely to survive when their attending physicians treat high volumes of AMI patients. Also, AMI patients whose physicians are board certified in family practice or in internal medicine are less likely to die compared to AMI patients with physicians not board certified. Similarly, AMI patients hospitalized in teaching facilities are less likely to die compared to AMI patients in hospitals not affiliated with a medical school.