The explosion of medical care costs within the United States has forced a careful examination of current health care practice. Health services researchers have identified wide variation in management strategy for adult diseases, but few studies have been performed in children. Evaluating variability in patient outcomes or resources use may suggest improved methods for providing high-quality, cost-effective patient care. However, understanding sources of variability requires appropriate adjustment for """"""""illness severity"""""""", or individual patient risk. This project studies health services of pediatric patients undergoing surgical correction of congenital heart defects. Our preliminary data from Massachusetts and California suggest that there is sufficient variability in mortality and resources use for these patients to warrant further study. Care for theses patients is specialized and expensive; mortality rates are high.
Our specific aims are as follows: 1) To study variability in clinical outcome and resource utilization, we will perform a population-based study using discharge abstract data from nine states for the calendar year 1992 for all patients under 18 years of age undergoing repair of congenital heart defects. Specifically, we will examine variations in mortality, morbidity, care strategy, total hospital contribution of patient and hospital characteristic to variation in these outcomes. 2) To develop methods for risk adjustment for in-hospital mortality, we will develop and validate a multi-dimensional method of risk adjustment for in-hospital mortality using a """"""""pane of experts"""""""" approach. Performance characteristic of the method will be assessed using a c-statistic. Calibration will be assessed by the ability of the model to correctly place patients into deciles of risk. Empiric methods of risk adjustment for length of hospital stay and total hospital charges will be derived and internally validated from the clinical data. Performance of these methods will be assessed using R2.
