The purpose of this proposal is to evaluate the impact of 5 obstetrical care practice guide lines implemented by the Maine Medical Liability Demonstration Project (MLDP) on physician practice patterns. The MLDP (PL 931) is a five year demonstration project aimed at reducing the cost of medical services associated with the practice of defensive medicine and lowering the risk of medical malpractice claims. These goals are to be achieved by providing physicians with an affirmative defense against a medical malpractice suit for incidents arising from medical care practiced within a specific set of practice guidelines. This study will evaluate the parameters, effective January 1, 1992, which concern the management and diagnoses of conditions of labor including: (1) Cesarean Delivery for Failure to Progress; (2) singleton Breech Presentation; (3) Intrapartum Fetal Distress; (4) Assessment of Fetal Maturity Prior to Repeat Cesarean Delivery or Elective Induction of Labor; and, (5) Antepartum Management of Prolonged Pregnancy. Data from Maine and two comparison states, New Hampshire and Vermont, will be used: 1) Uniform Hospital Discharge Data (UHDDS) over nine years, January 1, 1987 to December 31, 1995 (pre = 5 years, post = 4 years), 2) a stratified sample of 1,800 medical records will be abstracted from the same time period as UHDDS data, 3) a survey of all OBGYN physicians, and 4) a survey of hospital OB policies. Analyses will be conducted to measure changes in C-section rates, diagnoses and birth outcomes addressed by the parameters, changes in physician behavior relating to the parameters, estimates of reductions in defensive medicine resulting from these changes, and estimates of the impact of the MLDP on OB malpractice incidents and claims. The attitudes of physicians and the effect of hospital policies on behavior will be examined. Control factors to be assessed in the analyses include: physician specialty, source of payment, age of mother, hospital volume, and hospital policies. Potential confounders will be controlled for in the development of predictive models for physician behaviors using multiple logistic and multiple linear regression techniques. Odds ratios and 95% confidence intervals will be calculated from the multiple logistic regression models for each of the independent variables while controlling for the effects of all other variables in the model.
