The proposed research seeks to determine the institutional characteristics that lead to variations in medical practice patterns. This objective will be achieved by investigating the relationship between elements of health care organization and the use of cesarean section. Ideal for evaluating variations in practice patterns, cesarean section is the most common surgical procedure in U.S. hospitals. Considerable controversy surrounds its increasing use. Past research suggests that health care organization has a strong influence on cesarean section utilization. Previous studies, however, have failed to determine the independent effects of health care organization variables and have rarely controlled for the confounding effects of patient characteristics. The current study will fill an unmet need for research through an innovative approach to the analysis of cesarean section use. A database on obstetrical practices will be developed by combining information from hospital discharge abstracts and birth records for the 470,000 deliveries occurring in California in 1985. These two data sources are necessary because each contains important variables absent from the other. Despite potential drawbacks, the use of existing public data is both efficient and statistically powerful. A multivariate model of cesarean section use based on logistic regression will be developed to determine the independent effects of health care organization variables. Case-mix adjustment will include the medical characteristics of patients, an improvement over the use of demographic characteristics alone as proxies for the risk of cesarean section. This research will contribute to the study of variations in medical practice. It will also clarify current confusion regarding patterns of cesarean section use, evaluate potential policy strategies for controlling the increasing use of cesarean section, and develop general methods for employing public data in health services research.
